Thermal cycler for automatic performance of the polymerase chain reaction with close temperature control

~ 2~7~3 .

-- 1 - THERMAL CYCLER FOR AUTOMATIC PERFORMANCE OF THE PO~YMERASE CHAI~ REACTION WITH CLOSE TF~PERATU~ ÇO~TROL Backaround o~ the Invention The invention pertains to the field of computer directed instruments for performing the polymerase chain 10 reaction (hereafter PCR). More particularly, the invention pertains to automated instruments that can perform the polymerase chain reaction simultaneously on many samples with a very high degree of precision as to results obtained for each sample. This hi~h precision provides the 15 capability, among other things, of performing so-called "quantitative PCR".

To amplify DNA (Deoxyribose Nucleic Acid) using the PCR process, it is necessary to cycle a specially constituted liquid reaction mixture through a PCR protocol including 20 several different temperature incubation periods. The reaction mixture is comprised of various c~ ~ nents such as the DNA to be amplified and at least two primers selected in a predetermined way so as to be sufficiently complementary to the sample DNA as to be able to create extension products 25 of the DNA to be amplified. The reaction mixture includes various enZymes and/or other reagents, as well ~s several deoxyribonucleoside trip~osphates such as dATP, dCTP, dGTP and dTTP. Generally, the primers are oligonucleotides which are capable of acting as a point of initiation of synthesis 30 when placed under cond~itions in which synthesis of a primer extension product which is complimentary to a nucleic acid strand is induced, i.e., in the presence of nucleotides and inducing agents such as thermostable D~A polymerase at a suitable temperature and pH.

The Polymerase Chain Reaction (PCR) has proven a 2~567~ - 2 phenomenally successful technology for genetic analysis, largely because it i~ so ~imple and reguires relatively low cost instrumentation. A key to PCR is the concept of thermocycling: alternating steps of melting DNA, annealing 5 short primers to the resulting single ~trands, and extending those primers to make new copies of double stranded DNA. In thermocycling, the PCR re~ction mixture is repeatedly cycled from high te per2tures t~90~ C) for melting the DNA, to lower temperatures (40~C to 70~C) for primer annealing and lO extension. The first cc ?rcial system for performing the thermal cycling required in the polymerase chain reaction, the Perkin-Elmer Cetus DNA Thermal Cycler, was introduced in 1987.

Applications of PCR technology are now moving from basic 15 research to applications in which large numbers of similar amplifications are routinely run. These areas include diagnostic research, biopharmaceutical development, genetic an~ly~is, and environmental testing. User~ in these areas would benefit from a high performance PCR system that would 20 provide the user with high throughput, rapid turn-around time, and reproducible results. Users in these areas must be assured of reproducibility from sample-to-sample, run-torun, lab-to-lab, and instrument-to-instrument.

For example, the physical mapping process in the Human 25 Genome Project may ~ec~- ? greatly simplified by utilizing sequence tagged sites. An STS i8 a short, unique seguence easily amplified by PCR and which identifies a location on the chromosome. Checking for such sites to make qenome maps reguires amplifying large numbers of ~ample~ in a short time 30 with protocols which can be reproducibly run throughout the world.

As the number of PCR samples increases, it bec. -~ more i~portant to integrate amplification with sample preparation and post-amplification analysis. The sample vessels must 35 not only ~llow rapid thermal cycling but also permit more automated handling for operations such as solvent _ _ ~ a 5~7 .

-- 3 extractions and centrifugation. The vessels ~hould work consistently at low volumes, to reduce reagent costs.

Generally PCR temperature cycling involves at least two incubations at different temperatures. One of these 5 incubations is for primer hybridization and a catalyzed primer extension reaction. The other in~llh~tion is for denaturation, i.e., separation of the double stranded extension products into single strand templates for use in the next hybridization and extension incubation interval.

10 The details of the polymerase chain reaction, the temperature cycling and reaction conditions nece~-Fy for PCR as well as the various reagents and enzymes ne~essary to perform the reaction are described in U.S. patents 4,683,202, 4,683,195, EPO Publication 258,017 and 4,889,818 15 (Taq polymerase enzyme patent) ~nd all other PCR patents which are assigned to Cetus Corporation The purpose of a polymerase chain reaction is to manufacture a large volume of DNA which i8 identical to an 20 initially supplied ~mall volume of ~seed" DNAo The reaction involves copying the strands of the DNA and then using the copies to generate other copies in subsequent cycles. Under ideal conditions, each cycle will double the amount of DNA present thereby resulting in a geometric ~ro~L~6sion in the 25 volume of copies of the #target~ or ~seed" DNA ~trands present in the reaction mixture.

A typical PCR temperature cycle reguires that the reaction mixture be held accurately at each incubation temperature for a prescribed time and that the identical 30 cycle or a similar cycle be repeated many times. A typical PCR program starts at a sample temperature of 94 ~C held for 30 seconds to denature the reaction mixture. Then, the temperature of the reaction mixture is lowered to 37-C and held for one minute to permit primer hybridization. Next, 35 the temperature of the reaction mixture is raised to a temperture in the range from 50~C to 72~C where it is held .. :. .

, s r 2~67~3 for two minutes to promote the synthesis of extension products. This completes one cycle. The next PCR cycle then starts by raising the temperature of the reaction mixture to 94~C again for strand separation of the extension 5 products~ formed in the previous cycle (denaturation).

Typically, the cycle is repeated 25 to 30 times.

Generally, it is desirable to change the ~ample temperature to the next temperature in the cycle ~s rnpidly as possible for several rea~ons. Fir~t, the chemical 10 reaction has an optimum t~ ,erature for each of its ~tages.

Thus, less time spent at nonoptimum temperatures means a better chemical result is achieved. Another reason i8 that a minimum time for holding the reaction mixture at each incubation te~r~rature is reguired after each ~aid 15 incubation temperature is reached. These ~ni incubation times establish the "floor~ or minimum time it takes to complete a cycle. Any time transitioning between sample incubation temperatures is time which is added to this mini~um cycle time. Since the number of cycles is fairly 20 large, this additional time ~ ecessarily lengthens the total time needed to complete the a~mplification.

In ~ome prior automated PCR instruments, the reaction mixture was stored in a disposable plastic tube which is closed with a cap. A typical sample volume for such tubes 25 was approximately 100 microliters. Typically, such instruments used many such tubes filled with ~ample DNA and reaction mixture inserted into holes called ~ample wells in a me~al block. To perform the PCR p~c:ess, the temperature of the metal block was controll~d according to prescribed 30 temperature~ and times ~pQcifi~d by the user in a PCR protocol file. A computer ~nd associated electronics then controlled ~he temrer~ture of the metal block in accordance with the u~er ~uppl~ed data in the PCR protocol file defining the times, temperatures and number of cycles, etc.

35 As the met~l block changed temperature, the ~amples in the various tubes followed with similar changes in temperature.

- 5 - 2 ~ 7~ 3 However, in these prior art instruments not all samples experienced exactly the same temperature cycle. In these prior ~rt PCR instruments, errors in sample temperature were generated by nonuniformity of temperature from place to 5 place within the metal ~ample block, i.e., temperature gradients existed within the metal of the block thereby causing ~ome samples to have different temperatures than other samples at particular times in the cycle. Further, there were delays in transferring heat from the cample block lo to the ~ample, but the delays were not the ~ame for all samples. To perform the PCR process succes6fully and efficiently, and to enable ~o called ~quantitative" PCR, these time delays and temperature errors must be ini ized to a great extent.

The problems of minimizing time delays for heat transfer to and from the sample li~uid and minimizing temperature errors due to temperature gradients or nonuniformity in temperature at various points on the metal block become particularly acute when the ~ize of the region 20 containing samples becomes large. It is a highly desir~ble attribute for a PCR instrument to~have a metal block which is 1 rge enough to acc -date 96 ~ample tubes arranged in the format of an industry stAn~Ard microtiter plate.

The microtiter plate i a widely used means for 25 handling, processing and ~nalyzing large numbers of small samples in the bioc~- istry and biotechnology fields.

Typically, a microtiter plate i~ a tray which is 3 5/8 inches wide and 5 inches long and contains 96 identical ~ample well~ in an 8 well by 12 well rectangular array on 9 30 millimeter centers.

Although microtiter plates are zvailable in a wide variety of materials, ~pes and volumes of the sample wells, which are optimized for many different uses, all microtiter plate~ have the aame overall outside dimensions and the same 8 x 12 array of wells on 9 35 millimeter centers. A wide variety of equipment is available for automating the handling, processing and ~ - 6 - 2~67~3 analyzing of samples in this standard microtiter plate format.

Generally microtiter plates are made of injection molded or vacuum formed plastic and are in~Yr~ive and 5 considered disposable. Disposability is a highly desirable characteristic because of the legal liability arising out of cross cont~min~tion and the difficulty of washing and drying microtiter plates after use.

It is therefore a highly desirable characteristic for 10 a PCR instrument to be able to perform the PCR reaction on up to 96 samples simultaneously ~aid ~amples being arranged in a microtiter plate format.

Of course, the size of the metal block which is necessary to heat and cool 96 ~amples in an 8 x 12 well 15 array on 9 millimeter centers is fairly large. This large area block creates multiple challenging engineering problems for the design of a PCR instrument which is capable of heating and cooling such a block very rapidly in a temperzture range generally from 0 to 100~C with very little 20 tolerance for temperature variations between ~amples. These problems arise from several sour¢es. First, the large thermal mass of the block makes it difficult to move the block temperature up and down in the operating range with great rapidity. Second, the need to attach the block to 25 various external devices ~uch as manifolds for supply and withdrawal of cooling liguid, block support attachment points, and associated other peripheral eguipment creates the potential for t erature gradients to exist across the block which exceed tolerable limits.

There are al~o numerous other conflicts between the requirements in the design of a thermal cycling ~ystem for automated performance of the PCR reaction or other reactions requiring rapid, accurate t~ erature cycling of a large number of samples.

For example, to change the te ~rature 35 of a metal block rapidly, a large amount of heat must be added to, or removed from the sample block in a short period ~ 2 ~ 4 3 - 7 of time. Heat can be added from electrical resistance heaters or by flowing a heated fluid in contact with the block. Heat can be removed rapidly by flowing a chilled fluid in contact with the block. However, it is seemingly 5 impossible to add or remove large amounts of heAt rapidly in a metal block by these means without causing large differences in temperature from place to place in the block thereby forming temperature gradients which can result in nonuniformity of temperature among the samples.

Even after the process of addition or removal of heat i6 termin~ted, t~ ?rature gradients can persist for a time roughly proportional to the square of the distance that the heat stored in variou~ points in the block must travel to cooler regions to eliminate the temperature gradient. Thus, 15 as a metal block i~ made larger to acro~o~Ate more ~amples, the time it takes for temperature gradients exicting in the block to decay after a temperat~- change causes t~mrerature gradients which extend across the largest dimensions of the block can become markedly longer. This makes it 20 increasingly difficult to cycle the temperature of the ~ample block rapidly while maintain~ing accurate t~ erature uniformity among all the ~amples.

Because of the time required for t. erature gradients to dissipate, an important need has arisen in the design of 25 a high performance PCR instrument to prevent the creation of temperature gradients that extend over large distAnces in the block. Another need is to avoid, as much as possible, the requirement for heat to travel acrocs -cbAnical boundaries between metal parts or other peripheral equipment 30 attached to the block. It i5 difficult to join metal parts in a way that insures uniformly high thermal conductance everywhere across the joint.

Non~n~formities of thermal conductance will generate unwanted temperature gradients.

Sum~ary of the Invention According to the teachings of the invention, there is - 8 - ~ 7 ~ 3 disclosed herein a thin walled sample tube for decreasing the delay between changes in sample temperature of the sample block and corresponding changes in tP ?rature of the reaction mixture. Two different sample tube ~izes are 5 disclosed, but each has a thin walled conical section that fits into a matching conical recess in the sample block.

Typically, cones with 17~ angles relative to the longitudinal axis are used to prevent jamming of the tubes into the sample block but to allow snug fit. Other shapes 10 and angles would also suffice for purposes of practicing the invention.

Also, other types of heat exchangers can al~o be used other th~n sample blocks such ~8 li~uid baths, ovens, etc.

However, the wall thickness of the section of the sample 15 tube which is in contact with whatever heat exchange is being used should be as tr.ln as possible 80 long as it is sufficiently strong to withstand the thermal stresses of PCR cycling and the stresses of normal use. Typically, the sample tubes are made of autoclavable poly~ lene such as 20 Himont PD701 with a wall thi~kn~ss of the conical section in the range from 0.009 to 0.012 ~ches plus or ~inus 0.001 inches. Most preferably, the wall thi~k~ess is 0.012 inches.

In the preferred embodiment, the sample tube also has 25 a thicker walled cylindrical section which joins with the conical section. This conical section provide containment for the original reaction mixture or reagents which may be added after PCR processing.

- The sample tube shown in Figure 50 has industry 30 standard configuration ~ e~L for the thin walls for compatibility in other PCR sy~tem~. The ~ample tube of Figure 15 is a shorter tube which can be used with the system disclosed herein.

The 35 other subject matter of the system environment in which use of the thin walled sample tubes is preferred ~re summarized * Trade-mark ,. . .

2~7~3 g below.

There is also described herein a novel method and apparatus for achieving very accurate temperature control for a very large number of samples arranged in the S microtiter plate format during the performance o* very rapid temperature cycling PCR protocols. The teachings of the invention contemplate a novel structure for a sample block, sample tubes and supporting mounting, heating and cooling apparatus, control ¢lectronics and software, a novel user lo interface and ~ novel method of using said apparatus to perform the PCR protocol.

The instrument described herein is designed to do ~CR gene amplification on up to 56 samples with very tight tolerances of temperature control across the universe of 15 samples. This means that all samples go up and down in temperature simultaneously with very little difference in temperature between different wells cont~ining different samples, this being true throughout the polymerase chain reaction cycle. The instrument described herein i5 also 20 capable of very tight control of the reaction mixture concentration through control of the evaporation and condensation process~- in each sample well. Further, the instrument described herein is capable of proce~sing up to 96 samples of 100 microliters each from different donor 25 sources with substantially no cross-contA in~tion between sample wells.

The teachings of the invention herein includes a novel method of heating and cooling an aluminum ~ample block to thermally cycle ~amples in the ~tAnd~rd 96-well microtiter 30 plate format with thc result that excellent sample-to-sample uniformity exists despite rapid thermal cycling rates, noncontrolled varying ambient temperatures and variations in other operating conditions such as power line voltage and coolant temperatures.

3~ The teachings of the invention also contemplate a novel design for a disposable plastic 96~well microtiter plate for 2~7~3 -- 10 -accommodation of up to 96 individual sample tubes containing DNA for thermal cycling each sample tube having individual freedom of movement sufficient to find the best fit with the sample block under downward pressure from a heated cover.

5 The microtiter plate design, by allowing each tube to find the best fit, provides high and uniform thermal conductance from the sample block to each ~ample tube even if differing rates of thermal eYr~n~ion and contraction between the metal of the block and the plastic of the ~ample tube and lo microtiter plate structure cause the relative center-to center dimensions of the wells in the ~ample block to change relative to the center-to-center distance of the sample tubes in the disposable microtiter plate structure.

The teachings of the invention also contemplate a novel 15 method and apparatus for controlling the PCR instrument which includes the ability to continuously calculate and display the temperature of the samples being processed without directly measuring these t- -ratures. These calculated temperatures are used to control the time that 20 the samples are held within the given temperature tolerance band for each target temperature of incubation. The control system also controls a three-zone heater thermally coupled to the ~ample block and gates fluid flow through directionally interlaced ramp cooling channels in the sample 25 block which, when combined with a constant bias ~ooling flow of coolant through the sample block provides a facility to achi~ve rapid temperature changes to and precise temperature control at target temperatures specified by the user. The method and apparatus for controlling the three~zone heater 30 includes an apparatus for taking into accou~ mong other things, the line voltage, block temperature, coolant temperature and ambient temperature in calculating the amount of electrical energy to be ~upplied to the various zones of the three-zone heater. This heater has zones which 35 are separately controllzble under the edges or "guard bands" of the sample block so that excess heat lofises to the ~ 2~5~7~3 -- 11 ambient through peripheral equipment attached to the edges of the sample block can be compensated. This helps prevent ther~al gradients from forming.

The teachings of the invention also contemplate ~ novel 5 method and apparatus for preventing loss of solvent from the reaction mixtures when the samples are being incubated at temperatures near their boiling point. A heated platen covers the tops of the sample tubes and is in contact with an individual cap which provides a gas-tight ~eal for each 10 sample tube. The heat from the platen heats the upper parts of each sample tube and the cap to a temperature above the condensation point such that no condensation and refluxing occurs within Any sample tube. Condensation represents a relatively large heat transfer since an amount of heat equal 15 to the heat of vaporization is given up when water vapor condenses. This could cause large temperature variations from ~ample to sample if the condensation does not occur uniformly. The heated platen prevents any condensation from occurring in any sample tube thereby minimizing this source 20 of potential temperature errors. The use of the heated platen also reduces reagent consumption.

Furthermore, the heated platen provides a downward force for each sample tube which exceeds an experimentally determined inj downward force necessAry to keep all 25 sample tubes pressed firmly into the temperature controlled sample block 50 as to establish and maintain uniform blockto-tube thermal conductance for ~ach tube. This uniformity of thermal conductance i~ established regardless of variations ~rom tube to tube in length, diameter, angle or 30 other ~i ~n~ional errors which otherwi~e could cause some ~ample tubes to fit more ~nugly in their corresponding sample wells than other ~ample tubes.

The heated platen softens the plastic of each cap but does not totally destroy the caps elasticity. Thus, a 35 minimum threshold downward forced is successful].y applied to each tube despite differences in tube height from tube to 2Q5~7~3 - 12 tube.

The PCR instrument described herein reduces cycle times by a factor of 2 or more and lowers reagent cost by accommodating PCR volumes down to 20 uh but remains 5 compatible with the industry standard 0.5 ml microcentrifuge tube.

Brief Descri~tion of the Drawings Figure 1 is a block diagram of the thermal cycler according to the teachings of the invention.

Figure 2 is a plan view of a sample block according to the teachings of the invention.

Figure 3 is a side, elevation view of the ~ample block showing the bias and ramp cooling ~hAn~els.

Figures 4 and 5 are end, elevation view6 of the ~ample 15 block.

Figure 6 is a ~ectional view of the sample block taken alony section line 6-6' in Figure 2.

Figure 7 is a sectional view of the sample block taken along section line 7-7' in Figure 2.

Figure 8 is a section~l view of the sample block taken along section line 8-8' in Figure 2.

Figure 9 is a cross-section~l, elevation view of the sample bloclc structure after assembly with the three-zone film heater and block ~upport.

2S Figure 10 is a graph of power line voltage illustrAting the ~orm of power control to the three-zone film heater.

Figure 11 is a temperature graph showing a typical thre~ incub tion temperature PCR protocol.

Figure 12 i5 a cross-sectional view of the sample block 30 illustrating the local zone concept.

Figure 13 is a plan view of the three-zone heater.

Figure 14 is a graph of sample t~ ,?rature versus time illustrating the effect of an r of a sample tube seating force F which is too low.

Figure 15 is a cross-sectional view of a sample tube and cap seated in the sample block.

2~7~3 - 13 Figure 16A is a graph of the impulse response of an RC circuit.

Figure 16B is a graph of an impulse excitation pulse.

Figure 16C is a graph illustrating how the convolution 5 of the thermal impulse response and the temperature history of the block give the calculated sample temperature.

Figure 16D illustrates the electrical analog of the thermal response of the sample block/sample tube system.

Figure 17 illustrates how the calculated t~mperatures 10 of 5iX different samples all converge on a target temperature to within about 0.5~C of each other when the constants of proportionality for the equations used to control the three zone heater are properly set.

Figure 18 is a graph illustrating how the denaturation 15 t~rget te ?rature affects the amount of DNA generated.

Figure 19 is a cro~s-sectional view of the sliding cover and heated platen.

Figure 20 is perspective view of the sliding cover, sample block and the knob used to lower the heated platen.

Figure 2lA is a cross-sectional view of the assembly of one e~bodiment of the frame, retainer, nample tube and cap when seated on a sample block.

Figure 21B is a cross-~ectional view of the assembly of the pr.eferred emho~t ~nt of the frame, retainer, sample tube 25 and cap when seated on the nample block.

Figure 22 is a top, plan view of the plastic, disposable frame for the microtiter plate.

Figure 23 is a bottom, plan view of the frame.

Figure 24 is an end, elevation view of the frame.

Figure 25 i~ another end, elevation view of the frame.

Figure 26 is a cross-sectional view of the frame taken along section line 26-26' in Figure 22.

Figure 27 is a cross-sectional view of the frame taken along section line 27-27' in Figure 22.

Figure 28 is an edge elevation view and partial section of the frame.

~ - 14 - 205~7~3 Figure 29 is a sectional view of the preferred sample tube.

Figure 30 is a sectional view of the upper part of the sample tube.

Figure 31 is an elevation view of a portion of the cap strip.

Fiqure 32 is a top view of a portion of the cap strip.

Figure 33 is a top, plan view of the plastic, disposable retainer portion of the 96 well microtiter tray.

lo Figure 34 is a side, elevation view with a p~rtial section of the retainer.

Figure 35 is an end, elevation view of the ret~iner.

Figure 36 is a sectional view of the retainer taken along section line 36-36' in Figure 33.

Figure 37 is a ~ectional view of the retainer taken Along ~ection line 37-37' in Figure 33.

Figure 38 is a plan view of the plastic disposable ~upport base of the 96 well microtiter tray.

Figure 39 is a bottom plan view of the ba~e.

Figure 40 i5 a side elevation view of the base.

Figure 41 is an end elevation view of the base.

Figure 42 i8 a ~ectional view of the support base taken along section line 42-42' in Figure 38.

Figure 43 is a ~ectional view of the s~o~L base taken 25 along section line 43-43' in Fiqure 38.

Fiqure 44 is a section view of the base taken along section line 44-44' in Figure 38.

Figure 45 is a perspective ~xploded view of the plastic disposable items that comprise the microtiter tray with some 30 ~ample tube~ and caps in place.

Figure 46 i~ a diagram of the coolant control ~ystem 24 in Figure 1.

Figures 47A and 47B are a block diagram of the control electronics accordinq to the teachings of the invention.

Figure 48 is a schematic of a typical zener temperature sensor.

CA 020~6743 1999-02-04 ~_ - 15 Figure 49 is a time line diagram of a typical sample period.

Figure 50 is elevation sectional view of a tall thin walled sample tube marketed under the trademark MAXIAMP.

Figure 51 is a graph showing the difference in response time between the thin walled sample tubes and the thick walled prior art tubes.

Figure 52 is a plan view of a sample tube and cap.

Figures 53 and 54 are flow charts of the power up test sequence.

Detailed Description of the Invention Referring to Figure 1 there is shown a block diagram of the major system components of one embodiment of a computer directed instrument for performing PCR according to the teachings of the invention. Sample mixtures including the DNA or RNA to be amplified are placed in the temperature-programmed sample block 12 and are covered by heated cover 14.

A user supplies data defining time and temperature parameters of the desired PCR protocol via a terminal 16 including a keyboard and display. The keyboard and display are coupled via bus 18 to a control computer 20 (hereafter sometimes referred to as a central processing unit or CPU). This central processing unit 20 includes memory which stores the control program described below, the data defining the desired PCR protocol and certain calibration constants described below. The control program causes the CPU 20 to control temperature cycling of the sample block 12 and implements a user interface which provides certain displays to the user and which receives data entered by the user via the keyboard of the terminal 16.

In the preferred embodiment, the central processing unit 20 is custom designed. A block ~ 2~7~3 - 16 diagra~ of the electronics will be discussed in more detail below.

In alternati~e embodiments, the central processing unit 20 and associated peripheral electronics to control the various heaters and other electro-mechanical systems of the 5 instrument and read various ~ensors could be any general purpose computer ~uch as a suitably programmed personal computer or microcomputer.

The samples 10 are stored in capped ~i~pos~hle tubes which are seated in the sample block 12 and are thermally lo isolated from the ambient ~ir by a heated cover 14 which contacts a plastic disposzble tray to be described below to form a heated, enclosed box in which the ~ample tubes reside. The heated cover serves, among other things, to reduce undesired heat transfer~ to and from the ~ample 15 mixture by evaporation, con~en~tion and refluxing inside the Qample tubes. It also reduces the rhAnce of cross cont~ination by keeping the insides of the caps dry thereby preventing aerosol formation when the tubes are uncapped.

The heated cover is in contact with the sample tube caps and 20 keeps them heated to a temperature of approximately 104~C or above the condensation points of the various components of the reaction mixture.

The central processing unit 20 includes appropriate electronics to ~ense the temperature of the heated cover 14 25 and con~rol electric resistance heaters therein to maintain the cover 14 at a predetermined t~ ,?rature.

Sensing of the t~ ~-rature of the heated cover 14 and control of the resistance heaters therein $s aceomplished via a temperature sensor ~not shown) and bus 22.

A coolant control system 24 continuou61y circulates a chilled liguid coolant such a~ a mixture of automobile antifreeze and water through bias cooling rhAn~els (not shown) in the sample block 12 via input tubes 26 and output tube 28.

The coolant control ~ystem 24 also controls fluid 35 flow through higher volume ramp cooling fluid flow paths tnot shown) in the sample block 12. The ramp cooling 2~67~3 -- 17 -channels are used to rapidly change the temperature of the sample block 12 by pumping large volumes of chilled liquid coolant through the block at a relatively high flow r_te.

Ramp cooling liguid coolant enters the sample block 12 5 through tube 30 and exits the ~ample block through tube 32.

The details of the coolant control system are shown in Figure 46.

The coolant control system will be ~i~cl~csed more fully below in the description of the electronic~ and software of the control ~ystem.

o Typically, the liquid coolant used to chill the sample block 12 consists mainly of a mixture of water and ethylene glycol.

The liquid coolant is chilled by a heat exchanger 34 which receives liguid coolant which has extracted heat from the sample block 12 vi~ input tube 36.

The heat 15 exchanger 34 receives compressed liquid freon refrigerant via input tube 38 ~rom a refrigeration unit 40. This refrigeration unit 40 includes a compressor (not ~hown), a fan 42 and a fin tube heat radiator 44. The refrigeration unit 40 compresses freon gas received from the heat 20 exchanger 34 via tube 46. The gaseous freon is cooled and condensed to a liguid in the fin tube condenser 44. The pressure of the liquid freon i6 maintained above its vapor pressure in the fin tube condenser by A flow restrictor capillary tube 47. The output of this capillary tube is 25 coupled to the input of the heat ~Y~hAnger 34 via tube 38.

In the heat exchanger, the pressure of the freon i5 allowed to ~rop below the freon vapor pressure, and the freon eYr~nd~. In this pbc~e~ of -yr~n~ion~ heat is absorbed from the warmed liquid coolant circulating in the heat 30 exchanger and thi~ heat i~ tr~nsferrQd to the freon thereby causing the freon to boil. The warmed freon is then extracted from the heat ~YchAnger via tube 46 and is compressed and again circulated through the fin tube condensor 44. The fan 42 blows air through the fin tube 35 condensor 44 to cause heat in the freon from tube 46 to be exchanged with the ambient air. As symbolized by arrows 48.

2~56~3 - 18 The refrigeration unit 40 ~bould be capable of ~xtracting 400 watts of heat at 300C and 100 watts of heat at 10~C from the ].iquid coolant to ~upport the rapid temperature cycling according to the teachings of the invention.

In the preferred embodiment, the apparatus of Figure 1 is enclosed within A housing (not shown).

The heat 48 expelled to the _mbient air is kept within the housing to aid in evaporation of any condensation which oc~ D on the ~arious tubes carrying chilled liquid coolant or freon from lo one place to another. This conden6ation can cau~e corrosion of ~etals used in the conotruction of the unit or the electronic circuitry and should be removed.

Expelling the heat 48 inside the enclosure helps evaporate any condensation to prevent corrooion.

After exchanging its heat with the freon, the liquid c~olant exits the heat eY~hA~er 34 via tube 50 and reenters the cool_nt control system where it i8 gated as needed to the sample block during rapid cooling portion of the PCR cycle defined by data cntered by the user via te~ inAl 16.

As noted above, the PCR protocol involves incubations at at least two different tempèratures and often three different temperAtureo. A typic_l PCR cycle is shown in Figure 11 with _ denaturation incubation 170 done at ~ temperature near 94~C, a hybridization 1nrllhAtion 122 done 25 at A t~ ,-rature near room t~ ,-rature ~25-C-37-C) and an exten~ion inc--hAtion 174 done at a t- ,-rature near 50~C.

These temperatureo are cubotantially diff~rent, _nd, therefore means must be providsd to move the temperature of the reaction mixture of all the ~ample~ rapidly from one 30 temperature to another. The ramp cooling ~y6tem io the mean~ by which the temperature of the 6ample block 12 is brought down rapidly from the high temperature denaturation incubation to the lower temperature hybridization and extension incubation t~ ,cratures. Typically the coolant 35 temperature is in the range fro~ 10-20~C.

When the coolant is at 20~C it can pump out about 400 watts of heat from the 2~7~3 -- 19 -sample block Typically the ramp cooling channel dimensions, coolant temperature and coolant flow rate are set such that peak cooling of 5~-6~C per second can be achieved near the high end of the operating range (100~C) 5 and an average cooling rate of 2 5~C per 6econd is achieved in bringing the sample block temperature down from 940C to 37OC The ramp cooling system, in some embodiment6, may also be used to ~aintain the sample block tempQratur- ~t or near 10 the target incubation t _-rature al~o ~ow~v~, in the preferred embodiment, small temperature change~ of the sample bloc~ 12 in the downward direction to maintain target incubation temperature are implemented by the bias cooling ~ystem 1~ As seen in Figure 46, a pump 41 constantly pumps coolant from a filter/reservoir 39 (130 milliliter capacity) via 1/2" pipe and pu~ps it via a 1/2~ pipe to a branching inter6ection 47 The pump 41 aupplies coolant to pipe 45 at a constant flow rat- of 1-1 3 gallons per minute At the 20 intersQction 47, a portion of the flow in tube 45 ic diverted as the constant flow tprough the ~ias cooling channels 49 Another portion of the flow in tube 45 i6 diverted through a flow restrictor 51 to output tube 38 Flow restrictor 51 maintains sufficient pres~ure in the 25 ~y~tem ~uch that a positive pres~ure exi~t~ at the input 53 of a two ~tat- ~olenoid operated valve 55 under the control of the CPU 20 via bu~ S4 When ramp cooling i~ desired to impl~ment a rapid downward temperature change, the CPU 20 causes the ~olenoid operated valve 55 to open to allow flow 30 of coolant ~hrough the ramp cooling ~hAnn~ls 57 There are 8 ramp cooling channels so the flow rate through each ramp cooling ch~nnel is about 1/8 gallon per minute The flow rate through the bia~ cooling ~h~nel~ i8 much le88 because of the grea~ly restricted cross- ectional area thereof The bias cooling ~ystem provides a small con~tant flow of chilled coolant through bias cooling channels ~9 in the ~ 2~7~3 - 20 sample block 12. This cau~es ~ constant, small he~t loss from the sample block 12 which i~ compen~ated by a multizone heater 156 which is thermally coupled to the sample block 12 for i nct~hAtion Fe', ~~ts where the temperature of S the sample block is to m_intained ~t a steady value. The constant small heat 105s caused by the bias cooling flow allows the control system to implement proportional control both upward and downward in t _-rature for small temperatures. Thi~ ~eans both heating ~n~ ~ooling at 10 controlled, predictable, small rates i8 avail~ble to the temperature servo ~ystem to correct for block temperature errors to cause the block temperature to faithfully track a PCR temperature profile entered by the user. The alternative would be to cut off power to the fil~ heater and lS allow the ~ample block to cool by giving up ~eat to the ambient by radiation and con~__Lion when the block temperature got too high.

This would be too slow ~nd too unpredictable to meet tight temperature control specifications for guantitative PCR cycling.

This multi-zone heater 156 is controlled by the CPU 20 via bus 52 in Figure 1 ~nd i~ the means by which the temp~!rature of the ~ample block 12 i8 raised rapidly to higher incubation temperatures from lower incubation temperature~ and i8 the means by which bia~ cooling is 25 c ~n~-ted and t~ ,-rature ~rrors are corrected in the upward direction during t: ,-rature tracking and control duril~g ~ ~c~hAtions .

In alternative - ho~iment~, bias cooling may be eliminated or may be ~upplied by other mQans such ~s by the 30 u~e of a cooling fan and cooling fin~ formed in the metal of the sample block, peltier ~unctions or constantly circulating tap water. Care must be taken however in these alternative embo~i ~nt~ to inffure that t~ ,-rature gradients ~re not created in the ~ample block which would cause the 35 temperature of some samples to diverge from the temperature of other 6amples thereby possibly causing different PCR ~ 2~67~3 - 21 Amplification results in some sample tubes than in others In the preferred embo~i ?nt~ the bias cooling is proportional to the difference between the block t~ -rature and the coolant temperature The CPU 20 controls the temperature of the sample block 12 by sensing the temperature of the metal of the sample block via temperature sensor 21 and bus 52 in Figure 1 and by sensing the temperature of the circulating coolant liguid via bus s4 and a temperature ~en~or in the coolant cGn~ol 10 system The temperature sensor for the coolant i~ shown at 61 in Figure 46 The CPU also ~enses the internal ambient air temperature within the hou~ing of the system via an ambient air temperature sensor 56 in Figure 1 Further, the CPU 20 senses the line voltage for the input power on line 15 58 via a sensor symbolized at 63 All the~e items of d_ta together with items of data entered ~y the user to define the desired PcR protocol ~uch as target temperatures and times for incubations are used by a control program to be described in more detail below This control program 20 calculates the amount of power to apply to the various zones of the multi-zone sample block film heater 156 via the bus 52 and generates a coolant cG ~Lol ~ignal to open or close the ~olenoid operated valve 55 in the coolant control system 24 via bus 54 BO a~ to cause the temperature of the sample 25 block to follow the PCR protocol defined by data entered by the user Referring to Figure 2, there i- shown a top view of the sample block 12 The purpose of the ~ample block 12 is to provide a chAnical ~ppG~ and heat ~YrhA~e element for 30 an array of thin walled ~ample tubes where heat may be exchanged between the sample liquid in ~ach sa~ple tube and liguid coolant flowing in the bias cooling and ramp cooling channels formed in the sample block 12 Further, it is the function of the sample bloc~ 12 to provide this heat 35 exch2nge ~unction without creating large temperature gradients between various ones of the sample wells ~uch that 2~g7~ - 22 all sample mixtures in the array experience the same PCR cycle even though they are spatially separated. It is an overall objective of the PCR instrument described herein to provide very tight t~ erature control over the temperature of the sample liquid for a plurality of samples such that the temperature of any sample liquid does not vary appreciably (approximately plus or minus 0.5~C) from the temperature of any other sample liquid in Another well at any point in the PCR cycle.

There is an emerging branch of PCR t~hnology called "quantitative" PCR. In this technology, the objective is to perform PCR amplification as precisely as possible by causing the amount of target DNA to exactly double on every cycle. Exact doubling on every cycle is difficult or 15 impossible to achieve but tight te ,-rature control helps.

There are many sources of errors which can cause a failure of a PCR cycle to exactly double the amount of target DNA (hereafter DNA should be understood as also referring to RNA) during a cycle. For example, in some PCR 20 ampllfications, the process starts with a single cell of target DNA. An error that can easily occur results when this single cell sticks to the wall of the sample tube and does not amplify in the first several cycles.

Another type of error is the entry of a foreign 25 nuclease into the reaction mixture which attacks the ~foreign~ target DNA. All cells have some nonspecific nuclease th~t attacks foreign DNA that is 1008e in the cell.

When this happens, it interferes with or ~tops the replication pl GCeSS ~ Thus, if A drop of ~aliva or a 30 dandruff particle or material from another sample mixture were inadvertently to enter a sample mixture, the nuclease materials in these cells could attack the target DNA and cause an error in the amplification process. It is highly desirable to eliminate all such sources of cross35 contamination.

Another source of error is nonprecise control over - 23 - 2~7~3 sample mixture temperature as between various ones of 2 multiplicity of di~ferent ~amples. For example, if ~ll the samples are not precisely controlled to have the proper annealing t~mrerature (a user selected temperature usually 5 in the range from 50 to 60~C) for the extension incubation certain forms of DNA will not extend properly. This happens bacause the primers used in the extension proces~ anneal to the wrong DNA if the temperature is too low. If the annealing te~perature is too high, the primers will not 10 anneal to the target DNA at all.

One can easily imagine the conse~uences of performing the PCR amplification process inaccurately when PCR amplification is part of diagnostic testing ~uch as for the presence HIV antibodies, hepatiti~, or the presenre of 15 genetic diseases such as sickle cell anemia, etc. A false positive or false negative rssult in such diagnostic testing can have disastrous personal and legal conseguences.

Accordingly, it is an object for the design of the PCR instrument described herein to eliminate as many of these 20 ~ources of possible errors as possible such as cross cont~ ~n~tion or poor t~ ~rature control while providing an instrument which is compatible with the industry st~ndArd 96-well microtiter plate format. The instrument must rapidly perform PCR in a flexible manner with a simple user 25 interface.

In the preferred ~ ho~1 -nt, the sample block 12 is machi~ out of a solid block of relatively pure but corrosion resistant aluminum ~uch ~s the 6061 aluminum alloy. Machining the block ctructure out of a solid block 30 of aluminum results in a ~ore thermally homogenous structure. Cast aluminum stru~u-~s tend not to be as thermally homogenous as is neceFE~ry to meet the very tight desired t~np~rature control specifications.

Sample block 12 is capable of rapid changes in 35 temperature because the thermal mass of the block is kept low. This is done by the formation in the block of many . ~ ~ 2 ~ 3 - 24 cooling passagew~ys, sample wells, grooves and other threaded and unthreaded holes.

Some of these holes are used to attach the block to supports and to attach external devices such as manifolds and spillage trays thereto.

To best appreciate the "honeycomb" nature of the sample block structure, the reader should refer simultaneously to Figure 2 which shows the block in plan view as well as Figures 3 through 8 which show elevation views and strategically located sectional views of the ~aopl8 block.

10 For example~ ~igure 3 is a side elevation view ~howing the cooling channel positions taken from the vantage point of the view line 3-3' in Figure 2. The elevation view of the sample block 12, looking at the opposite edge, i8 identical.

Figure 4 is an elevation view of the edge of the ~ample 15 block 12 from the perspective of view line 4-4' in Figure 2.

Figure 5 is an elevation view of the end of the sample block 12 taken from the perspective of view line 5-5' in Figure 2.

Figure 6 i8 a ~ectional view of ~e ~ample block 12 taken along the cection line 6-6' in Figure 2. Figure 7 is a 20 ~ectional view of the sample block 12 taken along ~ection line 7-7' in Figure 2. Figure 8 i~ a sectional view of the sAmple block 12 taken along section line 8-8' in Figure 2.

The top surface of the sample block 12 is drilled with an 8 x 12 arrAy of conical sample wells of which wells 66 25 and 68 are typical. The conical configuration of each ~ample well is best ~een if Figure 8. The walls of each ~ample well are drilled at an angle of 17- to match the angle of the conical ~ection of each ~ample tube. This is done by drilling a pilot hole having the di~meter D~ in 30 ~igure 8. Then a 17~ countersink is used to form the conical walls 67.

The bottom of each ~ample well includes a ~ump 70 which has a depth which -Ycee~ the depth of penetration of the tip of the ~mple tube. The ~ump 70 i~ created by the pilot 35 hole and provides a small open space bene~th the sample tube when the sample tube is seated in the corresponding sample ~ 2~6~ - 25 well. This ~ump provides a space for liquid ~uch as condensation that forms on the well walls to reside without interfering with the tight fit of each sample tube to the walls of the sample well. ~his tight fit is necess~ry to 5 insure that the thermal conductance from the well wall to the sample liquid is uniform and high for each ~ample tube.

Any cont2~;nation in a well which causes a loose fit for one tube will destroy this uniformity of thermal conductance across the array. That is, because liguid is ~ubstantially lO uncompressible at the pressures involved in ~eating the sample tubes in the sample wells, if there were no ~ump 70, the presence of liquid in the bottom of the sample well could prevent a ~ample tube from fully seating in its sample well. Furthermore, the sump 70 provides a space in which a 15 gaseous phase of any liguid re~iding in the sump 70 can expand during high temperature incubations ~uch that large forces of such expansion which would be present if there were no ~ump 70 are not applied to the sample tube to push the 1:ube out of flush contact with the sample well.

It has been found experimentally that it i8 important for each sample tube to be in flush contact with its corresponding sample well and that a certain minimum threshold force be applied to each sample tube to keep the thermal conductivity between the walls of the ~ample well 25 and the reaction mixture uniform throughout the array. This i~i u~ threshold ~eating force is ~hown as the force vector F in Figure lS and i~ a key factor in preventing the thermal - conductivity through the walls of one sample tube from being different than the thermal conductivity through the walls of 30 another sample tube located elsewhere in the block. The minimum threshold seating force F is 30 grams and the preferred force level ~s between 50 and lO0 grams.

The array of 5ample wells is substantially completely surrounded by a groove 78, best seen in Figures 2, 6 and 8, 35 which has two functionS. The main function i~ to reduce the thermal conductivity from the central area of the sample ~ 2~6~ - 26 block to the edge of the block. The groove 78 extends about 2t3 through ~he thickness of the ~ample block. This groove minimizes the effects of unavoidable thermal gradients caused by the necess~ry ~eçhAnical connections to the block 5 of the support pins, manifolds, etc. A secondary function is to remove thermal mass from the sample block 12 so as to allow the temperature of the eample block 12 to be altered more rapidly and to ~imulate a row of wells in the edge region called the "guard band". The amount of m~tal removQd lo by the portion of the groove 78 between points 80 and 82 in Figure 2 is designed to be substantially egual to the amount of metal removed by the adjacent column of eight sample wells 83 through 90. The purpose of this is to match the thermal mass of the guard band to the thermal mass of the 15 adjacent "local zone", a term which will be explained more fully below.

Referring specifically to Figures 3, 6 and 8, there is sho~ the number and relative positions of the var~ous bias cooling and ramp cooling channels which are formed in the 20 metal of the sample block 12. ~here are nine bias cooling ch~nnels marked with reference numerals 91 through 99.

Likewise, there are eight ramp cooling channel~ marked with reference numerals 100 through 107.

Each of these bias cooling and ramp cooling channels is 25 gun drilled through the aluminum of the sample block. The gun drilling process i~ well known and provides the ability ~o drill a long, very ~traight hole which is as close as possible to the bottom surface llO of the sample block 12.

Since the gun drilling process drills a straight hole, this 30 process is preferred ~o as to prevent any of the bias cooling or ramp cooling ch~nn~ls ~rom strayinq during the drilling process and penetr~ting the bottom surface 110 of the ~ample block or otherwi~e altering it~ position relative to the other cooling channels. Such mispositioning could 35 cause undesirable t~ ~rature gradients by upsetting the "loc:al balance" and n local symmetry" of the local zones.

~ 2 ~ 3 - 27 These concepts are explained below, but for now the reader should understand that these notions and the structures which implement them are key to achieving rapid temperature cycling of up to 96 samples without creating excessive 5 temperature errors as between different ~ample wells.

The bias cooling channels 91 through 99 are lined with silicone rubber in the preferred embodiment to reduce the thermal conductivity across the wall of the bias cooling channel. Lowering of the thermal conductivity ~cross the 10 chamlel wall in the bias cooling chAnnels is preferred so as to prevent too rapid of a change in tempera~ure of the sample block 12 when the multi-zone heater 156 is turned off and heat loss from the sample bloc~ 12 is primarily through the bias cooling channels. This is the situation during the 15 control process carried out when the ~ample block temperature has strayed slightly above the desired target incubation temperature and the control system ic attempting to bring the sample block temperature back down to the user's specified incubation temperature. Too fast a cooling 20 rate in this situation could cause overshoot of the desired incubation temperature before the control sy~tem' 8 servo feedback loop can respond although a "controlled overshoot" algorithm is used as will be described below. Since the block tD ~erature servo feedbAck loop has a time constant 25 for reacting to sti~uli, it i8 desirable to control the amount of heating and cooling and the resulting rate of temperature change of the sample block such that overshoot is i ni ized by not changing the sample block temperature at a rate faster than the control ~ystem can respond to 30 temperature errors.

In the preferred emho~i -nt, the bias cooling channels are 4 millimeters in diameter, and the silicone rubber tube has a one millimeter inside diameter and a 1.5 millimeter wall thickness. This provides a bias cooling rate of 35 approximately 0.2~C per second when the block i~ at the high end of the operating range, i.e., near 100~C, and a bias 29~67~ - 28 cooling rate of approximately 0.1~C per second when the sample block 12 is at a temperature in the lower end of the operating range. The coolant control system 24 in Figure 1 causes a flow rate for coolant in the bias cooling cha~nels 5 of approximately 1/20th to 1/30th of the flow rate for liquid coolant through the rA~p cooling channels, 100 through 107. The bias cooling and rAmp cooling channels are the same size, i.e., 4 millimeters in di~meter, ~nd extend completely through the sample block 12.

The bias cooling channels are lined by inserting a stiff wire with a ~ook at the end thereof through the bi~s cooling channel and hooking it through a hole in the end of a silicone rubber tube which has an outside diameter which i5 slightly greater than 4 millimeters. ~he hook in the 15 wire is then placed through the hole in the silicone rubber tube, and the silicone tube i~ pulled through the bias cooling channel and cut off flush with the end surfaces of the sample block 12.

Threaded holes 108 through 114 are used to bolt a 20 coolant manifold to each side of the sample block 12.

There is a coolant manifold bolted to each end of the block.

These two coolant manifolds are coupled to the coolant chAm~els 26, 28, 30 and 32 in Figure 1, and are affixed to the sample block 12 with a gasket material (not shown) 25 interposed between the manifold ~nd the sample block metal.

This ga~ket prevents leaks of coolant and limits the thermal conductivity between the sample block 12 and the manifold which represents a heat sink.

Any ga~ket material which serves the above ~tated pu~oses will suffice for practicing ~he invention.

The positions of the bias cooling and ramp cooling channels relative to the position of the groove ~8 are best 35 seen in the sectional view of Figure 6. The positions of the bias cooling and ramp cooling channels relative to the 7 ~ 3 - 29 positions of the sample wells is best seen in Figure 8. The bias cooling and ramp cooling channels are generzlly interposed between the positions of the tips of the sample wells.

Further, Figure 8 reve~ls that the bias cooling And 5 ramp cooling channels such as chAnnels 106 and 97 c_nnot be moved in the positive z direction very far without risking penetration of the walls of one or more ~ample wells.

Likewise, the cooling c~Annels cannot be moved in the negative z direction very far without creating the 10 possibility of penetrating the bottom surface 116 of the sample bloc~ 12. For clarity, the positions of the bias and ramp cooling c~nnel8 are not shown in hidden lines in Figure 2 relative to the positions of the sample wells and other structures. However, there is either a bias cooling 15 channel or a ramp cooling channel between every column of sample wells.

Referring to Figure 2, the holes 118, 119, 120 and 121 are threaded and are used to attach the sample block 12 to machinery used to machine the various holcs and grooves 20 fonned therein. In Figures 2, 4 and 5, the holes 124, 125, 126 and 127 are used to attach the sample block 12 to a support bracket shown in Figure 9 to be described in more detail b4low. Steel bolts extend through this support bracket into the threaded holes 124 through 127 to provide ~c-~Anical ~po~ of the ~ample block 12. These steel bolts Also represent heat sink~ or heat sourcea which tend to add thermal mass to the sample block 12 and provide additional pathways for tr~nsfer of the2-mal energy between the ~ample block 12 ~nd the surrounding envi. 6 ent. These 30 support pins and the manifolds are two important factors in creating the need for the guard bands to prevent the thermal energy transferred back and forth to these peripheral structures from affecting these sample temperatures.

Referring to Figure 5, the holes 128, 130 and 132 are 35 mounting holes for an integrated circuit temperature sensor (not shown) which is inserted into the sample block through ~ 2~7~3 - 30 hole 128 znd secured thereto by bolts which fasten to threaded holes 130 and 132. The extent of penetration of the hole 128 and the relative position of the temperature sensor to the groove 78 and the adj~cent column of ~ample wells is best seen in Figure 2.

Referring to Figure 2, holes 134 through 143 are mountinq holes which are used to mount a spill collar 147 (not shown). This spill collar 147 is shown in Figure 19 detailing the structure of the heated platen 14, sliding 10 cover 316 and lead screw assembly 312. The ~ul~ose of the spill collar is to prevent any liquid spilled from the sample tubes from getting inside the instrument casing where it could cause corrosion.

Referring to Figure 9, there is shown in cross-section 15 a view of the support system ~nd multi-zone heater 156 configuration '~r the sample block 12. The ~ample block 12 i5 supported by 'our bolts of which bolt 146 is typical.

These four bo~ts pass through upright member~ of a steel support bracket 148. Two l~rge coil springs _0 and 152 are 20 compressed between a horizontal portion of the support bracket 148 and a steel pressur~ pl~ate 154. The springs 150 and 152 are compressed ~ufficiently to supply approximately 300 lbs. per square inch of force in the positive z direction acting to compress a film heater 156 to the bottom 25 surface 116 of the sample block 12. This three layer film heater structure is comprised of a multi-zone film heater 156, a silicone rubber pad lS8 and a layer of epoxy resin foam 160. In the preferred ~mbodiment the film heater 156 has three separately controllable zones. The purpo~e of the 30 film heater 156 i8 to supply heat to the ~ample block 12 under the control of the CPU 20 in Figure 1. The purpose of the silicone rubber pad 158 is to lower the thermal conductivity from the film heater layer 156 to the structures below. These lower structures ~erve as heat 35 sinks and heat sources between which undesired heat energy may be transferred to and from the sample block 12. The ~ - 31 - 2~ silicone rubber pad 158 has the additional function of compensating for surface irregularities in the film heater 156 since some film heaters embody nichrome wires and may not be not perfectly flat.

~ he purpose of the steel plate 154 and the epoxy resin foam 160 is to transfer the force from the springs 150 and 152 ~o the silicone rubber pad 158 and the multi-zone film heater 156 ~o as to compress the film heater to the bottom surface 116 of the sample block with a8 flush a fit as 10 possible. The epoxy resin foam ~hould be stiff so as to not be crushed under the force of the ~prings but it ~hould al80 be a good i~sulator and should have low thermal m~ss, i.e., it should be a nondense structure. In one embodiment, the foam 160 is m~nufactured under the trademark ECK0 foam. In alternative e ho~i ents, other ~tructures may be ~ubstituted for the silicone rubber layer 158 and/or the epoxy resin foam layer 160. For example, a ~tiff honeycomb structure ~uch a~ i8 used ~n airplane 20 construction could be placed between the pressure plate 154 nnd the film heater 156 with insulating layers therebetween.

Whatever ~tructure is used for layers 158 and 160 should not absorb substantial amounts of heat from the sample block 12 while the block is being heated and should not transfer 25 substantial amounts of heat to the sample bloc~ 12 when the block is being cooled. Perfect isolation of the block from its surrou~ing ~tructures however, is virtually impossible.

Every effort should be made in designing alternative ~tructures that will be in contact with the sample block 12 30 so ~s to thermally isolate the sample block from its environment as much as possible to minimize the thermal mass of the block and enable rapid temperature changes of the sample block ~nd the sample mixtures stored therein.

Precise temperature control of the nample block 35 temperature i8 achieved by the CPU 20 in Figure 1 by controlling the amount of heat applied to the sample block 2 ~ 7 ~ 3 - 32 by the multi-zone film heater 156 in Figure 9. The film heater is driven using a modified form of pulse width modulation. First, the 120 volt waveform from the power line is rectified to preserve only half cycles of the same 5 polarity. Then portions of each half cycle are gated to the appropriate zones of the foil heater, with the percentage of each half cycle which is applied to the various zones of the foil heater being controlled by the CPU 20.

Figure 10 illustrates one embo~; ?nt of a power control 10 concept for the film heater 156. ~igure 10 is a diagram of the voltage waveform of the supply line voltage.

Rectific~tion to eliminate the negative half cycle 162 occurs. Only positive half cycles remain of which half cycle 164 is typical. The CPU 20 ~nd its associated 15 peripheral electronic circuitry then controls the portion of each half cycle which is applied to the various zones of the film heater 156 by selecting a portion of each half cycle to Apply according to a power level computed for each zone based upon equations given below for each zone. That is, 20 the dividing line 166 is moved forward or backward along the time axis to control the amount of power to the film heater based upon a number of factors which are related in a special eguation for each zone. The cross-hatched area under the positive half cycle 164 represents the amount of 25 power applied to the film heater 156 for the illustrated position of the dividing line 166. A~ t~e dividing line 166 is ~oved to the right, more power is applied to the film heater, and the sample block 12 gets hotter. As the dividing line is moved to the left along the ti.me axis, the 30 cross-hatched Area becomes smaller and less power is applied to the film heater. How the CPU 20 and its associated software and peripheral circuitry control the temperature of block 12 will be described in more detail below.

The amount of power supplied to the film heater is 35 continuously variable from 0 to 600 watts. In alternative embodiments, the amount of power supplied to the film heater _ _ _ _ ~ 2~67~L3 156 can be controlled using other schemes such as computer control over the current flow through or voltage ~pplied to a DC film heater or by the zero crossing switching scheme described below.

In other embodiments, heating control of the sample block 12 may be performed by control over the flow rate and/or temperature of hot gases or hot liquid which is gated through heating control channels which are formed through the metal of the sample block 12. Of course in such 10 alternative embodiments, the number of sample wells in the block would have to be reduced &ince there is no room for additional heating channels in the sample block 12 shown in Figures 2 through 8. Such alternative embodiments could still be compatible with the 96-well microtiter plate format 15 if, for example, every other well were removed to make room for a heating chAnnel in the ~ample block. This would provlde compatibility only as to the ~i ?n~ions of such microtiter plates and not as to the simultaneoufi processing of 96 different samples. Care must be taXen to preserve 20 local balance and local symmetry in these alternative embodiments.

In the embodiment described herein, the maximum power that can be delivered to the block via the film heater is 1100 watts. This limitation arises from the thermal 25 conductivity of the block/heater interface. It has been found experimentally that the ~upply of more than approximately 1100 watts to the film heater 156 will frequently cause self-destruction of the device.

Typical power for heating or cooling when controlling 30 block temperatures at or near target i~cl~b~tion temperatures is in the range of plus or minus 50 watts.

Referring to Figure 11, there is shown a time versus temperature plot of a typical PCR protocol. Large downward changes in block temperature are accomplished by gating 35 chilled liquid coolant through the ramp cooling channels while monitoring the sample block temperature by the _ _ * 2 ~ - 34 temperature sensor 21 in Figure 1. Typically these rapid downward temperature changes are c~rried out during the ramp following the denaturation incubation 170 to the temperature of hybridization incubation 172. Typically, the user must 5 specify the protocol by defining the t~mrerAtures and times in one fashion or another so as to describe to the CPU 20 the positions on the temperature/time plane of the checkpoints symbolized by the circled intersections between the ramp legs and the ~ tion legs. Generally, the 10 incubation legs are marked with reference numer~ls 170, 172 and 174 and the ramps are marked with reference numerals 176, 178 and 180. Generally the incubation intervals are conducted at a single temperature, but in alternative embodiments, they may be ~tepped or continuously ramped to 15 different temperatures within a range of temperatures which is acceptable for performing the particular portion of the PCR cycle involved. That i8, the denaturation incubation 170 need not be carried out at one temperature as shown in Figure 11, but may be carried out at any of a plurality of 20 different temperatures within the range of temperatures acceptable ~or denaturation. In some . ho~i ?nts, the user may ~pecify the length of the ramp ~c; ~tc 176, 178 and 180. In other embodiments, the user may only specify the temperature or temperatures and duration of each incubation 25 $nterval, and the instrument will then move the t- ?rature of the sample block as rapidly as possible between incubation t~mr~ratures upon the completion of one incubation and the start of another. In the preferred embodiment, the user can also have t~ -ratures and/or 30 incubation times which are difSerent for each cycle or which automatically increment on ev~ry cycle.

The average power of ramp cooling during a transition from a 95~C denaturation incubation to a 35-C hybridization incubation is more than one kilowatt typically. This 35 results in a temperature change for the cample block of approximately 4-6~C per second when the block temperature is 2 ~ - 35 at the high end of the operating range, and approximately 2~c per second when the block temperature is at the low end of the operating range.

Generally it is desirable to have as high a cooling rate as possible for ramp cooling.

Because so much heat is being removed from the sample block during ramp cooling, temperature gradients across the sample block from one end of a ramp cooling channel to the other could occur.

To prevent this ~nd 1ni ize these types of temperature gradients, the ramp cooling rhAnnel~ are 10 directionally interlaced. Th~t is, in Figure 3, the direction of coolant flow through ramp cooling channels 100, 102, 104, and 106 is into the page as symbolized by the x~5 inside these ramp cooling channel holes. Ramp cooling liquid flow in interlaced ramp cooling channels 101, 103, 15 105, and 107 is out of the page a~ symbolized by the ~ingle points in the center of these rAmp cooling channel holes.

This interlacing plus the high flow rate through the ramp cooling channels minimizes any temperature gradients which might otherwise occur using noninterlaced flow patterns or 20 lower flow rates because the di~tAnces between the hot and cold ends of the channels is made~maller. A slower flow rate result~ in most or all of the heat being t~ken from the block in the first inch or ~o of travel which means that the input side of the block will be at a lower temperature than 25 the ouL~ut ~ide of the block. A high flow rate minimizes the temperature gradient along the channel. Interlacing means the hot end of the chAnnels running in one direction are "sandwiched" between the cold ends of channels wherein flow is in the opposite direction. This i~ a smaller 30 distance than the length of the channel. Thus, temperature gradients are reduced because the distAnces heat must travel to eliminate the temperature gradient are reduced. This causes any t~ ,erature gradients that for~ because of cooling in the ramp channels to be quickly eliminated before 35 they have time to differentially heat some samples and not others. Without interlacing, one side of the sample ~lock ~ - 36 - 2 ~ ~ 7~ 3 would be approximately 1~C hotter than the other side.

Interlacing results in dissipation of any temperature gradients that result in less than approximately 15 seconds.

In order to accurately estimate the amount heat added S to or removed from the block, the CPU 20 measures the block temperature using temperature sensor 21 in ~igure 1 and measures the coolant temperature by way of temperature sensor 61 in Figure 46 coupled to bus 54 in Figure 1. ~he ~mbient air temperature is al80 measured by way of 10 temperature sensor 56 in Figure 1, and the power line voltage, which controls the power applied to the film heaters on bus 52, is also measured. The thermal conductance from the sample block to ambient and from the sample block to the coolAnt are known to the CPU 20 as a 15 result of measurements made during an initializ~tion process to ~et control parameters of the ~ystem.

For good temr~rature uniformity of the ~ample population, the block, at const nt temperature, can have no net heat flow in or out. However, temperature gradients can 20 occur within the sample block arising ~rom local flows of heat from hot spots to cold spots which have zero net heat transfer relative to the block borders. ~or instance, a slab of material which is heated ~t one end and cooled at the other is at a constant average temperature if the net 25 heat flow into the block is zero. However, in this situation a significant temperature nonuniformity, i.e., a temperature gradient, can be established within the slab due to the flow of heat from the hot edge to the cold edge.

When heating and cooling of the edges of the block are 30 ~topped, the flow of heat ~rom the hot edge to the cold edge eventuAlly dissipates this t~ ~rature gradient ~nd the block reaches a uniform temperature throughout which is the average between the hot temperature and cool temperature at the beginning of ~eat flow.

3S If a slab of cross sectional area A in length L has a uniform thermal conductivity K, and the slab is held at _ _ ' 2~67~3 - 37 constant average temperature because heat influx from a heat source Qin is matched by heat outflow to a heat sink Q~t the steady state temperature profile which results from the heat flow is:

Qin L (1) Delta T = K Where, Delta T = the temperature gradient L ~ the thermal path length A = the area of ~he thermal path R ~ the thermal conductance through the path In general, within any material of uniform thermal conductance, the temperature gradient will be established in 15 proportion to the heat flow per unit area. Heat flow and temperature nonuniformity are thus intimately linked.

Practically speaking, it is not possible to control the temperature of a sample block without some heat flow in and out. The cold bias control cooling reguires ~ome heat flow 20 in from the strip heaters to h~lAnce the heat re~oved by the coolant flowing through the bias cooling channels to maintain the block t erature at a stable value. The key to a uniform sample block temperature under these conditions is a geometry which has "local balance" and "local symmetry" 25 of heat sources and heat sinks both statically and dynamically, and which is arranged such that any heat flow from hot spots to cold spots occur~ only over a short distance.

Stated briefly, the concept of "static lo~al balance" 30 means that in a block ~t constant temperature where the total heat input egu~ls the total heat output, the heat sources and heat sinks are arranged such that within a distinct local region, all heat sources are completely balanced by heat sinks in terms of heat flows in and heat 35 flows out of the block. Therefore, each local region, if isolated, would be maintained at a constant temperature.

~ - 38 - 2 ~ ~ 6 7 ~ 3 The concept of "static local symmetry" means that, within a local region and for a constant temperature, the center of mass of heat sources is coincident with the center of mass of heat sinks. If this were not the case, within 5 each local region, a ~emperature gradient across each local region can exist which can add to a temperature gradient in an adjacent local region thereby causing a gradient across the sample block which is twice as lArge as the size of a single local region because of lack of local symmetry even lo though local balance within each local region ~xists. The concepts of local balance and local symmetry ~re importznt to the achievement of a static temperature balance where the temperature of the sample block is being maintained at a constant level during, for example, an incubation interval.

For the dynamic case where rapid temperature changes in the ~ample block are occurring, the thermal mass, or heat capacity of each local region heC~ ?S important. This is because the amount of heat that must flow into each local region to change its t~ ~ature is proportional to the 20 thermal mass of that region.

Therefore, the concept of static local balance can be expanded to the dynamic case by requiring that if a local region includes x percent of the total dynamic heat source and heat sink, it must also include x percent of the thermal 25 mass for "dynamic local balance" to exist. Likewise, "dyn~mic local symmetryN requires that the center of mass of heat capacity be coincident with the center of mass of dynamic heat sources and sinks. What this means in simple terms is th~t the thermal mass of the sample block is the 30 metal thereof, ~nd the ma~hining of the sample block must be symmetrical and balanced such that the total mass of metal within each local zone is the same. Further, the center of mass of the metal in each local zone should be coincident with the center of mass of the dynamic heat sources and 35 sinks. Thus, the center of mass of the multi-zone heater 156, i.e., its geometric center, and the geometric center of 2~7~3 ~ - 39 the bias and ramp cooling channels must coincide.

From a study of Figures 2-9, it will be seen from the detailed discussion below that both static and dynamic local balance and local symmetry exist in sample block 12.

Figure 12 illu5trates two local regions ~ide by side for the design of the ~ample block 12 according to the teachings of the invention.

In Figure 12, the boundaries of two local regions, 200 and 202, are marked by dashed lines 204, 206 and 208.

Figure 12 shows that each local region 10 which is not in the guard band i~ comprised of: two columns of cample wells; a portion of the foil heater 156 which turns out to be 1/8th of the total area of the heater; one ramp cooling channel ~uch a~ ramp cooling chAnnels 210 and 212; and, one bias cooling channel. To preserve local 15 symmetry, each local region is centered on its ramp cooling channel and has one-half on a bias cooling çhAn~e] at each boundary. Fo~ example, local region 200 has a center over the ramp cooling channel 210 and bias cooling channels 214 and 216 are dissected by the local region boundaries 204 and 20 206, respectively. Thus the center of mass of the ramp cooling channel (the middle thereof), coincides (horizontally) with the center of mas~ of the bias cooling channels ~the center of the local region) and with the center of mass of the film heater portion coupled to each 25 local region. Static local hAlAnce will exist in each local region when the CPU 20 is driving the film heater 156 to input an amount of heat ~nergy that i~ egual to the ~mount of h~at energy that i~ being removed by the ramp cooling and bias cooling channels. Dynamic local balance for each local 30 region exists because each local region in the center portion of the block where the 96 ~ample mixture~ reside contains approximately 1/8th the total thermal mass of the entire sample block, contain~ 1/8th of the total number of ramp cooling channels and contains l/8th of the total number 35 of bias cooling channels. Dynamic local symmetry ~xists for each local region, because the center of mass of the metal 2 ~ 3 - 40 of each local region is horizontally coincident with: the center of film heater portion underlying the local region;

the center of the ramp cooling channel; and, the center of mass of the two half bias cooling channels.

By virtue of these physical properties characterized as static ~nd dynamic local balance and local symmetry, the sample block heats and cools all samples in the population much more uniformly than prior art thermal cyclers.

Referring to Figure 2, the plan view of the boundaries lo of the local regions are illustrated by dashed lines 217 through 225. Inspection of Figure 2 reveals that the central region of the 96 sample well~ are divided into six adjacent local regions bounded by boundaries 218 through 224. In addition, two guard band local reyions are added 15 at each edge. The edge local region (local regions are sometimes herein also called local zones) having the most negative x -oordinate is bounded by boundary lines 217 and 218. The edge local region having the most positive x coordinate is bounded by boundary lines 224 and 225. Note 20 that the edge local regions cont in no sample well columns but do contain the groove 78 simulating a column of wells.

The depth and width of the groove 78 i8 designed to remove the same metal mass as a column of wells thereby somewhat preserving dynamic local symmetry. The edge local zones are 25 therefore different in thermal mas3 (they also have additional thermal mass by virtue of the external connections ~uch as manifolds and ~u~pG~ pins) than the six local zones in the central part of the ~ample block. This difference is accounted for by heating the edge local zones 30 or guard bands with separately controllable zones of said multizone heater so that more energy may be put into the guard band than the central zone of the block.

The local regions at each edge of the block approximate, but do not exactly match the thermal properties 35 of the six centrally located local regions. The edge local regions are called Nguard band" regions because they 2~7~3 - 41 complete a guard band which runs around the periphery of the sample block 12. The purpose of thiC guard band is to provide some thermal i~olation of the central portion of the sample block containing the 96 ~ample wells from 5 uncontrolled heat sinks and sources inherently embodied in mechanical connections to the block by such things as support pins, manifolds, drip collars and other devices which must be mechanically affixed to the sAmple block 12.

For example in Figure 2, the edge surfaces 228 and 230 of 10 the ~ample block have plastic manifolds attached thereto which c~rry coolant to and from the ramp ~nd bias cooling passages. The guard band along edges 228 and 230 consists of portions of the slot 78 which are parallel to and closest to the edges 228 and 230. The depth of the groove 78 is 15 such that the bottom of the groove is as close to the perimeters of the bias and ramp cooling channels as is possible without actually intersecting them. The width of the groove 78 coupled with this depth is such that the volume of metal removed by the slot 78 between points 82 and 20 232 in Figure 2 approximately eguals the volume of metal removed by the adjacent row of sample wells starting with sample well 234 and ending with cample well 83. Also, the slot 78 all around the perimeter of the block ifi located approximately where such an additional row of wells would be 25 if the periodic pattern of sample wells were exte~P~ by one row or column of wells in each direction.

Along the edges 250 and 252 where the support connections are made to the sample block, the guard band local regions contain, in addition to a portion of the slot 30 78, the full length of several cooling channels. Referring to Figure 3, these include: 1/2 of a bi~s cooling channel (e.g., g2) which merges with the adjacent 1/2 bias cooling channel of the adjacent local region to form a whole bias cooling channel; a ramp cooling channel (e.g., 100); and a 35 whole bias cooling channel (e.g., 91). ~or the edge local region at edge 250, these cooling channels are 107, 198 and ~ 2~6~3 - 42 99 .

The whole bias cooling channels in the guard bands are slightly displaced inward from the edge of the block. The reason that these whole bias cooling channels are used is 5 because a "half" cooling channel is impractical to build.

Since the bias cooling channels reguire such a thick walled rubber lining, it would be difficult to keep a hole through a lining of a "half" bias cooling channel reliably open.

This asymmetry in the edge local regions cau~es a ~mall 10 excess loss of heat to the coolant from the edge guard band local regions, but it is ~ufficiently remote from the central region of the ~ample block containing the ~ample wellc that its contribution to ~ample temperature nonuniformities is small. Also, ~ince the temrerature 15 affects of this small asymmetry ~re predictable, the effect can be further minimized by the use of a separately controllable zone of the multi-zone heater system under each guard band.

Referring to Figure 13, there are ~hown three 20 separately controlled zones within the film heater layer 156 in Figure 9. These separately cont~olled zones include edge heater zones which are ~ituated under the guard bands at the exposed edges of the sample block 12 which are coupled to the support bracket 148. There are also ~eparately 25 controlled ~anifold heater zones ~ituated under the guard band~ for the edges 228 and 230 which are attached to the coolant manifold~. F~nally, there i~ a central heater zone that underlies the ~ample wells. The power applied to each of these zones is separately controlled by the CPU 20 and 30 the control software.

The film heater 156 is composed of a pattern of electrical conductors formed by etching a thin ~heet of metal alloy such as Inconel~. The metal alloy selected should have high electrical resistance and good resistance 35 to heat. The pattern of conductors so etched i5 bonded between thin sheets of an electrically insulating polymeric 2~ 7~ _ - 43 material such ~s Kapton~.

Whatever material is used to insulate the electrical resi~tance heating element, the material must be resistant to high temperatures, have a high dielectric strength and good mechanical ~tability.

The central zone 254 of the film heater has approximately the same ~i e~cions as the central portion of the sa~ple block inside the guard bands.

Central region 254 delivers a uniform power density to the ~ample well area.

Edge heater region~ 256 and 258 are ~bout a~ wide as 10 the edge guard bands but are not quite as long.

Manifold heater regions 260 and 262 underlie the guard bands for edges 228 and 230 in Figure 2.

The manifold heater zones 260 and 262 Are electrically connected together to form one 6eparately controllable 15 heater zone. Also, the edge heater sections 256 and 258 are electrically coupled together to form a ~aoon~ separately controllable heater zone. The third ~eparately controllable heater zone i8 the central section 254. Each of these three separately controllable heater zones has separate electrical 20 leads, and each zone is controlled by a separate control algorithm which may be run on ~eparate microprocessors or a shared CPU as is done in the preferred em~o~; -nt.

The edge heater zones 256 and 258 are driven to compensate for heat lost to the support brackets. ThiC heat 25 10s6 is proportional to the temperature difference between the 6ample block 12 and the ambient air ~urro~Aing it. The edge heater zones 256 and 258 also compensate for the excess 108s of heat from the sample block to the full bias cooling channels at each edge of the block. This heat loss is 30 proportional to the temperature difference between the sample block 12 and the coolant ~lowing through these bias cooling ~hAnnels.

The manifold heater ~ections 260 and 262 are also driven 80 as to compen5ate for heat 106t to ~he plastic 35 coolant manifolds 266 and 268 in Figure 13 which are attached to the edges of the sample block 12. The power for ~ 44 2Q~7~ _ the manifold heater sectionc 260 and 262 compensates for heat loss which is proportional m~inly to the temperature difference between the sample block and the coolant, and to a lesser degree, between the sample block and the ambient 5 air.

For practical reasons, it is not possi~le to match the thermal mass of the guard band local regions with the thermal masses of the local regions which include the sample wells overlying central heater ~ection 254. For example, 10 the plastic coolant manifolds 266 and 268 not only conduct heat away from the guard band, but they also add a certain amount of thermal mass to the guard band local regions to which they are attached. The result of thi~ i8 that during rapid block ~emperature changes, the rates of rise and fall 15 of guard band temperature do not exactly match that of the sAmple well local regions. Thi~ generates a dynamic t~mrerature gradient between the guard bands and sample wells, which if allowed to become larqe, could persist for a time which is longer than is tolerable. This t~ erature 20 gradient effect is roughly proportional to the rate of change of block t~ -rature and is i n i ized by adding or deleting heat from each guard band local zone at a rate which is proportional to the rate of change of block temperature.

The coefficients of proportionality for the guard band zone heaters are relatively ~table properties of the design of the system, and are determined by engineering measurements on prototypes.

The values for these coefficients of proportionality ~re given below in 30 connection with the definitions of the terms of Equations (3) through (5). These equations define the amounts of power to be applied to the manifold heater zone, the edge heater zone and the c~ntral zone, respectively in an alternative embodiment. The equations used in the preferred 35 e~bodiment are given below in the description of the software (Equations (46)-(48), power distributed by ~rea).

2~7~ -- 45 - (3) P", = Am P ~ K~l (TBL~ ~ TA~) + K~2 (TU~ ~ To~L) + K~3(dt~L~/dt) where, P~ = power supplied to the manifold heater zones 260 and 262.

A~ ~ area of the manifold heater zone.

P ~ power needed to cause the block temperature to stay at or move to the desired t~ _er~ture at any particular time in a PCR thermal cycle protocol.

K~1 ~ an experimentally determined constant of proportionality to c ,~nC~te for ~Yce~fi heat loss to ambient through the mani~olds, equal to 0 watt~/ degree Kelvin.

15 X~2 = an experimentally determined constant of proportionality to c ,~n~te for ~Ysess heat loss to the coolant, equal to 0. 4 watts/degree Kelvin.

K~ ~ an experimentally determined constant of proportionality to provide extra power to compensate for additional thermal mass of the manifold edge guard bands caused by the attachment of the plastic manifolds etc., egual to 66.6 watt-secon~c/degree Kelvin.

25 T~ ~ the temperature of the sample block 12.

T~ ~ the t~ _?rature of the a~bient air.

T~~ e the temperature of the coolant.

dt~/dt ~ the change in sample block t~ ature per unit time.

30 (4) PE = A~P ~ KE1(T~ - T~) I KE2 (T-LK T~L~ + KE3 (dt~/dt) where, PE = power to be applied to the edge heater zones AE ~ the area of the edge heater zones 3 5 KE1 - an experimentally determined constant of ' ' 2~7~3 - 46 proportionality to compensate for excess heat 10s6 to ambient through the manifolds, equal to 0.5 watts/degree Xelvin.

KE2 = an experimentally determined constant of proportionality to compensate for excess heat loss to the coolant, equal to 0.15 watts/degree Kelvin.

~3 ~ an experimentally determined constant of proportionality to provide extra power to compensate for additional thermal mass of the exposed edge guard bands caused by the attachment of the ~ample block 12 to the support pins and bracket, the t- ,erature sensor etc., equ~l to 15.4 watt-sec/degree Kelvin.

(5) Pc ~ A~ P where Pc ~ the power to be applied to the central zone 254 of the multi-zone heater.

20 Ac ~ the area of the central zone 254.

In each of Equations (3) through (5), the power term, P is a variable which is calculated by the portion of the control algorithm run by the CPU 20 in Figure 1 which reads the user defined setpoints and determines what to do next to 25 cause the ~ample ~lock temperature to stay at or ~e~:l? the proper t~ ,erature to implement the PCR temperat~e protocol defined by the time and temperature setpoint~ stored in memory by the user. The manner in which the setpoints are read and the power density i~ calculated will ~e described 30 in more detail below.

The control algorithm run by CPU 20 of Figure 1 senses the temperature of the sample block via temperature sensor 21 in Figure 1 and Figure 9 and bus 52 in Figure 1. This temperature is differentiated to derive the rate of change ~ 2~5~7~3 - 47 of temperature of the ~ample block 12. The CPU then measures the t~mp~rature of the ambient air via temr~rature sensor 56 in Figure 1 and measure~ the temperature of the coolant via the temperature ~ensor 61 in the coolant control 5 system 24 shown in Figure 46. The CPU 20 then computes the power factor correspon~ing to the particular segment of the PCR protocol being implemented and makes three calculations in accordAnce with Eguations (3), (4) and (5) by plugging in all the measured temperatures, the constant~ of 10 proportionality (which are ~torQd in nonvolatile memory), the power factor P for that particular iteration of the control program and the areas of the various heater zones (which are ~tored in nonvolatile memory). The power factor i5 the total power needed to move the block temperature from 15 its current level to the t~ _-rature level epecified by the user via a setpoint. ~ore details on the calculations performed by the CPU to control heating and cooling are given below in the description of the control ~oftware "PID task".

After the required power to be applied to each of the three zones of the heater 156 is calculated, another calculation i- made regarding the ~Lo~G~Lion of ~ach half cycle of input power which i8 to be applied to each zone in &ome 1 ho~ ?~t~.

In the preferred - ho~ t described 25 below, the calculation mode i~ how many half cycles of the total nl her of half cycles which occur during a 200 millisecond ~ample period are to be applied to ~ach zone.

Thi~ process is de~cribed below in conne~tion with the discussion of Figureff 47A and 47B (hereafter referred to as 30 Figure 47) and the ~PID Task" of the control ~oftware. In the alternative ~ ho~ ?nt ~ymbolized by Figure 10, the computer calculates for ~ach zone, the position of the dividing line 166 in Figure 10. After this calculation is performed, appropriate control signal~ are generated to 35 cause the power ~upplies for the multi-zone heater 156 to do the appropriate ~witching to cause the calculated amount of 2~7~3 - 48 power for each zone to be applied thereto.

In alternative embodiments, the multi-zone heater can be implemented using a single film heater which delivers uniform power density to the entire sample block, plus one 5 or two additional film heaters with only one zone apiece for the guard bands. These additional heaters are ~uperimposed over the single film heater that covers the entire ~ample block. In ~uch an . ho~ -nt, only the power nece~s~ry to make up the guard band losses is delivered to the additional 10 heater zones.

The power factor P in Equations (3) through (5) is calculated by the CPU 20 for variou~ points on the PCR temperature protocol based upon the set points and ramp times specified by the user. However, a limitation is 15 imposed based upon the ~Yi power delivery capability of the zone heater mentioned above.

The constants of proportionality in Equations (3) through (5) must be properly ~et to adequately compensate for excess heat 106se6 in the guard band for good 20 temperature uniformity.

Referring to Figure 17, there~is ~hown a graph of the differences between calculated ~ample tcmperatures for a plurality of different sample in response to a step change in block te~erature to raise the t~ -r~ture of the sample 25 block toward a denaturation jnc~ tion target t~ -rature of approximately 94-C from a ~ubstantially lower ~emperature.

Figure 17 illustrates the calculated sample liquid temperature~ when the multi-zone heater 156 i~ properly managed u~ing the constants of ~.o~GLLionality given above 30 in the definitions of the terms for Eguations (3) through (5). The various wells which were used to derive the graph of Figure 17 are indicated thereon by a single letter and number combination. The 8 x 12 well array ~howing Figure 2 is coded by lettered columns and numbered row6. Thus, for 35 example, sample well 90 is also designated sample well A12, while sample well 89 is also designated ~ample well B12.

~ 2~743 - 49 Likewi~e, ~ample well 68 is al~o designated ~ample well D6, and r,o on. Note that the well temperatures ~ettle in asymptotically at temperatures which are within approximately 0.5-C of each other ~ecau~e of the overall 5 thermal design described herein to eliminate temperature gradients.

The foregoing description illustr~tes how the sample block t~ ~rature may be controlled to be uniform and to be quickly changeable. ~owever, in the PCR ~ GCe~ it is the 10 temperature of the sample reaction mixture and not the block temperature that is to be ~ G~L ammed. In the preferred embodiment according to the teachings of the invention, the user specifies a sequence of target temperature8 for the s~mple li~uid itself and ~pecifie~ the inc~h~tion ti~es for 15 the samDle li~uid at each of these tarset te~era~res for each stage in the PCR process. The CPU 20 then manages the snmple ~lock temperature so as to get the ~ample reaction mixture~ to the ~pecified target incu~-tion tQmperatures and to hold the ~ample mixtures at the~e target t~ _-ratures for 20 the ~pecified incubation times. The u~er interf~ce code run by the CPU 20 di~play~, at all ~tage~ of this ~c_~s~, the current calculated ~ample liguid temperature on the di~play of terminal 16.

The difficulty with displ&ying an actual measured 25 s~mple tumperature i8 that to measure the actual temperature of the reaction mixture reguires ~nsertion of a temperature ~ lring probe therein. The thermal mass of th~ probe can significantly alter the t. erature of any w-ll in which it is placed ~ince the sample reaction nixture in any 30 particular well is often only 100 ~icrolit~r~ ~n volume.

Thus, the mere insertion of ~ temperature probe into a reaction mixture can cause a temperature gradient to exist between that reaction mixture ~nd neighboring mi~u~es.

Since the extra thermal mass of the temperature rrn~or would 35 cause the reaction mixture in which it i8 immer-ed to lag behind in te~perature from the temperatures of the reaction 2 ~ 5 ~ 7 ~ 3 - 50 mixtures in other wells that have les6 therm~l mas~, errors can result in the amplification ~imply by attempting to measure the t~ erature Accordingly, the instrument described herein calculates S the sample temperature from known factors ~uch a~ the block temperature history and the thermal time constant of the system and displays this sample temperature on the display It has been found experimentally for the ~yst~m described herein that if the ~ample tube~ are pre~-d down into the 10 sample wells with at lea~t a mini~um thre~hold force F, then for the size and shape of the sample tubes used in the preferred embodiment ~nd the sample volumes of approximately 100 ~icroliters, thermally driven convection o~ s within the sample reaction mixture and the ~ystem acts thermally 15 like a single time constant, linear ~y~tem Experiments have shown that each ~ample tube ~ust be p~he~ down with approximately 50 grams of force for good well-wall-to-liquid thermal conductivity from well to well The heated platen design described below i~ designed to push down on each 20 sample tube with about 100 gram~ of force This minimum force, symbolized by force vector F in Figure 15, is n*cessAry to insure that regardless of slight differences in external ~i -n~ions ~s between vAriou~ ~ample tubes and variou~ ~ample wells in the ~ample block, they all will be 25 pushed down ~ith ~ufficient force to guarantee t~e ~nug and flush fit ~or ~ach tube to guarant-- uniform thermal conductivity Any de~ign which ha~ ~ome ~ample tubes with loose fits in their COL~C~O~A;n7 ~ample well8 and ~ome tube~ with tight fit~ will not be ablc to achieve tight 30 temperature control for all tubes b~cau~e of non-uniform thermal conductivity An insufficient level of force F result~ in a t: ~rature response of the sample liquid to a step change in block temperature a8 ~hown at 286 in Figure 14 An adequate level of force F results in the temperature 35 response shown at 282 The result achieved by the ~pparatus constructed CA 020~6743 l999-Ol-l~ - 51 according to the teachings of the invention is that the temperature of each sample mixture behaves as if the sample is being well mixed physically during transitions to new temperatures. In fact, because of the convection currents caused in each sample mixture, the sample reaction mixture in each sample tube is being well mixed.

The surprising result is that the thermal behavior of the entire system is like an electrical RC circuit with a single time constant of 9 seconds which is about 1.44 times the half-life of the decay of the difference between the block temperature and the sample temperature. A GeneAmp* sample tube filled with 50 milliliters of sample has a time constant of about 23 seconds.

In other words, during an upward change in temperature of the sample block, the temperature of the reaction mixture acts like the rise in voltage on the capacitor C in a series RC electrical circuit like that shown in Figure 16D in response to a step change in the voltage output of the voltage source V.

To illustrate these concepts, refer to Figure 14 which shows different temperature responses of the sample liquid to a step change in block temperature and to Figure 15 which shows a cross section through a sample well/sample tube combination. It has been found experimentally that when the volume of sample liquid 276 is approximately 100 microliters and the dimensions of the tube are such that the meniscus 278 is located below the top surface 280 of the sample block 12, and the force F pushing the sample tube into the sample well is at least 30 grams, the thermal time constant I (tau) of the system shown in Figure 15 is approximately nine seconds for a sample tube wall thickness in the conical section of 0.009 inches (dimension A). It has also been found experimentally that for these conditions, the thermal time constant I varies by about 1 second for every 0.001 inch change in wall thickness for the sample tube frustum (cone).

The thin-walled sample tubes described herein have been found to have thermal time constants of from about 5 to about 14 seconds when containing from 20 to 100 microliters of sample. Thicker tube walls result in longer time constants and more lag between a change in sample block temperature ~ 2 ~ 3 - 52 and the resulting change in ~ample liquid temperature.

M~thematically, the expre~sion for the thermal response of the sample liquid temperature to a change in temperature of the ~ample block is:

s ( 6 ) T~ T (l-e~~ where T,~pl. ~ the tr ?rature of the ~ample liquid ~T ~ the t~ -r~ture diffcrence between the temperature of the samplc block 12 ~nd the temperature of the ~ample l$quid t ~ elapsed time r ~ thermal time constant of the sy~tem, or the heat capacity of cample divided by the lS thermal conductance from sample well wall to the ~ample liquid In Figure 14, the curve 282 represents this exponential temperature response to a theoretical ~tep change in sample block temperature when the force F p-~h i ng down on the 20 ~ample tube is sufficiently high. The step change in temperature of the ~ample block i8 shown as function 284, with rapid rise in t~ erature starting at time T~.

Note how the temperature of the cample liquid exponentially increases in re~ponse to the step change and asymptotically approaches 25 the final s~pl- block t~- -rature. As ~entioned briefly u~ov~, the curve 286 ~p~esents the thcrmal ~eapor.se when the downward ~eating force F in Figure lS i~ in~ufficient to cause a ~nuq, flush fit between ~he cone of the sample tube and the w~ll 290 of the ~ample well. Generally, the thermal 30 response of curve 286 will result if the force F i~ less than 30 grams. Note that although Figure lS ~hows a ~mall layer of air between the cone of the sample t:u~e and the sample well wall for clarity, this i~ ~xactly the opposite of the desired situation ~ince air i~ a good in~ulator and 35 would substantially increase the thermal time constant of the system.

2~67~ - s3 The thermal time constant r i~ analogous to the RC time constant in a ~eries RC circuit where R corresponds to the thermal reslstance between the wall of the sAmple well and the sample liquid and C is the heat capacity of the sample 5 liquid. Thermal re~istance is egual to the inverse of thermal conductance which is expressed in units watt~seconds per degree Xelvin.

Because of the convection current~ 292 shown in the sample liquid in Figure 15, everywhere in the reaction lo mixture the sample liquid i~ at very near~y the aame temperature, and the flow of heat between the block and the sample is very nearly proportional to the difference in temperature between the sample block and the sample reaction mixture. The constant of proportionality i8 the thermal 15 conductance between the wall of the ~ample well in the ~ample block 12 and the reaction mixture. For different sample volumes or different tubes, i.e., different wall thicknesses or materials, the th¢rmal time constant will be different. In ~uch a case, the user can as part of his 20 6pecification of the PCR protocol enter the ~ample volume or tube type and the machine will automatically look up the correct thermal time con~tant for use in calculating the ~ample temperature. In some embodiments, the user may enter the actual time constant, and the machine will use it for 25 sample t~- eratur~ temperature calculation.

To keep the thermal time constant a~ ~mall a8 possible, the conical wall6 of the ~ample tubes ~hould be as thin as possible.

In the preferred embodiment, these conical walls are 0.009 inches thick whereas the wall~ of the cylindrical 30 portion of the s~mple tube ~re 0.030 ~nch~6 thick. The conical shape of the ~ample tube provide~ a relatively large ~urface area of contact with the metal of the sample well wall in relation to the volume of the 6ample mixture.

Molding of the ~ample tubes i8 done using a Hcold 35 runner" sys~em and a four cavity mold such that four ~ample tubes are molded at each injection. The molten plastic is 2 ~ 3 - 54 injected at the tip of the ~ample tube cone 80 that any remnant of plastic will project into the cavity 291 between the tip of the sample tube and the tip of the ~ample well.

This prevents any remnant from interfering with the flush 5 fit between the tube and the well. A ~xi limit of 0.030 inches is placed on the size of any remnant plastic.

In various embodiments, 3 different grades of polypropylene each with different advantages can be used.

The preferred polypropylene is PD701 from Himont ~ecause it 10 is autoclavable. However this plastic is difficult to mold because it has a low melt index. This plastic has a melt index of 35 and a molecular density of 9. PD701 tends to leave flash and creates ~omewhat spotty quality parts but would work better if it was injected into the thick walled 15 part of the mold instead of at the tip of the conical section as i~ currently done. Generally, it is desirable to have a high melt index for ease of molding but also a high molecular density to ~aintain good ~trength and to prevent crazing or cracks under the thermal ~tress of the 20 autoclaving process at 260~F. Another plastic, PPW 1780 from American Hoescht has a melt index of 75 and a molecular density of 9 and i3 autoclavable. Another plastlc which may be used in some e:ho~ments is Himont ~44. This plastic is not autoclavable and needs to be sterilized in another ~nner.

In alternative . ho~iments, the tubes may be molded using a ~hot runner" or ~hot nozzle" system where the temperature of the molten plastic is con~olled right up to the gate of the mold. Also, in some . ho~ -nts, ~ultiple 30 gates may be u~ed. However, neither of thess t~chn~gues has been experimentally proven at the time of filing to be better than the currently used ~cold runner" sy~tem.

The fact that the ~ystem acts thermally like a si~gle time constan~ RC circuit is an important result, because it 35 means that if the thermal conductance from the sample blo~k to the sample reaction mixture is known and uniform, the , 2~7~3 ~ - 54A thermal response of the sample mixtures will be known and uniform. Since the heat capacity of the sample reaction mixture is known and constant, the temperature oP the sample reaction mixture can be computed accurately using only the 5 measured history of the block temperature over time. This eliminates the need to measure the ~ample temperature thereby eliminating the errors and mechanical difficulty of putting a probe with nonnegligible thermal mass into a sample well to measure the sample t~ raturQ directly lO thereby changing the thermal mass of the sa~ple in the probed well.

The algorithm which makes this calculation ~~el 5 the thermal behavior of the system after a ~ingle time constant series R-C electrical circuit. This model uses the ratio of 15 the heat c pacity of the liquid sample divided by the thermal conductance from the ~a_ple block to the sample reaction mixture. The heat capacity of the sample reaction mixture is equal to the specific heat of the liquid times the mass of the liquid. The thermal resistance i8 equal to 20 one over the thermal conductance from the ~ample block to the liquid ~eaction mixture through the cample tube walls.

When this ratio of heat capacity divided by thermal conductance i6 expressed in consistent units, it has the dimension of time. For a fixed ~ample volume and a fixed 25 sample composition both of which are the came in every sample well and a fixed thermal conductance, the ratio is also a constant for every sample well, and i8 called the thermal time constant of the ~ystem. It i~ the time required for the ~ample t: ,erature to come within 36.8% of 30 the block temperature a~ter a sudden ~tep change in the blocX temperature.

There is a mathematical theorem used in the analysis of electronic circuits that holds that it i~ possible to calculate the output response of a filter or other linear 35 system if one knows the impulse response of the system.

This impulse response is also known as the transfer 1 3 - 54B function. In the case of a series RC circuit, the impulse response is an exponential function as shown in Figure 16A.

~he impulse stimulu~ resulting in the response of Figure 16A is a5 ~hown in Figure 16B. The mathematical theorem 5 referred to above holds that the output response of such a linear system can be determined by calculating the convolution of the input signal and a weighting function where the weighting function is the impulse response of the system reversed in time. The convolution is otherwise known lo as a running weighted average although a convolution is a concept in calculus with infinitely small step sizes whereas a running weighted average has discreet step ~izes, ~.e., multiple samples. The impulce response of the series RC circuit shown in Figure 16D as such that when the voltage of 15 the voltage generator V ~uddenly rises and falls with a spike of voltage as shown in Figure 16B, the voltage on the capacitor C suddenly rises to a peak at 294 in Figure 16A which is equal to the peak voltage of the impul~e shown in Figure 16B and then exponentially decays back to the steady 20 state voltage V1. The resulting weighting function is the impulse respon~e of Figure 16A turned around in time as shown in Figure 16C at 385.

Superi~posed upon Figure 16C is a hypothetical curve 387 illustrating a typical t~mp~rature history for the 25 temperature of the ~ample block 12 for an approximate step change in t. ,?rature. Also shown superi ,s~ed upon Figure 16C are the times of five t- ,-rature sample periods labelled T1 through T5. According to the teachings of the invention, the sample temperature is calculated by 30 multiplying the t- ,-rature at ~ach one of the~e times T1 through T5 by the value of the weighting function at that particular time and then summing all these products and dividing by 5. The fact that the thermal system acts like a single ti~e constant linear circuit is a surpri~ing result 35 based upon the complexities of thermal heat transfer considerations for this complicated thermal system.

2~7~ In one embodiment, the calculation of the sample temperature is adjusted by a ~hort delay to account for transport lag caused by different thermal path lengths to the block temperature sensor And the sample liquid. The S calculated sample temperature is displayed for the user's information on the terminal 16 shown in Figure 1.

Figure 17 shows the t ,erature response results for six different wellfi spread throughout the 96 well sample block for a ~tep change in sample block temperature from a 10 relatively lower temperature in the hybridization/extension temperature r~nge to the relatively higher temperature of approximately 94-C used for denaturation. The graph of Figure 17 ~hows good agreement between the predicted exponential rise in sample t~ ,erature if the ~ystem were 15 perfectly analogous to the series RC circuit shown in Figure 16D, and al~o shows excellent uniformity of temperature response in that the temperatures of the six ~ample wells used for this study asymptotically ~ettle in at t~ ratures very elose to each other and in a denaturation temperature 20 Ntolerance" band which is approximately O.S-C wide.

In one - hoA ~ ~ent, t~he ten most recent block temperature samples are used for the running weighted average, but in other embodiments a different nl her of tempersature history ~amples may be used. The good 25 agree~ent with theoretically predicted results stems from the fact that the thermal convection current~ make the sample liquids well mixed thereby causing the ay~tem to act in a linear ~ashion.

The uniformity between ~ample temperatures in various 30 ~ample wells spread throughout the 96 well array results from dynamic and static local bal~nce and local ~ymmetry in the sample block ~tructure as well as all the other thermal design factors detailed herein. Note however that during rapid t~ ~erature changes all the cample well~ will h~ve 35 temperatures within o.soc of each other only if the user has carefully loaded each sample well with the ~ame mass of ~ 2 ~ 3 - 56 sample liquid. Inequality of mass in different wells does not cause u~equal temperatures in ~teady ~tate, ~nchAnging conditions, only during rapid changes. The ~as6 of the sample liguid in each well is the ~ ;n~nt factor in 5 determining the heat capacity of each sample and, therefore, is the dominant factor in the thermal time constant for that particular sample well.

Note that the ability to cause the sample liguid in all the sample wells to cycle up and down in temperature in 10 unison and to stabilize at target temperature~ very near each other, i.e., in tolerance bands that are only 0.5~C wide, al50 depends upon the force F in Figure 15. This force must eYcee~ a minimum threshold force before the thermal time constants of all sample wells loaded with 15 similar mas~es of sample liguid will have the ~ame time constant. This minimum threshold force has been experimentally determined to be 30 grams for the 6ample tube and ~ample well configuration described herein. For higher levels of accuracy, the ; n; threshold force F in Figure 20 15 should be established at at leAst 50 grams and preferably 100 grams for an additional margin of safety as noted above.

The importance of thermal uniformity in ~ample well temperature can be appreciated by reference to Figure 18.

This figure ~hows the relationship between the amount of DNA 25 generated in a PCR cycle and the actual ~ample t~ ?rature during the denaturation inter~al for one instance of ampl$fication of a certain F~_ ~nt of DNA. The slope of function 2g8 betw-en temperature~ 93 and 95 degrees centigrade i5 approximately 8~ per degree centigrade for 30 this particular segment of DNA and primers. Figure 18 ~hows the general shape of the curve which relates the amount of DNA generated by amplification, but the details of the shape of the curve vary with every different case of primer~ and DNA target. Temperatures for denaturation above 97 degrees 3s centigrade are generally too hot and result in decreasing amplification for increasing denaturation temperature.

~ - 57 Temperatures between 95 and 97 degrees centigrade A re generally just right.

Figure 18 illustrates that any ~ample well containing this particular DNA target and primer combination which 5 ~tabilizes at a denaturation t~ ,~rature of approximately g3~C is likely to have 8S less DNA generated over the course of a typical PCR protocol than wells denatured at 94~C.

Likewise, ~ample liquids of this mixture that ~tabilize at denaturation te eratures of g5-C are likely to hav- 8~ more 10 DNA generated therein than i~ generated in ~ample wells which stabilize at denaturation t~ -rature6 of 94~C.

Because all curves of this nature have the ~ame general shape, it is important to have uniformity in ~ample temperature.

The sample t. ,eratures calculated as described above are used by the CO~LO1 algorithm for controlling the heaters and flow through the ramp cooling ~hA~els and to determine how long the samples have been held at various target t- ~ratures. The control algorithm u~es these times 20 for c~ ,~rison with the desired times for each i~C~h~tion period as entered by the user. ~hen the times match, the control algorithm takes the appropriate StQp~ to heat or cool the samplQ block toward the target to ,-r~ture defined by the user for the next ~nc~h~tion.

When t~e calculated ~ample temperature is within one degree centigrade of the setpoint, i.e., the incllh~tion te ,~rature programmed by the user, the control pLG~r,~ causes a timer to start.

This timer may be preset to count down from A n~er set ~o a~ to time out the interval 30 specified by the user for the ~ncllh~tion being performed.

The timer ~art~ to count down from the preset count when the calculated sample t~ ,-rature i8 within one degree centigrade.

When the timer reaches a z~ro CouAt, a ~ignal is activated which causes the CPU to take actions to 35 implement the next ~egment of the PCR protocol. Any way to time the specified interval will suffice for purposes of 2~7~ - 58 practicing the invention.

Typically, the tolerance band around any particular targe~ temperature i~ plus or minus 0.5~C. Once the target temperature is reached, the computer holds the 6ample block 5 at the target temperature using the bias cooling channel~ and the film heater such that all the samples remzin close to the target te erature for the specified interval.

For the thermal ~ystem described herein to work well, the thermal conductance from the ~ample block to each sample 10 must be known and uniform to within a very close tolerance.

Otherwise, not all samples will be held within the specified tolernnce band of the target temperature when the timer starts and, not all the samples will experience the same incubation intervals at the target temperature.

~lso, for this thermal sy~tem to work well, all 6ample tubes must be isolated from variables in the ambient environment. That is, it is undesirable for some sample tubes to be cooled by drafts while other sample tubes in different physical positions do not experience the 6ame 20 cooling effects. For good uniformity it is highly desirable that the temperatures of all the samples be determined by the t-- ~rature of the sample block and by nothing else.

Isolation of the tubes from the ambient, and application of the ;~i ~m threshold force F pll~h~ng down on 25 the sample tubes i8 achieved by a heated cover over the sample tubes and sample block.

I~ven though the sample liquid is in a ~ample tube pre~sed tightly into a tempera~u~L controlled metal block, tiyht}y capped, with a men~c~l~ well below the surface of 30 the temperaturc coi.Lrolled metal block, the ~amples ~till lose their heat upward by convection. Significantly, when the 6 mple is very hot (the denaturation temperature is typically near the boiling point of the sample llquid), the sample liguid can lose a very significant amount of hcat by 35 refluxing of water vapor. In this process, water evaporates from the surface of the hot sample liguid and condenses on ~ 7 4 3 - 59 the inner walls of the cap and the cooler upper parts of the sample tube above the top surface of the sample block. If there is a relatively large volume of ~ample, condensation continues, ~nd condensate builds up and runC back down the 5 walls of the sample tube into the reaction mixture. This "refluxing" process carries about 2300 joule~ of heat per gram of water refluxed. This process can cause a drop of ~everal degrees in the ~urface t~ ?rature of a 100 microliter reaction mixture theraby causing a large 10 reduc-tion of efficiency of the reaction.

If the reaction mixture is small, say 20 ~icroliters, and the sample tube has a relatively large surface area above the top surface of the sample block, a ~ignificant fraction of the water in the reaction ~ixture may evaporate.

15 Thi~ water ~ay then con~n~? inside the upper part of the sample tube and L~ there by surface tension during the remainder of the high temperat~re part of the cycle.

This can ~o concentrate the rem~ining reaction mixture that the react:ion i8 impaired or fails completely.

In the prior art PCR thermal cyclers, this refluxing problem was dealt with by overlaying the reaction mixture with a layer of oil or melted wax.

This ~ i~cible layer of oil or wax floated on the agueous reaction mixture and prevented rapid evaporation.

However, labor was reguired to 25 add the oil which raisQd p~c~ ing costs. Further, the pre6ence of oil interfered with later step~ of ~ ocessing and nnalysis and created a po~sibility of cont~ in~tion of the sample.

In fact, it i8 known that industrial grade mineral oil~ have in the past contaminated samples by the 30 unknown pre~ence of conta inating factors in the oil which were unknown to the users.

The need for an oil overlay is eliminated, and the problems of heat loss and concentration of the reaction mixture by evaporation and UnPL edictable thermal effects 35 cAused by refluxing are avoided according to the teac~ings of the invention by enclosing the volume a~ove the sample ~ 2~5~7~3 - 60 block into which the upper parts of the sample tubes project and by heating this volume fro~ above by a heated cover sometimes hereafter also called the platen.

Referring to Figure 19, there is shown a cross 5 sectional view of the ~tructure which is used to enclose the sample tubes and apply downward force thereto ~o as to supply the minimum threshold force F in Figure 15. A heated platen 14 is coupled to a lead ~crew 312 80 as to move up and down along the axis symbolized by arrow 314 with 10 rotation of the lead screw 312. The lead screw i8 threaded through an opening in a sliding cover 316 and is turned by a knob 318. The platen 314 is heated to a temperature above the boiling point of w~ter by resistance heaters (not shown) controlled by computer 20.

~he sliding cover 316 slides back and forth along the Y axis on rails 320 and 322. The cover 316 includes vertical sides 317 and 319 and also includes vertical sides parallel to the X-Z plane (not shown) which enclose the sample block 12 and sample tubes. This structure 20 substantially prevent drafts from acting on the ~ample tubes of which tubes 324 ~nd 326 are typical.

Figure 20 is a perspective view of the sliding cover 316 and sample block 12 with the sliding cover in retracted position to allow access to the ~ample block. The sliding 25 cover 316 ~P~ hles the lid of a rectangular box with vertical wall 328 having a portion 330 removed to allow the sliding cover 316 to slide over the ~ample block 12. The sliding cover i5 moved along the Y axis in Figure 20 until the cover i8 centered over the sample block 12. The user 30 then turns the knob 318 in a direction to lower the heated platen 14 until a mar~ 332 on th~ knob 318 lines up with a mark 334 on an escutcheon plate 336. In ~ome embodiments, the escutcheon plate 336 may be permanently affixed to the top eurface of the sliding cover 316. In other . hoAi ?nts, 35 the escutcheon 336 may be rotatable such that the index mark 334 may be placed in different positions when different size 2~5~7~3 ~ -- 61 -sample tu~eE are used In other words, if taller sample tubes are used, the heated platen 14 need not be lowered as much to apply the minimum threshold force F in Figure 15 In use, the user screws the screw 318 to lower the platen 14 5 until the index marks line up The user then knows that the minimum threshold force F will have been applied to each sample tube Referring jointly to Figures 15 and 19, prior to lowering the heated platen 14 in Figure 19, the plastic cap 10 338 for each ~ample tube ~ticks up about 0 5 millimeter~ above the level of the top of the walls of a plastic tray 340 (Figure 19) which holds all the sample tubes; in a loose 8x12 array on 9 millimeter centers The array of sample wells can hold up to 96 MicroAmp~ PCR tubes of 100 ~L 15 capacity or 48 larger GeneAmp~ tubes of 0 5 ml capacity The details of this tr~y will be cl;-cl~se1 in gr~ater detail below The tray 340 hAs a planar surface having an 8x12 array of holes for ~;ample tubes This planar surface is shown in Figures lS and 19 a8 a horizontal line which 20 intersects the sample tubes 324 and 326 in Figure 19 Tray 340 nlso has four vertical walls t,wo of which are shown _t 342 and 344 in Fis~ure 19 The top level of these vertical walls, shown at 346 in Figure lS, establishes a rectangular box which defines a reference plane As be~t ~een in Figure 15, the caps 338 for all the sample tubes project above this reference plane 346 by ~ome small amount which is designed to allow the caps 338 to be softened and deformed by the heated platen 14 and "s~ Ashe~" down to the level of the referencs plane 346 In the 30 preferred ~ ho~j ?nt, the heated platen 14 i~ kept at a temperature of 105~C by the CPU 20 in Figure 1 and the bus 22 coupled to resi~tance heaters (not chown) in the platen 14 In the preferred embodiment, the knob 318 in Figure 19 and the lead screw 312 are turned until the heat~d platen 14 35 descends to and makes contact with the tops of the caps 338 In the preferred embodiment, the CAps 338 for the sample 2~7~ - 62 tube~ are made of polypropylene .

These cap~ soften ~hortly after they come into contact with the heated platen 14. As the caps soften, they deform, but they do not lose all of their elasticity. After contacting the caps, the heated 5 platen is lowered further until it rests upon the reference plane 346. This further lowering deforms the caps 338 and causes a minimum threshold force F of at least 50 grams to push down on each ~ample tube to keep each tube well ~eated firmly in its sample well. The a~ount by which t~e caps 338 10 project above the reference plane 346, and the ~mount of deformation and residual elasticity when the heated platen 14 rests upon the reference plane 346 i8 designed such that a minimum threshold force F of at least 50 grams and preferably 100 grams will have been achieved for all sample 15 tubes then present after the heated platen 14 has descended to the level of the reference plane 346.

The heated platen 14 and the four vertical walls and planar surface of the tray 340 form a heated, sealed compartment when the platen 14 is in contact with the top 20 edge 346 of the tray. The plastic of the tray 340 has a relatively poor thermal conductivity property. It has been found experimentally that contacting the heated platen 14 with the caps 338 and the i~olation of the portion of the sample tubes 288 which project above the top level 280 of 25 the ~ample block 12 by a wall of material which has relatively poor thermal conductivity has a beneficial roesult. With this structure, the entire upper part of the tube and cap are brought to a t~ erature which is high enough that littlo or no co~e~-tion forms on the in~ide 30 surfaces of ~he tube and cap ~ince the heated pl~ten i~ kept at a temperature above the boiling point of water. This is true even when the ~ample liquid 276 in Figure 15 iG heated to a temperature near its boiling point. This eliminates the need for a layer of immiscible material ~uch as oil or 35 wax floating on top of the sample mixture 276 thereby reducing the amount of labor involved in a PCR reaction and ~ 2~7~3 - 63 eliminating one source of po~sible cont~ ;nAtion of the sample.

It has been found experimentally that in spite of the very high temperature of the heated cover and its close 5 proximity to the sample block 12, there is little affect on the ability of the sample block 12 to cycle accurately and rapidly between high and low temperatures.

The heated platen 14 prevents cooling of the samples by the refluxing process noted earlier because it keeps the 10 temperature of the caps above the condensation point of water thereby keeping the insides of the caps dry. This also prevents the formation of aerosol~ when the caps are removed from the tubes.

In alternative embodiments, any means by which the 15 minimum acceptable downward force F in Figure 15 can be applied to each individual sample tube regardless of the number of sample tubes present and which will prevent condensation and refluxing and convection cooling will suffice for purpo~es of practicing the invention. The 20 application of this downward force F and the use of heat to prevent refluxing and undesired ~ample liguid concentration need not be both implemented by the same system a8 is done in the preferred embodiment.

The sample tubes may vary by a few thousandths of an 25 inch in their overall height. Further, the caps for the sample tubes may also vary in height by a few thousandths of an inch. Al60, each conical ~ample well in the ~ample block 12 may not ~e drilled to exactly the ~ame depth, and each conical sample well in the sample block may be drilled to a 30 slightly different diameter ~nd angle. Thus, when a population of capped tubes i~ pl~ced in the ~ample block so as to be seated in the corresponding sample well, the tops of the caps will not all nec~s-rily be at the fiame height.

The worst case discrepancy for this height could be as much 35 as 0.5 millimeters from the highest to the lowest tubes.

If a perfectly flat unheated platen 14 mowlted so that ~ 2~7~3 - 64 it is free to find its own position were to be pressed down on such an array of caps, it would first touch the three tallest tubes. As further pressure was applied and the tallest tubes were compressed somewhat, the platen would 5 begin to touch some caps of lower tubes. There is a distinct possibility that unless the tube and cap assemblies were compliant, the tallest tubes would be dam ged before the shortest tubes were contacted at all. Alternatively, the force necessary to compress all the tall tubes 10 sufficiently so as to contact the shortest tube could be too large for the device to apply. In either case, one or more short tubes might not be pres~ed down at all or ~ight be pressed down with an in~ufficient amount of force to gu~rantee that the thermal time constant for that tube was 15 equal to the thermal time constants for all the other tubes.

This would result in the failure to achieve the ~ame PCR cycle for all tubes in the ~ample block since some tubes with different thermal time constants woulc not be in step with the other tube~. Heating the platen and ~ofteni~g the 20 caps eliminates these risks by eliminating the manufacturing tolerance crrors which lead to dif~ering tube heights as a factor.

In an alternative embo~i ?nt, the entire heated platen 14 iç covered with a compliant rubber layer. A compliant 25 rubber layer on the heated platen would solve the height tolerance problem, but would al~o act as a thermal insul~tion layer which would delay the flow of heat from the heated platen to the tube caps. Further, with long use at high temperatures, most rubber materials deteriorate or 30 become hard. It is therefore de~irable that the heated platen surface be a metal and a good conductor of heat.

In another alternative embodiment, 96 individual springs could be mounted on the platen 80 that each spring individually presses down on a single ~ample tube. This is 35 a complex and costly ~olution, however, and it requires that the platen be aligned over the tube array with a mechanical - 65 - ~ 7 4 ~ precision which would be difficult or bother~ome to achieve.

The necessary individual compliance for each sample ~ube in the preferred embodiment is supplied by the use of plastic caps which collapse in a predictable way under the 5 force from the platen but which, even when collapsed, still exert a downward force F on the sample tubes which is adequate to keep each sample tube seated firmly in its well.

In the sample tube cap 338 shown in Figure 15, the surface 350 should be free of nicks, flash and cuts ~o that 10 it can provide a hermetic seal with the inner walls 352 of the sample tube'288. In the preferred embodiment, the material for the c~p is polypropylene. A suitable material might be Valtec*HH-444 or PD701 polypropylene manufactured by Himont as described above or PPW 1780 by American 15 Hoescht. In the preferred ~mho~iment, the wall thickness for the domed portion of the cap is 0.130 + .000 - 0.005 inches.

The thickness of the shoulder portion 356 is 0.025 inches and the width of the domed shaped portion of the cap is 0.203 inches in the preferred ~ ho~i -~t.

Any material and configuration for the caps which will cause the in; threshold force F in Figure 15 to be applied to all the sample tubes and which will allow the CAp and upper portions of the sample tubes to be heated to a temperature high enough to prevent condensation and 25 refluxing will ~uffice for purposes of practicing the invention. The dome ~haped cap 338 has a thin wall to aid in deformation of the cap. Because the heated platen is kept at a high tr _-rature, the wall thickness of the domed ~haped cap can be thick enough to be easily manufactured by 30 injection molding since the necessA~y compliance to account for differences in tube height is not necessA y at room temperature.

The platen can be kept at a temperature anywhere from 940C to 110~C according to the teachings of the invention 35 although the range from 100~C to llO-C is preferred to prevent refluxing since the boiling point of water is 100~C.

* Trade-mark B 2 ~ 3 ~ - 66 In this temperature range, it has been experimentally found that the caps soften just enough to collapse easily by as much as 1 millimeter. Studies have ~hown that the elastic properties of the polypropylene used ~re ~uch ~hat even at 5 these temperatures, the collapse is not entirely inelastic.

That is, even though the heated platen causes permanent deform2tion of the caps, the material of the caps still retain a significant enough fraction of their room temperature elastic modulus that the minimum thre~hold rorce 10 F is applied to each ~ample tube. Further, the heated platen levels all the caps that it contacts without excessive force regardless of how many tubes are present in the sample block because of the ~oftening of the cap.

Because the cap temperature i~ above the boiling point 15 of water during the entire PCR cycle, the inside surfaces of each cap remain completely dry.

Thus, at the end of a PCR process, if the samples are cooled to room temperature before being removed from the ~ample block, if the caps on each ~ample tube are opened, there i~ no po~sibility of 20 creating an aerosol sprzy of the ~ample tube contents which could result in cross contamination. This is because there is no liguid at the cap to tube seal when the seal is broken.

This i~ extremely advantageous, because tiny particles 25 of aerosol cont~ining amplified product DNA can cont~ inate a laboratory and get into ~ample tubes cont~ ng samples from other ~ources, ~.g., other patient~, thereby po~sibly causing fal~e positive or negative diagno~tic result6 which can be very troublesome. User~ of the PCR amplification 30 process are extremely concerned that no aerosols that can contaminate other samples be created.

A system of disposable plastic items is used to convert the individual ~ample tubes to an 8x12 array which is compatible with microtiter plate format lab e~uipment but 35 which maintains sufficient individual freedom of movement to compensate ~or differences in the various rates of thermal 2~7~3 .

- 67 expansion of the sy~tem components. The relationship of the thermally compliAnt cap to the rest of this system i6 best seen in Figure 21A which is a cross ~ectional ~iew of the sample block, and two sample tubes with caps in place with 5 the sample tubes being held in place by the combination of one embodiment of a plastic 96 well microtiter tray and a retainer. Figure 21B is an alternative, preferred embodiment showing the ~tructure and interaction of most of the various plastic disposable items of the 6ystem. The 10 rectangular plastic 96 well microtiter plate tray 342 rests on the surface of the sample block 12. The top edge 346 of the frame 342 has a height which is approximately 0.5 millimeters shorter than the height of the caps of which cap 364 is exemplary. All of the capped tubes will project 15 higher than the edge 346 of the frame 342. The ~rame 342 is configured ~uch that a downward exten~ing ridge 366 extends into the guardband groove 78 through its entire length. The frame 342 does however have a gap (not shown) which corresponds to the gap in the groove 78 for the temperature 20 sensor ~hown in Figure 2 in plan view and in Figure 7 in cross-~ectional view.

The reference plane 346 mentioned above is established by the top of the frame 342. How this reference plane interacts with the heated platen is as follows. Prior to 25 screwing down the knob 318 in Figure 20 to line up the index mark~ 332 and 334 to ~tart an amplification run, a calibration process will have been performed to locate the position of the index mark on the e~cutcheon platen 336 in Figure 20. This calibration i6 ~tarted by placing the frame 30 342 in Figure 21 in po~ition on thc ~ample block. The frame 342 will be empty however or ~ny sample tubes therein will not have any caps in place. Then, the knob 318 is screwed down until the heated platen 14 i~ firmly in contact with the top edge 346 of the frame 342 around its entire 35 parameter. When the knob 318 has been ~crewed down sufficiently to allow the heated platen to rea~t on the ' 2~7~ - 68 reference plane 346 and to press the frame 342 firmly against the top surface 280 of the sample block, the rotatable escutcheon 336 of the preferred r ho~i -nt will be rotated until the index mark 334 on the escutcheon lines up 5 with the index mark 332 on the knob 318. Then, the knob 318 is rotated counterclockwise to raise the platen 14 and the cover 316 in Figure 19 is slid in the negative Y direction to uncover the frame 342 ~nd the sample block 12. Sample tubes with caps loaded with a ~ample mixture ~ay then be 10 placed in position in the frame 342. The heated cover 316 is then placed back over the sample block, and t:he knob 318 is turned clockwise to lower the heated platen 14 until the index mark 332 on the knob lines up with the index mark 334 as previously positioned. This guarantees that all tubes 15 have been firmly seated with the minimum force F applied.

The use of the index marks gives the user a simple, verifiable task to perform.

If there are only a few 6ample tubes in pl~ce, it will take only a small amount of torque to line up the index 20 marks 332 and 334. If there are many tubes, however, it will take more torque on the knob~318 to line up the index marks. This is because each tube is resisting the downward movement of the heated platen 14 as the caps deform.

However, the user is assured that when the index marks 332 25 and 334 are aligned, the heated platen will once ag2in be tightly placed against the top edge 346 of the frame 342 and all tubes will have the i ni ~m threshold force F applied thereto.

This virtually guarantees that the thermal time constant for all the tubes will be substantially the same.

In alternative embodiments, the index marks 332 and 334 may be dispensed with, and the knob 318 may simply be turned clockwise until it will not turn any more.

This condition will occur when the heated platen 314 has reached the top edge or reference plane 346 and the plastic frame 342 has 35 stopped further downward movement of the heated platen 14.

Obviously in this alternative embodiment, and preferably in 2~7~ ~ - 69 the index mark embodiment described above, the plastic of the frame 342 will have a melting temperature which is sufficiently high to prevent deformation of the plastic of the frame 342 when it is in contact with the heated platen 5 14. In the preferred embodiment, the plastic of the frame 342 is celanese nylon 1503 with a wall thickness of 0.05 inches.

An advantage of the above described ~ystem is that sample tubes of different heights may be used ~imply by 10 u~ing frames 342 having different heights. The frame 342 should have a height which is approximately 0.5 millimeters shorter than the plane of the tips of the capped tubes when both are seated in the sample block. In the preferred embodiment, two different tube heights are used. The range 15 of motion of the lead screw 312 which drives the heated platen 14 in Figure 19 must be ~ufficient for all the different sizes of sample tubes to be used.

Of course, during any particular PCR processing cycle, all tubes must be the same height.

The ~ystem described above provides uniform temperatures in the sample block, uniform thermal conductance from block to ~ample, and isolation of the sample tubes from the vagaries of the ambient environment.

Any number of sample tubes up to 96 may be arrayed in the 25 microtiter plate format. The sy~tem allows accurate t~ ~rature control for a very large -- hDr of samples and a visual indication of the sample temperatures for all samples without actually measuring the temperature of any sample.

As the container for PCR reactions, it has been common in the prior Art to use polypropylene tubes which were originally designed for microcentrifuges.

This prior art tube had a cylindrical cross-section closed at the top by a snap-on cap which makes a gas-tight seal.

This prior art 35 tube had a bottom section which comprised the frustrum of a cone with an included angle of approximately 17 degrees.

2~7~ ~ - 70 When such a conical sample tube is pressed down into a sample well of a sample block with a conical cavity with the same included angle, and when the ~ample mixture in the tube lies entirely within the conical volume and below the top 5 surface of the sample block, the thermal conduct~nce between the block and the liguid can be made adeguately predictable for good uniformity of sample temperature throughout the array. To achieve ade~uate control of the thermal conductance between the sample block and the ~ample mixture, 10 the included angles of the conical tube and the sample well must match closely, and the conic~l ~urfaces of the tube and well must be smooth and held together in flush relation.

Further, the minimum threshold force F must be applied to each sample tube to press each tube tightly into the sample 15 well ~o that it does not rise up or loosen in the well for any reason during thermal cycling, such as steam formation from trapped liquid in space 291 in Figure 15. Finally, each tube must be loaded with the same amount of ~ample liguid. If the above listed conditions are met, the thermal 20 conductance between the sample block and the ~ample liquid in each tube will be predc inA~tly determined by the cond~ctance of the conical plastic wall 368 in Figure 15 and a boundary layer, ~not ~hown) of the sample liquid at the inside surface 370 of the conical 6ample tube wall.

The thermal conductance of the plastic tube walls is determined by their thickness, which cAn be closely controlled by the injection molding method of manufacture of the tubes.

The sample liguid in all the sample tubes has virtually identical thermal properties.

It has been found by experiment and by calculation that a molded, one-piece, 96-well ~icrotiter plate is only marginally feasible for PCR because the differ~nces in the thermal ~Yr~n~ion coefficients between aluminum and plastic lead to dimensional changes which can destroy the uniformity 35 of thermal conductance to the sample liguid across the array. That is, since each well in such a one piece plate ~ 2~7~3 - 71 is connected to each other well through the surface of the plate, the distances between the wells are determined at the time of initial manufacture of the plate but change with changing temperature since the plastic of the plate has a 5 significant coefficient of thermal expansion. Also, distances between the sample wells in the metal sample block 12 are dependent upon the temperature of the sample block since aluminum also has a significant coefficient of thermal expansion which is different than that of plastic. To have 10 good thermal conductance, each ~ample well in a one-piece 96-well microtiter plate would have to fit almost perfectly in the corresponding well in the sample block ~t all temperatures. since the temperature of the ~ample block changes over a very wide range of temperatures, the 15 distances between the sample wells in the sample block vary cyclically during the PCR cycle. Because the coefficients of thermal exp~nsion for plastic and aluminum are substantially different, the distances of the well separation in the ~ample block would vary differently over 20 changing temperatures than would the di~tances between the sample wells of a plastic, one-piece, 96-well microtiter plate.

Thus, as an important criteria for a perfect fit between a sample tube and the corresponding sample well over 25 the PCR temperature range, it is nec~ss~ry that each sample tube in the 96-well array be individually free to move laterally and each tube must be individually free to be pressed down vertically by whatever ~mount is rlecessary to make flush contact with the wall~ of the ~ample well.

The sa~ple tubes used in the invention are different from the prior art microcentrifuge tubes in that the wall thickness of the conical frustrum position of the sample tube is much thinner to allow fnster heat transfer to and from the sa~ple liquid.

The upper part of the~e tubes has 35 a thicker wall thickness than the conical part. In Figure 15, the wall thickness in the cylindrical part 2~8 in Figure 2~7~3 - 72 l5 is ~enerally 0.030 inches while the wall thickness for the conical wall 368 is 0.009 inches. Because thin pArts cool faster than thick parts in the injection molding process, it is important to get the mold full before the s thin parts cool off.

The material of the sample tubes must be compatible chemically with the PCR reaction. Glass is not a PCR compatible material, because DNA sticks to gla~s and will not come off which would interfere with PCR ~mplification.

10 Preferably an autoclavable polypropylene i8 ufied. Three types of suitable polypropylene were identified earlier herein. Some plastics are not compatible with the PCR process because of outgassing of materials from the plastic or because DNA sticks to the plastic walls. Polypropylene 15 i5 the best known clas~ of plastics at this time.

Conventional injection ~olding t~chnique~ and mold manufacture techniques for the injection mold will suffice for purposes of practicing the invention.

The use of cone shaped sample tubes translates 20 substantially all manufacturing tolerance errors to height errors, i.e., a variance from tube to tube in the height of the tip of the cap to the top of the sample block when the ~ample tube i5 ~eated in the sample well. For example, an angle error for the angle of the ~ample tube walls is 25 converted to a height error when the tube is placed in the sample block because of the mi~match between the tube wall angle and the sample well wAll angle. Likewise, a diameter error in the ~i ~n~ions of the cone would al~o translate into a height error since the conical part of the tube would 30 either penetrate deeper or not as much as a properly dimensional tube.

For good uniformity of thermal conductance across the array, a good fit between the sample tubes and the sample well must exist for all 96-wells over the full temperature 35 range of 0 to 100~C regardless of differences in thermal expansion rates. Also, each of the 56 sample tubes must 2~67~3 - 73 have walls with dimensions and wall thicknesses which are uniform to a very high degree. Each sample tube in which sample mixture is to be held should be fitted with a removable gas-tight cap that makes a gas-tight seal to 5 prevent loss of water vapor from the reaction ~ixture when this mixture is at or near its boiling point such that the volume of the sample mixture does not decrease. All these factors combine to make a one-piece microtiter plate with 96 individual fiample wells extremely difficult to manufacture 10 in a manner 50 as to achieve uniform thermal conductance for all 96 wells.

Any structure which provides the necessary individual lateral and vertical degrees of freedom for each sample tube will ~uffice for purposes of practicing the invention.

According to the teachings of the preferre~ embodiment of the invention, all the above noted requirements have been met by using a 4 piece disposable plastic system. This system gives each sample tube sufficient freedom of motion in all necessary directions to c~ ,?nC~te for differing 20 rates of thermal expansion and yet retains up to 96 ~ample tubes in a 96 well microtiter plate format for user convenience and compatibility with other laboratory equipment which is sized to work with the industry standard 96-well microtiter plate. The multi-piece disposable 25 plastic system is very tolerant of manufacturing tolerance error~ and the differing thermal expansion rates over the wide t. ,erature range encountered during PCR thermal cyclinq.

Figures 21A and 21~ show alternative embc~i ~nts of 30 most of the four piece plastic ~ystem ~ nts in cross ~ection as assembled to hold a plurality of sample tubes in their sample wells with sufficient freedom of motion to account for differing rates of thermal expansion.

Figure 45 shows all the parts of the disposable plastic microtiter 35 plate emulation system in an exploded view. This figure illustrates how the parts fit together to form a microtiter 2~67~3 plate with all the sample tubes loosely retained in an 8x12 microtiter plate format 96 well array. Figure 22 ~hows a plan view of a microtiter plate frame 342 according to the teachings of the invention which i~ partially shown in 5 cross-section in Figures 21A and 21B. Figure 23 shows a bottom view plan view of the frame 342. Figure 24 is an end view of the frame 342 taken from view line 24-24' in Figure 22. Figure 25 is an end view of the frame 342 taken from view line 25-25' in Figure 22. Figure 26 i8 a cros~ ~ection 10 through the frame 342 at section line 26-26' in Figure 22.

Figure 27 is a cross sectional view through the frame 342 taken alony section line 27-27' in Figure 22. Figure 28 is a side view of the frame 342 taken along view line 28-28' in Figure 22 with a partial cut away to show in more detail the 15 location where a retainer to be described below clips to the frame 342.

Referring jointly to Figures 21A, 21B and 22 through 28, the frame 342 i5 comprised of a horizontal plastic plate 372 in which there are formed 96 holes spaced on 9 20 millimeter centers in the standard microtiter plate format.

There are 8 rows labeled A through ,H and 12 colu~ns labeled 1 through 12. Hole 374 at row D, column 7 i8 typical of these holes. In each hole in the frame 342 there is placed a conical sample tube such as the sample tube 376 shown in 25 Figure 15. Each sample tube is smaller in diameter than the hole in which it is placed by about 0.7 millimeters, ~o that there i~ a loose fit in the hole. This i8 best seen in Figures 21A and 2lB by ob6erving the distance betwQen the inside edge 378 of a typical hole and the side wall 380 of 30 the ~ample tube placed therein. Reference numeral 382 in Figures 21A and 21B ~hows the opposite edge of the hole which is also spaced away from the outside wall of the cylindrical portion of the ~ample tube 376.

Each sample tube has a ~houlder shown at 384 in Figures 35 15, 2lA and 2lB. This shoulder i5 molded around the entire circumference of the cylindrical portion 288 of each sample 2 ~ - 75 tube.

The diameter of this shoulder 384 is large enough that it will not pass through the holes in the frame 342, yet not ~o large as to touch the ~houlders of the adjacent tubes in neighboring holes.

once all the tube~ are placed in their holes in the frame 342, a plastic retainer 386 (best seen in Figures 21A and 21B and Figure 45) is snapped into apertures in the frame 342.

The purpose of this retainer i~ to keep all the tubes in place such that they cannot fall out or be knocked 10 out of the frame 342 while not interfering ~ith their looseness of fit in the frame 342. The retainer 386 is sized and fitted to the frame 342 such that each sample tube has freedom to move vertically up and down to some extent before the ~houlder 384 of the tube encounters either the 15 retainer 386 or the frame 342. Thus, the fr~me and retainer, when coupled, provide a microtiter plate format for up to 96 sample tubes but provide sufficient horizontal and vertical freedom such that each tube is free to find its best fit at all t- ~ratures under the influence of the 20 minimum threshold force F in Figure 15.

A more clear vi-w of the sample tube and shoulder may be had by reference to Figures 29 and 30. Figures 29 and 30 are an elevation ~ectional view and a partial upper section of the shoulder portion, respectively, of a typical sample 25 tube. A plastic dome-6~pe~ cap such as will be described in more detail below is in~erted into the sample tube shown in Figure 29 and forms a hermetic 6eal with the inside w~ll 390 of the top at the sample tube. A ridge 392 formed in the inside wall of the sample tube acts as a stop for the 30 dome shaped cap to prevent further penetration. Normally, the dome-~haped c~ps come in str~ps connected by web.

Figure 31 shows three caps in elevation view connected by a web 394 and terminated in a tab 396. The tab aids the user in removing an entire row of caps by a single pull.

35 Normally, the web 394 rests on the top surface 398 of the sample tube and prevents further penetration of the cap into 2 ~ 3 - 76 the sample tube.

Each cap includes ~ ridge 400 which forms the hermetic seal between the cap and the inside wall of the sample tube.

Figure 32 shows a top view of three caps in a typical strip of 12 connected caps.

For a more detailed under~tanding of the retainer, refer to Figures 33 through 37.

Figure 33 is a top view of the plastic retainer.

Figure 34 is an elevation view of the retainer taken along view line 34-34' in Figure 33.

Figure 35 is an end elevation view of the retainer taken along view 10 line 35-35' in Figure 33. Figure 36 i~ a 6ectional view taken along section line 36-36' in Figure 33. Figure 37 is a sectional view through the retainer taken along section line 37-37' in Figure 33.

Referring jointly to Figures 33-37, the retainer 386 is 15 comprised of a single horizontal plastic plane 402 surrounded by a vertical wall 404. The plane 402 has an 8 x 12 array of 96 holes formed therein divided into 24 groups of four holes per group. These groups are set off by ridges formed in the plane 402 such as ridges 406 and 408. Each 20 hole, of which hole 410 is typical, has a diameter D which i5 larger than the diameter Dl in Fig. 29 and smaller than the diameter D2. This allows the retainer to be ~lipped over the ~ample tubes after they have been placed in the frame 342 but prevents the sample tubes from falling out of the 25 frame since the shoulder 384 is too large to pass through the hole 410.

The retainer snAps into the frame 342 by means of plastic tabs 414 ~hown in Figures 34 and 36. ~hese plastic tabs are pushed through the ~lot~ 416 and 418 in the frame 30 as shown in Figure 23. There are two plastic tabs 414, one on each long edge of the retA~ner. These two plastic tabs are shown as 414A and 414B in Figure 33.

The frame 342 of Figures 22-28, with up to 96 sample tubes placed therein and with the retainer 386 ~napped into 35 place, forms a single unit such as is shown in Figures 2lA and 21B which can be placed in the sample block 12 for PCR 2~5~7~3 - 77 processing.

After processing, all the tubes may be removed simultaneou~ly by lifting the frame 342 out of the ample block. For convenience and ~torage, the frame 342 with S sample tubes and retainer in place can be inserted into another plastic eomponent called the base. The base has the outside dimensions and footprint of a st~n~rd 96-well microtiter plate and i8 shown in Figures 38 ~hrough 44.

Figure 38 is a top plan view of the ba~e 420, while Figure 10 39 is a bottom plan view of the base. Figure 40 i8 an elevation view of the base taken from view line 40-40' in Figure 38. Figure 41 is an end elevation view taken from view line 41-41' in Figure 38. Figure 42 is a ~ectional view taken through the base along section line 42-42' $n 15 Figure 38. Figure 43 is a sectional view through the base taken along ~ection line 43-43' in Figure 38. Figure 44 is a sectional view taken along section line 44-44' in Figure 38.

The base 420 includes a flat plane 422 of plastic in 20 which an 8 x 12 array of holes with ~loped edges is formed.

These holes have dimension6 and ~pàcing such that when the frame 342 i~ ~eated in the base, the bottoms of the sample tubes fit into the conical holes in the ba~e such that the sample tubes are held in the same relationship to the frame 25 342 a6 the ~ample tubes are held when the frame 342 i5 mounted on the ~ample block. Hole 424 is typical of the 96 holes formed in the base and i~ shown in Figures 38, 44 and 43. The individual ~ample tubes, though loo~ely captured between the tray and retainer, become firmly ~eated and 30 immobile when the frame is inserted in the base. The -nner in which a typical cample tube 424 fits in the ba~e i~ ~hown in Figure 44.

In other words, when the frame, sample tubes and retainer are ~eated in the base 420 the entire assembly 35 becomes the exact functional equivalent of an industry standard 96-well microtiter plate, and can be placed in 2~743 .

- 78 virtually any automated pipetting or sampling system for 96well industry standard microtiter plates for further processing.

~fter the sample tubes have been filled with the S necessary reagents and DNA sample to be amplified, the sample tubes can be capped. In an alternative embodiment of the cap strip shwon in Figures 31 and 32, an entire mat of 96 caps with a compliant web connecting them in an 8 x 12 array may be used. This web, shown at 394 in Figure 31 must 10 be sufficiently compliant 80 that the caps do not L~s~rain the sample tubes from making the small motions these sample tubes must make to fit perfectly in the conical wells of the sample block at all temperatures.

The as~embly of tubes, caps frames, retainer and base 15 is brought after ~illing the tubes to the ther~al cycler.

There, the Srame, capped tubes and retainer plate are removed from the base as A unit. This unit is then placed in the sample block 12 to make the ~ss hly shown in Figure 21A or 21B with the tubes loosely held in the conical wells 20 in the sample block. As shown in Figure 21, the frame 342 i5 ceated on the top surface 280 o~ the guardband. In the preferred ~ l~o~i e~t, the ridge 366 extends down into the groove 78 of the guardband, but this is not essential.

Next, the heated cover is slid over the s~mples, and 25 the heated platen is screwed down as previously described until it contacts the top edge 346 of the frame 342.

Within ~econ~ after the heated platen 14 in Figure 19 touches the caps, the caps begin to soften and yield under the downward pres~ure from the lead screw 312 in Figure 19.

30 The user then continues to turn to knob 318 until the index marks 332 and 334 in Figure 20 line up which indicates that every sample tube has been tightly pressed into the sample block with at least the ninimum threshold force F and all air gaps between the heated platen 14, the sample block and 3s the top edge 346 of the frame 342 have been tightly closed.

The sample tubes are now in a completely closed and 2~7~ ,9 controlled environment, and precision cycling of temperature can begin.

At the end of the PCR protocol, the heated platen 14 is moved upward and away from the sample tubes, and the heated 5 cover 316 is slid out of the way to expose the frame 342 and sample tubes. The frame, ~ample tubes and retainer are then remo~ed and replaced into an empty base, and the caps can be removed. As each cap or Gtring of caps is pulled off, the retainer keeps the tube from coming out of the tray. Ribs 10 formed in the base (not shown in Figures 38-44) contact the retainer tabs 414A and 414B shown in Figure 33 to keep the retainer ~napped in place ~uch that the force exerted on the tubes by removing the caps does not dislodge the retainer 386.

Obviously, the frame 342 may be used with fewer than 96 tubes if desired. Also, the retainer 386 can be removed if desired by unsnapping it.

A user who wiches to run only a few tubes at a time and handle these tubes individually can pl_ce an empty frame 342 20 without retainer on the ~ample block. The user may then use the base a~ a "test tube rack~ and ~et up a ~mall number of tubes therein. These tubes can then be filled ~anually and capped with individual caps. The user may then transfer the tubes individually into wells in the ~ample block, close the 25 heated cover and 6crew down the heated platen 14 until the mark~ line up. PCR cycling may then commence.

When the cycling i5 complete, the cover 316 i8 ~ ed and the sample tubec are ~ndividually placed in an available base.

The retainer is not nece6s-ry in thi~ type of u~age.

Referring to Figur-s 47A and 47~ (hereafter Figure 47), there i~ shown a block diAgram for the electronics of a preferred embodiment of a control ~ystem in a class of control ~ystems represented by CPU block 10 in ~igure 1.

The purpose of the control electronics of Figure 47 is, 35 inter alia, to receive and 6tore user input dat_ defining the desired PCR protocol, read the various t~- srature ~ 2~7~3 - 80 sensors, calculate the sample temperature, compare the calculated sample t~ ,~,rature to the desired temperature as defined by the user defined PCR protocol, monitor the power line voltage and control the film heater zones and the ramp 5 cooling valves to carry out the desired temperature profile of the user defined PCR protocol.

A microprocessor ~hereafter CPU) 450 executes the control program described below and given in Appen~ iY C in source code form. In the preferred embodiment, the CPU 450 10 is an OKI CMOS 8085. The CPU drives an addre~s bua 452 by which various ones of the other circuit element~ in Figure 47 are addressed. The CPU also drives a data bus 454 by which data i~ transmitted to various of the other circuit elements in Figure 47.

The control program of Appendix C and ~ome system constants are stored in EPROM 456. User entered data and other system constants and characteristics measured during the in~tall process (install program execution described below) are stored in battery backed up RAM 458. A system 20 clock/calendar 460 supplies the CPU 450 with date and time information for purpoees of recording a history of events durinq PCR runs and the duration of power failures as described below in the description of the control ~oftware.

~n address decoder 462 receives and decodes addresses 25 from the address bus 452 and activates the appropriate chip select lines on a chip ~elect bus 464.

The user enters PCR protocol data via a keyboard 466 in response to information di~played ~y CPU on d$splay 468.

The two way - ~.ication between the user and the CPU 450 30 is described in more detail below in the user interface section of the description of the control software. A keybo~rd interface circuit 470 converts u~er keystrokes to data which i~ read by the CPU via the data ~us 454.

Two programmable interval timers 472 and 474 each 35 contain counters which are loaded with counts calculated by the CPU 450 to control the intervals during which power is ~ 2~7~ - 81 applied to ~he various film heater zones.

An interrupt controller 476 ~ends interrupt requests to the CPU 450 every 200 mill;~cond~ causing the CPU 450 to run the PID task described below in the description of the 5 control ~oftware. This task reads the temperature sensors and calculates the heating or cooling power necessary to move the sample temperature from its current level to the level desired by the user for that point in time in the PCR protocol being executed.

A UART 478 services an RS232 interface circuit 480 such that data stored in the RAM 480 may be output to a printer.

The control ~oftware maintains a record of ~ach PCR run which is performed with respect to the actual t, ,~ratures which existed at various times during the run for purposes 15 of user validation that the PCR protocol actually executed corresponded to the PCR protocol desired by the user. In addition, user entered data defining the specific times and temperatures desired during a particular PCR protocol is also ~tored. All this data and other data ac well may be 20 read by the CPU 450 and output to a printer coupled to the RS232 port via the UART 478. The RS232 interface also allows an external computer to take control of the address and data buses for purposes of testing.

A peripheral interface chip (hereafter PIC) 482 serves 25 as a programmable set of 4 input/output registers. At power-up, the CPU 450 selects the PIC 482 via the address deco~r 462 and the chip ~elect bu~ 464. The CPU then write~ a data word to the PIC via data ~u~ 454 to program the PIC 482 regarding which regicter~ are to be output ports 30 and which ~re to be input ports. Subsequently, the CPU 450 uses the output registers to ctore data words written therein by the CPU via the data bus 454 to control the internal logic ctate of a ~oylammable array logic chip (PAL) 484.

The PAL 484 is a state machine which ha~ a plurality of input signals and a plurality of output signals. PAL's in - 82 - 2~5~3 general contain an array of logic which has a number of different states. Each state i8 defined by the array or vector of logic states at the inputs and each ~tate results in n different array or vector of logic states on the 5 outputs. The CPU 450, PIC 482, PAL 484 and several other circuits to be defined below cooperate to generate different states of the various output ~ignals from the PAL 484.

These different states and associated ou~u~ signals are what control the operation of the electronics shown in 10 Figure 47 a~ will be described below.

A 12 bit analog-to-digital converter (A/D) 486 converts analog voltages on lines 488 and 490 to digital ~ignal~ on data bus 454. These are read by the CPU by generating an address for the A/D converter 6uch that a chip select signal 15 on bus 464 coupled to the chip select input of the A/D converter goes active and activates the converter. The analog ~ign ls on lines 488 and 490 are the output lines of two ~multiplexer~ 492 and 494. Multiplexer 492 has four inputs port~, each having two ~ignal lines. Each o~ these 20 port~ i6 coupled to one of the four temperature sQnsors in the ~ystem. The first port is coupled to the sample block t~ ,~rature sensor. The second ~nd third ports are coupled to the coolant and ambient temperature sen50rs, re~pectively and the fourth port is coupled to the heated cover 25 t~ erature ~en~cr. A typicAl circuit for ~ach one of these t: , rature ~en~or~ i~ shown in Figure 48. A 20,000 ohm resi~tor 49C receives at a node ~97 a regulated +15 volt regulated power ~upply 498 in Figure 47 via a bu~ connection line which i8 not shown. This +lS volts D.C. signal reverse 30 biases a zener diode 500. The reverse bias current and the voltage drop across the zener diode are functions of the temperature.

The voltage drop acros~ the diode i~ input to the multiplexer 292 via lines 502 and 504.

Each t~ _erature sensor has a 6imilar connection to the multiplexer 292.

Multiplexer 494 al80 has 4 input ports but only three are connected.

The first input port is coupled to a , 1~ 2 ~ 3 - 83 calibration voltage generator S06. This voltage generator outputs two precisely controlled voltage levels to the multiplexer inputs and is very thermally ~table. That is, the reference voltage output by voltage ~ource 506 drifts 5 very little if at all with temperature. This voltage is read from time to time by the CPU 450 and compared to a stored constant which represent6 the level thi~ reference voltage had at a known temperature as -a~llrQd during execution of the install proce~ described below. If the 10 reference voltage has drifted from the level ~ea~ured and stored during the install proces~, the CPU 450 knows that the other electronic circuitry u~ed for sen~ing the various temperatures and line voltages has also drifted ~nd adjusts their outputs accordingly to maintain very accurate control 15 over the te~perature measuring process.

The other input to the multiplexer 494 i3 coupled via line 510 to an RMS-to-DC converter circuit 512.

This circuit has ~n input 514 coupled to a step-down transformer 516 and receives an A.C. voltage at input 514 which is 20 proportional to the then existing line voltage at A.C. power input 518.

The ~MS-to-DC converter 512 rectifies the A.C.

voltage and averages it to develop a D.C. voltage on line 510 which also iB proportional to the A.C. input voltage on line 518.

Four optically coupled triac driver~ 530, 532, 534 and 536 receive input control signals via control bu~ 538 from PAL logic 484.

Each of the triac drivers 530, 532 and 534 controls power to one of the three film heater zones.

These heater zones ~re represented by blocks 254, 260/262 and 30 256/258 (the ~ame reference numerals used in Figure 13) The triac driver 536 control~ power to the hQated cover, represented by block 544 via a thermal cut-out ~witch 546.

The heater zones of the film heater ~re protect~d by a block ther~al cutout ~witch 548.

The purpo~e of the thermal 35 cutout swi~ches is to prevent meltdown of the film heater/sample block on the heated cover in case of a failure CA 020~6743 1999-02-04 - 84 leading to the triac drivers being left on for an unsafe interval. If such an event happens, the thermal cut-out switches detect an overly hot condition, and shut down the triacs via signals on lines 552 or 554.

The main heater zone of the film heater is rated at 360 watts while the manifold and edge heater zones are rated at 180 watts and 170 watts respectively. The triac drivers are Motorola MAC 15A10 15 amp triacs. Each heater zone is split into 2 electrically isolated sections each dissipating ~ the power. The 2 halves are connected in parallel for line voltages at 518 less than 150 volts RMS. For line voltages greater than this, the two halves are connected in series. These alternate connections are accomplished through a "personality" plug 550.

The AC power supply for the film heater zones is line 559, and the AC supply for the heated cover is via line 560.

A zero crossing detector 566 provides basic system timing by emitting a pulse on line 568 at each zero crossing of the AC power on line 518. The zero crossing detector is a National LM 311N referenced to analog ground and has 25 mV of hysteresis.

The zero crossing detector takes its input from transformer 516 which outputs A.C. signal from 0 to 5.52 volts for an A.C. input signal of from 0 to 240 volts A.C.

A power transformer 570 supplies A.C. power to the pump 41 that pumps coolant through the ramp and bias cooling channels.

The refrigeration unit 40 also receives its A.C. power from the transformer 570 via another portion of the personality plug 550.

The transformer 550 also supplies power to three regulated power supplies 572, 498 and 574 and one unregulated power supply 576.

For accuracy purposes in measuring the temperatures, the calibration voltage generator 506 uses a series of very precise, thin-film, ultralow temperature drift 20K ohm resistors (not shown in Figure 47).

These same ultralow drift ~ 2~7~3 - 85 resistors are used to set the gain of an analog amplifier 578 which amplifies the output voltage from the ~elected temperature 6ensor prior to conversion to a digital value.

These resistors drift only S ppm/C~.

S All the temperature 6ensors are calibrated by placing them (~eparated from the structures whose temperatures they measure) first in a stable, stirred-oil, temperature controlled bath at 40~C and measuring the actual ou~uL voltages at the inputs to the ~ultiplexer 492.

The 10 temperature sensors are then placed in 2 ~ath at a temperature of 95~C and their GU~y~ voltages are again measured at the ~ame points. The GU~pU~ voltage of the calibration voltage generator 506 i6 also measured at the input of the multiplexer 494. For each temperature, the 15 digital output difference from the A/D converter 486 between each of the temperature sensor outputs ~nd the digital output that results from the voltage generated by the calibration voltage generator S06 is measured. The calibration constants for each te ?r~ture ~ensor to 20 calibrate each for changes in temperature may then be calculated.

The ~ample block temperature ~ensor is then subjected to a further c~l~hration procedure. This procedure involves driving the Qample block to two different temperatures. At 25 each t. ~~ature level, the actual temperature of the block in lG different sample wells is measured using 16 RTD thermocouple probes accurate to within 0.02~C. An average profile for the temperature of the block is then generated and the output of the A/D converter 464 is me~sured with the 30 block t. -rature ~ensor in its place in the ~ample block.

Thi~ is done at both temperature levels. From the actual block te~perature a~ measured by the RTD probes and the A/D output for the block temperature ~ensor, a further calibration factor can be calculated. The temperature 35 calibration factors so generated are stored in battery backed up RAM 458. Once these calibration factors are CA 020~6743 1999-01-1 - 86 determined for the system, it is important that the system not drift appreciably from the electrical characteristics that existed at the time of calibration. It is important therefore that low drift circuits be selected and that ultralow drift resistors be used.

The manner in which the CPU 450 controls the sample block temperature can be best understood by reference to the section below describing the control program. However, to illustrate how the electronic circuitry of Figure 47 cooperates with the control software to carry out a PCR protocol consider the following.

The zero crossing detector 566 has two outputs in output bus 568.

One of these outputs emits a negative going pulse for every positive going transition of the A.C. signal across the zero voltage reference. The other emits a negative pulse upon every negative-going transition of the A.C. signal across the zero reference voltage level. These two pulses, shown typically at 580 define one complete cycle or two half cycles. It is the pulse trains on bus 568 which define the 200 millisecond sample periods. For 60 cycle/sec A.C. as found in the U.S., 200 milliseconds contain 24 half cycles.

A typical sample period is shown in Figure 49.

Each "tick" mark in Figure 49 represents one half cycle. During each 200 msec sample period, the CPU 450 is calculating the amount of heating or cooling power needed to maintain the sample block temperature at a user defined setpoint or incubation temperature or to move the block temperature to a new temperature depending upon where in the PCR protocol time line the particular sample period lies. The amount of power needed in each film heater zone is converted into a number of half cycles each heater zone is to remain off during the next 200 msec sample period. Just before the end of the current sample period in which these calculations are ~ - 87 - 2 ~ ~ ~ 7 ~ 3 made, the cPu 450 nddresses each of the 4 timers in the programmable interval timer (PIT) 472. To each timer, the CPU writes data constituting a "present" count representing the m h~r of half cycles the heater zone associated with 5 that timer is to remain off in the next sample period. In Figure 49, this data is written to the timers during interval 590 just preceding the starting time 592 of the next ~ample period. Assume that a rapid ramp up to the denaturation t~ erature of 94~C is called for ~y the user 10 setpoint data for An interval which includes the 6ample interval between times 592 and 594. Accordingly, the film heaters will be on for most of the period. Assume that the central zone heater is to be on for all but three of the half cycle6 during the sample period. In this case, the CPU 15 450 writes a three into the counter in PIT 472 associated with the central zone heater during interval 590. This write operation automatically causes the timer to issue a "shut off~ ~ignal on the particular control line of bus 592 which controls the central zone heater. This ~shut off" 20 signal cause~ the PAL 484 to issue a "shut off" signal on the particular one of the signal lines in bu~ 538 a~ociated with the central zone. The triac driver 530 then shuts off at the next zero crossing, i.e., at time 592. The PIT receives a pulse train of positive-going pulse6 on line 594 25 from the PAL 484. These pulses are translation6 of the zero~crossing pulses on 2-line bus 568 by PAL 484 into positive goinq pul6es at all zero cros6ing pulses on 2-line bu~ 568 by PAL 484 into po6itive going p~l Pefi at all zero crossings on a ~ingle line, i.e., line 594. The timer in 30 PIT 472 associated with the central film heater zone start6 counting down from $ts present count of 3 using the half cycle marking pulses on line 594 as it6 clock. At the end of the third half cycle, this timer reaches 0 and causes its output signal line on bus 592 to change states. This 35 transition from the off to on ~tate is ~hown at 596 in Figure 49. This transition is communicated to PAL 484 and ~ 2 ~ 4 ~ - 88 causes it to change the state of the appropriate output signal on bus 538 to switch the triac driver 530 on at the third zero-crossing. Note that by switching the triacs on at the zero crossings as is done in the preferred 5 embodiment, switching off of a high current flowing through an inductor (the film heater conductor) is avoided. This minimizes the generation of radio freguency interference or other noise. Note that the technique of swit~h ~ ng a portion of ç~ç~ half cycle to the film heater in accordance with the 10 calculated amount of power needed will also ~ork as an alternative embodiment, but is not preferred because of the noise generated by this te~hn;gue.

The other timers of PIT 472 and 474 work in a similar manner to manage the power applied to the other heater zones 15 and to the heated cover in accordance with power calculated by the CPU.

Ramp cooling is controlled by CPU 450 directly through the peripheral interface 482. When the heating/cooling power calculations performed during each sample period 20 indicate that r~mp cooling power is needed, the CPU 450 addre~ses the peripheral interface controller (PXC) 482. A data ~ord is then written into the appropriate register to drive output line 600 high. This output line triggers a pair of monostable multivibratorfi 602 and 604 and causes 25 each to emit a single pulse, on lines 606 and 608, respectively. These pulses each have peak currents just under ~ ampere and a pulse duration of approximately 100 milli~econds. The purpose of these pul~es i~ to drive the solenoid valve coils that control flow through the ramp 30 cooling channels very hard to turn on ramp cooling flow quickly. The pulse on line 606 causes a driver 610 to ground a line 612 coupled to one side of the solenoid coil 614 of one of the solenoid operated valves. The other terminal of the coil 614 is coupled to a power supply "rail" 35 616 at +24 volts DC from power supply 576. The one shot 602 controls the ramp cooling solenoid operted valve for flow in CA 020~6743 l999-0l-l~ - 89 one direction, and the one shot 604 controls the solenoid operated valve for flow in the opposite direction.

Simultaneously, the activation of the RCOOL signal on line 600 causes a driver 618 to be activated. This driver grounds the line 612 through a current limiting resistor 620.

The value of this current limiting resistor is such that the current flowing through line 622 iS at least equal to the hold current necessary to keep the solenoid valve 614 open. Solenoid coils have transient characteristics that require large currents to turn on a solenoid operated valve but substantially less current to keep the valve open. When the 100 msec pulse on line 606 subsides, the driver 612 ceases directly grounding the line 612 leaving only the ground connection through the resistor 620 and driver 618 for holding current.

The solenoid valve 614 controls the flow of ramp cooling coolant through the sample block in only ~ the ramp cooling tubes, i.e., the tubes carrying the coolant in one direction through the sample block. Another solenoid operated valve 624 controls the coolant flow of coolant through the sample block in the opposite direction. This valve 624 iS driven in exactly the same way as solenoid operated valve 614 by drivers 626 and 628, one shot 604 and line 608.

The need for ramp cooling is evaluated once every sample period. When the PID task of the control software determines from measuring the block temperature and comparing it to the desired block temperature that ramp cooling is no longer needed, the RCOOL signal on line 600 is deactivated. This is done by the CPU 450 by addressing the PIC 482 and writing data to it which reverses the state of the appropriate bit in the register in PIC 482 which is coupled to line 600.

.. . . . .. ,.. ~ ...... . . .... ... . ..

CA 020~6743 1999-01-1 - 90 The PIT 474 also has two other timers therein which time a Hz interrupt and a heating LED which gives a visible indication when the sample block is hot and unsafe to touch.

The system also includes a beeper one shot 630 and a beeper 632 to warn the user when an incorrect keystroke has been made.

The programmable interrupt controller 476 is used to detect 7 interrupts; Level 1 - test; Level 2-20 Hz; Level 3 - Transmit Ready; Level 4 - Receive ready; Level 5 - Keyboard interrupt;

Level 6 - Main heater turn on; and, Level 7 - A.C. line zero cross.

The peripheral interface controller 482 has four outputs (not shown) for controlling the multiplexers 492 and 494. These signals MUX1 EN and MUX2 EN enable one or the other of the two multiplexers 492 and 494 while the signals MUX 0 and MUX 1 control which channel is selected for input to the amplifier 578.

these signals are managed so that only one channel from the two multiplexers can be selected at any one time.

An RLTRIG* signal resets a timeout one shot 632 for the heaters which disables the heaters via activation of the signal TIMEOUT EN* to the PAL 484 if the CPU crashes. That is, the one shot 632 has a predetermined interval which it will wait after each trigger before it activates the signal TIMEOUT EN* which disables all the heater zones. The CPU 450 executes a routine periodically which addresses the PIC 482 and writes data to the appropriate register to cause activation of a signal on line 634 to trigger the one shot 632. If the CPU 450 "crashes~' for any reason and does not execute this routine, the timeout one-shot 632 disables all the heater zones.

The PIC 482 also has outputs COVHTR EN* and BLKHTREN* (not shown) for enabling the heated cover and the sample block heater.

Both of these signals are active low and are , ~ 91- 2~7~ controlled by the CPU 450.

They are output to the PAL 484 via bus 636.

~he PIC 482 also outputs the signals BEEP and BEEPCLR* on bus 640 to control the beeper one shot 630.

The PIC 482 also outputs a signal MEM1 ~not shown) which is used to switch pages between the high address section of EPROM 456 and the low address section of battery RAM 458.

Two other signals PAGE S~L O and PAGE SEL 1 ~not shown) are output to select between four 16K pages in EPROM 10 456.

The four temperature sensors are National LM 135 zener diode type sensors with a zener voltage/temperature dependence of lO mV/~K. The zener diodes are driven from the regulated power supply 498 through the 20K resistor 496.

15 The current through the zener~ varies from approximately 560 ~A to 615 ~A over the 0~C to 100~C operating range. The zener self heating varies from 1.68 mW to 2.10 mW over the same range.

~he multiplexers 492 ~nd 494 are DG409 ~nalog switches.

20 The voltages on lines 488 and 490 are amplified by an AD6Z5KN instrumentation amplifier with 2 transfer function 0~ V~T~ 3*VI~ - 7.5.

The A/D converter 486 is an AD7672 with an input range from 0-5 volts. With the zener temperature ~ensor output from 2.73 to 3.73 volts over the 0~C to 100~C 25 range, the output of the amplifier 578 will be 0.69 volts to 3.69 volts, which is comfortably within the A/D i.nput range.

The key to highly accurate sy6tem performance are good accuracy and low drift with changes in ambient te erature.

Both of these goals are ~chieved by using a precision 30 voltage reference ~ource, i.e., calibration voltage generator S06, and continuously monitoring its output through the ~ame chain of electronics as are used to monitor the outputs of the temperature sen60r6 and the AC line voltage on line 510.

3S The calibration voltage generator 506 outputs two precision voltages on lines 650 and 652.

One voltage is CA 020~6743 l999-Ol-l~ - 92 3.75 volts and the other is 3.125 volts. These voltages are obtained by dividing down a regulated supply voltage using a string of ultralow drift, integrated, thin film resistors with a 0. 05~ match between resistors and a 5 ppm/degree C temperature drift coefficient between resistors. The calibration voltage generator also generates -5 volts for the A/D converter reference voltage and -7.5 volts for the instrumentation amplifier offset.

These two voltages are communicated to the A/D 486 and the amplifier 578 by lines which are not shown. These two negative voltages are generated using the same thin film resistor network and OP 27 GZ op-amps (not shown). The gain setting resistors for the operational amplifier 578 are also the ultralow drift, thinfilm, integrated, matched resistors.

The control firmware, control electronics and the block design are designed such that well-to-well and instrument-toinstrument transportability of PCR protocols is possible.

High throughput laboratories benefit from instruments which are easy to use for a wide spectrum of lab personnel and which require a minimal amount of training. The software for the invention was developed to handle complex PCR thermocycling protocols while remaining easy to program.

In addition, it is provided with safeguards to assure the integrity of samples during power interruptions, and can document the detailed events of each run in safe memory.

After completing power-up self-checks shown in Figures 53 and 54, to assure the operator that the system is operating properly, the user interface of the invention offers a simple, top-level menu, inviting the user to run, create or edit a file, or to access a utility function. No programming skills are required, since pre-existing default files can be quickly edited with customized times and temperatures, then stored in memory for later use. A file protection scheme prevents unauthorized changes to any user's programs. A file normally consists of a set of instructions to hold a desired .... ....

~ - 93 - 2~5~ temperature or to thermocycle. Complex programs are created by linking files together to form a method. A commonly used file, such as a 4~C incubation following a thermocycle, can be stored and then incorporated into methods created by 5 other users. A new type of file, the AUT0 file is a PCR cycling program which allows the user to specify which of several types of changes to control parameters will occur each cycle: time incrementing (auto FC'_ ?nt ex~ension, for yield enhancement), time decrementing, or temperature 10 incrementing or decrementing. For the highest degree of control precision and most reliable methods tran6ferability, temperatures Are setable to 0.1~C, and times are programmed to the nearest second. The invention has the ability to program a 6cheduled PAUSE at one or more setpoints during a 15 run for reagent additions or for removal of tubes at specific cycles.

The system of the invention has the ability to store a 500 record history file for each run. This feature allows the user to review the individual steps in each cycle and to 20 flag any special status or error messages relating to irre gularities. With the optional ~printer, the invention provides hardcopy documentation of file and method parameters, run-time time/t~ erature data with a time/date stamp, configuration parameters, and 60rted file 25 directories.

In order to assure reproducible thermocycling, the computed cam~le temperature is displayed during the ramp and hold segments of each cycle. A temperature one degree lower than the set temperature is normally used to trigger the 30 ramp time and hold-time clocks, but this can be altered by the user. Provided the proper time constant for the type of tube and volume is used, the sample will always approach the desired sample temperature with the ~ame accuracy, regardless of whether long or short sample incubation times 35 have been programmed. Users can program slow ramps for the specialized annealing requirements of degenerate primer CA 020~6743 1999-01-1 - 94 pools, or very short (1-5 sec) high-temperature denaturation periods for very GC rich targets. Intelligent defaults are preprogrammed for 2- and 3-temperature PCR cycles.

Diagnostic tests can be accessed by any users to check the heating and cooling system status, since the software gives Pass/Fail reports. In addition, a system performance program performs a comprehensive subsystem evaluation and generates a summary status report.

The control firmware is comprised of several sections which are listed below:

- Diagnostics - Calibration - Install - Real time operating system - Nine prioritized tasks that manage the system - Start-up sequence - User interface The various sections of the firmware will be described with either textual description, pseudocode or both.

Features of the firmware are:

1. A Control system that manages the average sample block temperature to within +/- 0.1~C as well as maintaining the temperature non-uniformity as between wells in the sample block to within +/- 0.5~C.

2. A temperature control system that measures and compensates for line voltage fluctuations and electronic temperature drift.

... .... . ..... . ... ... ... .. ..

2~67~ 3.

Extensive power up diagnostics that determine if system components are working.

4.

Comprehensive diagnostics in the install program which qualify the heating and cooling systems to insure they are working properly.

5. A logical and organized user interface, employing a menu driven system that allows in~-~, ~nt operation with minimal dependency on the operators manual.

6.

The ability to link up to 17 PCR protocols and store them as a method.

7.

The ability to store up to 150 PCR protocols and methods in the user interface.

8. A history file that records up to 500 events of the previous run as part of the sequence task.

9.

The ability to define the reaction volume and tube size type at the start of a run for temperature accuracy and control as part of the user interface and which modifies tau (the tube time constant) in the PID task.

10.

Upon recovery from a power failure, the system drives the sample block to 4~C to save any samples that may be loaded in the sample compartment.

The analyzer also reports the duration of the power failure as part of the sequence task.

11. The a~ility to print history file contents, "run time" parameters and stored PCR protocol parameters as part of the print task.

2~7~3 -95a 12. The ability to configure to which the apparatus will return during any idle state.

13. The ability to check that the set point temperature is reached with a reasonable amount of time.

14. The ability to control the instrument remotely via an RS232 port.

.

- 96 There are several levels of diagnostics which are described below:

~ series of power-up tests are automatically performed each time the instrument is turned on. They evaluate 5 critical areas of the hardware without user intervention.

Any test that detects a component failure will be run again.

If the test fails twice, an error message is displayed and the keyboard is electronically locked to prevent the user from continuing.

The following areas are tested:

Programmable Peripheral Interface device Battery RAM device Battery RAM checksum EPROM devices Programmable Interface Timer devices Clock / Calendar device Programmable Interrupt Controller device Analog to Digital ~ection Temperature sensors Verify proper configuration plug A Series of service only diagnostics are available to final testers at the manufacturer's location or to field service engineers through a "hidden" keyctrolce ~equence (i.e. unknown to the customer). Many of the tests are the 25 same as the ones in the start up diagnostics with the exception that they can be continually executed up to 99 times.

The following areas are tested:

Programmable Peripheral Interface device Battery RAM device Battery RAM checksum EPROM devices Programmable Interface Timer devices ~ 2~5(~7~3 Clock / Calendar device Programmable Interrupt Controller device ~nalog to Digital section RS-232 ~ection Display section Keyboard Beeper Ramp Cooling Valves Check for EPROM mismatch Firmware version level Battery RAM Checksum and Initialization Autostart Program Flag Clear Calibration Flag Heated Cover heater and control circuitry Edge heater and control circuitry Manifold heater and control circuitry Central heater and control circuitry Sample block thermal cutoff test Heated cover thermal cutoff test User diagnostics are also available to allow the user to perform a quick cool and heat ramp verification test and an extensive confirmation of the heating and cooling system.

These diagnostics also allow the user to view the history file, which is a sequential record of events that occurred 25 in the previous run. The records contain time, temperature, setpoint nu~ber, cycle number, program number and status message~.

Remote Diagnostics are available to allow control of the Gystem ~rom an external computer via the RS-232 port.

30 Control is limited to the service diagnostics and instrument calibration only.

Calibr~tion to determine various parameters ~uch ~s heater resistance, etc. is performed. Access to the calibration screen is limited by a ~hidden" key ~equence 35 (i.e. unknown to the customer). The following parameters ~ 2~7~3 - 98 are calibrated:

The configuration plug is a module that rewires the chiller unit, sample block heaters, coolant pump and power supplies for the proper voltage and frequency (lOOV/50Hz, 5 loO/60Hz, 120/60Hz, 220/SOHz or 230/50Hz). The user enters the type of configuration plug installed. The firmware uses this information to compute the equivalent resistance of the sample block heaters. Upon power-up, the system verifies that the configuration plug selected is consistent with the 10 current line voltage and frequency.

The heater resistance must be determined in the calibration process so that precise calculations of heater power ~elivered can be made. The user enters the actual resistances of the six sample block heaters (two main 15 heaters, two manifold heaters and two edge heaters). The configuration plug physically wires the heater in series for 220-230 VAC and in parallel for 100-120 VAC operation. The firmware computes the equivalent resistance of each of the three heaters by the following formula:

(7) For 100-120 VAC: R~ = (R1 * R2) / Rl + Rz (8) For 220-230 VAC: R~ - R1 + R2 The equivalent resistance is used to deliver a precise amount of heating power to the sample block (Power = Voltage2 x Resistance).

The calibration of the A/D circuit is necessary so that temperatures can be precisely measured.

This is performed by measuring two test point voltages (TP6 and TP7 on the CPU board) and entering the measured voltages.

The output of the A/D at each voltage forms the basis of ~ two point 30 calibration curve.

These voltages are derived from a 5 volt precision source and are ~ccurate and temperature independent. At the start of each run, these voltages are read by the system to measure electronic drift due to ~ 2 ~ 99 temperature because any changes in A/D output is due to temperature dependencies in the analog chain (multiplexer, analog amplifier and A~D converter).

Calibration of the four temperature sensors (sample S block, ambient, coolant and heated cover) is performed for accurate temperature measurements. Prior to installation into an instrument, the ambient, coolant, and heated cover temperature sensors are placed in a water bath where their output ifi recorded (XX.X~C at YYYY mV). These values ~re 10 then entered into the ~ystem. Since temperature accuracy in these areas is not critical, a one point calibration curve is used.

The sample block sensor is calibrated in the instrument. An array of 15 accurate temperature probes is 15 strategically placed in the sample block in the preferred embodiment. The output of the temperature probes is collected and averaged by a computer. The firmware commands the block to go to 40~C. After a brief stabilizing period the user enters the average block temperAture as read by the 20 15 probes. This procedure is repeated at 9S~C, forming a two point calibration curve.

Calibration of the AC to DC line voltage sampling circuit is performed by entering into the system the output of the AC to DC circuit for two given AC input voltages, 25 forming a two point calibration curve. The ou~u~ of the circuit is not linear over the required range (90 - 260 VAC) and therefore requires two points at each end (100 and 120, 220 and 240 VAC), but only uses one set based on the current input voltage.

An accurate measure of AC voltage is necessary to deliver a precise amount of power to the sample block (Power = Voltage2 x Resistance).

The Install program is a diagnostic tool that performs an extensive test of the cooling and heating systems.

35 Install measures or calculates control cooling conductance, ramp cooling conductance at 10~C and 18~C, cooling power at ~ 2~5~7~3 - 100 10~C and 20~C, sample block thermal and coolant capacity and sample block ~ensor lag. The purpose of install i5 three fold:

1. To uncover marginal or faulty components.

2.

To use some of the measured values as system constants stored in battery backed up RAM to optimize the control system for a given instrument.

3.

To measure heating and cooling system ~egradation over time Install is executed once before the system is shipped and fihould also be run before use or whenever a major component i8 replaced. The Install program may also be run by the user under the user diagnostics.

The heater ping test verifies that the heaters are properly configured for the current line voltage (i.e. in parallel for 90-132 VAC and in seriès for 208-26~ VAC). The firmware supplies a burst of power to the sample block and then monitors the ri~e in temperature over A 10 second time 20 period. If the temperature rise is outside a specified ramp rate window, then the heaters ~re incorrectly wired for the current line voltage and the install process is terminated.

The control cooling conductance tests measures the thermal conductance Kcc across the sample block to the 25 control cooling passages. This test is performed by first driving the sample block temperature to 60~C (ramp valves are closed), then integrating the heater power required to maintain the block at 60~C over a 30 second time period.

The integrated power is divided by the sum of the difference 30 between the block and coolant temperature over the interval.

2~67~3 -- 101 - (9) Kcc - ~ Heater Power ~ c / ~ Block - coolant Temp Typical values are 1.40 to 1.55 Watts/~C. A low Kcc may indicate a clogged liner(s). A high R~ may be due to a ramp 5 valve that is not completely closed, leakage of the coolant to the outside diameter of the liner, or a liner that has shifted.

The block thermal capacity (Blk Cp) test measures the thermal capacity of the sample block by first controlling 10 the block at 35~C then applying the maximum power to the heaters for 20 ~econds. The block thermal capacity is egual to the integrated power divided by the difference in block temperature. To increase accuracy, the effect of bias cooling power is subtracted from the integrated power.

(lo) Blk Cp - ramp time * (heater - control cool pwr) / delta temp.

where:

ramp time = 20 seconds heater power = 500 watts control cool ~ (~ block - coolant temp) K~c delta temp = TBlockt,20 - TBlockt.o The typical value of Block Cp is 540 watt-~econds/~C + 30.

Assuming a normal Kcc value, an increase in block ther~al 25 capacity is due to an increase in thermal loads, such as moisture in the foam backing, loss of insulation around the sample block, or a decrease in heater power such as a failure of one of the cix heater zones or a failure of the electronic circuitry that drives the heater zones, or an 30 incorrect or an incorrectly wired voltage configuration module.

0 2~ 7~3 - 102 A chiller test measures the system cooling output in watts at 10~C and 18~C. The system cooling power, or chiller output, at a given temperature is equal to the summation of thermal loads at that temperature~ The main 5 components are: 1. heating power required to maintain the block at a given temperature, 2. power dissipated by the pump used to circulate the coolant around the system, and 3.

losses in the coolant lines to the ambient. The chiller power parameter is measured by controlling the coolant 10 temperature at either 10~C or 18~C and integrating the power applied to the sample block to maintain a constant coolant temperature, over a 32 second interval. The difference between the blocX and coolant temperature is also integrated to compute losses to ambient temperature.

(11) Chiller power - ~ Heating power + Pump power + (Kamb * ~ (blk-cool temp)) where:

heating power - Sum of heating power required to maintain coolant at 10~C or 18~C over time 32 seconds.

Pump Power = Circulating pump, 12 watts ~amb = Conductance to ambient, 20 watts/~C blk-cool temp = Sum of difference in block and coolant temp over time 32 seconds The typical value for chiller power is 230 watts + 40 at 10~C and 370 watts + 30 at 18~C. Low chiller power may be due to an obstruction in the fan path, a defective fan, 30 or a marginal or faulty chiller unit.

It may also be due to a miswired voltage configuration plug.

A ramp cooling conductance (Kc) test measures the thermal conductance at 10~C and 18~C across the sample block 2a~7~3 - 103 to the ramp and control cooling passages. This test is performed by first controlling the coolant te~perature at 10~C or 18~C, then integrating, over a 30 second time interval, the heating power applied to maintain the coolant 5 at the given temperature divided by the difference of block and coolant temperature over the time interval.

(12) Kc = ~ Heating power / ~ (block - coolant temperature) Typical values for Kc are 28 watts/~C + 3 at 10~C and 31 10 watts/~C + 3 at 18~C. A low Kc may be due to a closed or obstructed ramp valve, kinked coolant tubing, weak pump or a hard water/Prestone~ mixture.

A sensor lag test measures the block sensor lag by 15 first controlling the block temperature to 35~C and then applying 500 watts of heater power for 2 seconds and measuring the time required for the block to rise 1~C.

Typical values zre 13 to 16 units, where each unit is equal to 200 ms. A slow or long sensor lag can be due to a poor 20 interface between the sensor and the block, such as lack of thermal grease, a poorly machinP~ sensor cavity or a faulty sensor.

The remaining install tests are currently executed by the install program but have limited diagnostic purposes due 25 to the fact that they ~re calculated values or are a funcl:ion of 50 many variables that their results do not detenmine the source of a problem accurately.

The install program calculates the slope of the ramp cooling conductance (Sc) between 18~C and 10~C. It is a 30 measure of the linearity of the conductance curve. It is also used to approximate the ramp cooling conductance at ooc. Typical values are 0.40 + 0.2. The spread in values attest to the fact that it is just an approximation.

2~5g7~3 - 104 (13) Sc = (Kc_18~ - Kc_10~) / (18~C - 10~C) The install program also calculates the cooling conductance Rco~ R~o is an approximation of the cooling conductance at ooc. The value is extrapolated from the 5 actual conductance at 10~C. Typical values are 23 watts/~C + 5. The formula used is:

(14) Kco = Kc_10 - (Sc * 10~C) The install program also calculates coolant capacity (Cool Cp) which is an approximation of thermal capacity of lo the entire coolant stream (coolant, plumbing lines, heat exchanger, and valves). The cooling c~pacity is equal to components that pump heat into the coolant minus the components that remove heat from the coolant. The mechanics u~ed to measure and calculate these components are complex 15 and ~re described in detail in the source code description section. In this measurement, the coolant is allowed to stabilize at 10~C. M~Y; 1~ heAter power is applied to the sample block for a period of 128 seconds.

(15) Cool Cp = Heat Sources - Coolant sources 20 (16) Cool Cp ~ Heater Power + Pump Power + Kamb * (~Tamb ~Tcool) - Block Cp * (Tblockt,0 - Tblockt~20) - Average Chiller Power between Tcoolt,0 And TC~~lt~l2~ Characters enclosed in { } indicate the variable names used in the source code.

Heater-Pinq Test Pseudocode:

The heater ping test verifies that the heaters are properly wired for the current line voltage.

2 ~ 3 -- 105 - Get the sample block and coolant to a known and stable point.

Turn ON the ramp cooling valves Wait for the block and coolant to go below 5~C Turn OFF ramp cooling valves Measure the cooling effect of control cooling by measuring the block temperature drop over a 10 second time interval. Wait 10 seconds for stabilization before taking any measurements.

Wait 10 seconds templ ~ block temperature Wait 10 seconds temp2 s block temperature {tempa} - temp2 - templ Examine the variable {linevolts} which contains the actual measured line voltage.

Pulse the heater with 75 watts for a line voltage greater then 190V or with 300 watts if it less than 140V.

if ({linevolts} > 190 Volts) then deliver 75 watts to heater else deliver 300 watts to heater Measure the temperature rise over a 10 ~econd time period.

The result is the average heat rate in 0.01 ~/second.

templ - block temperature Wait 10 seconds temp2 = block temperature {tempb} = temp2 - templ ~ 2~6743 - 106 Subtract the average heat rate {tempb} from the control cooling effect to calculate true heating rate (17) heat rate ~ {tempb} - {tempa} Evaluate the heat_rate.

For 220V-230V, the heat rate should be less than 0.30 ~/second.

For lOOV-120V the heat rate should be greater than 0.30 ~/second.

if (linevoltage = 220V and heat rate > 0.30 ~/second) then Error -> Heaters wired for 120 Lock up keyboard if (linevoltage = 120V and heat_rate < 0.30 ~/second) then Error -> Heaters wired for 220 Lock up keyboard 15 KCC Test Pseudocode:

This test measures the control cooling conductance also known as Rcc.

R~c is measured at a block t~ ,erature of 60~C.

Drive block to 60~C Maintain block temperature at 60~C for 300 seconds Integrate the power being applied to the sample block heaters over a 30 second time period. Measure and integrate the power required to maintain the ~lock temperature with control cooling bias.

{dt sum} ~ 0 (delta temperature sum) {main pwr_sum} = O (main heater power sum) {aux pwr_sum} - 0 (auxiliary heater power ~um) for (count = 1 to 30~ CA 020~6743 l999-Ol-l~ - 107 {dt_sum} = {dt_sum} + (block temperature - coolant temperature) wait 1 sec Accumulate the power applied to the main and auxiliary heaters. The actual code resides in the PID control task and is therefore summed every 200ms.

{main_pwr_sum} = {main_pwr_sum} + {actual_power} {aux_pwr_sum) + {aux_pwr_sum} + {auxl_actual} + {aux2_actual} } Compute the conductance by dividing the power sum by the temperature sum. Note that the units are 10 mW/~C.

(18) KCC = ({main_pwr_sum} + {aux_pwr_sum}) / {dt_sum} BLOCK CP Test Pseudocode:

This test measures the sample block thermal capacity.

Drive the block to 35~C Control block temperature at 35~C for 5 seconds and record initial temperature.

initial_temp = block temperature Deliver maximum power to heaters for 20 seconds while summing the difference in block to coolant temperature as well as heater power.

Deliver 500 watts {dt_sum} = 0 ~ 2~7~3 - 108 for (count = 1 to 20 seconds) { {dt sum} = {dt sum} + (block temperature - coolant temperature) wait 1 second } tl9) delta temp ~ block temperature - initial_temp Compute the joules in cooling power due to control cooling which occurs during ramp.

10 (20) cool joule ~ Control cooling conductance (Kc~) * {dt sum} Compute the total joules applied to the block from the main heater and control cooling. Divide by temp change over the interval to compute thermal capacity.

15 (21) Block CP = ramptime * (heater power - cool joule) / delta temp where: ramptime - 20 seconds heater power = 500 Watts COOL_PWR 10:

20 This test measures the chiller power at 10~C.

Control the coolant temperature at 10~C and stabilize for 120 secs.

count = 120 do while (count != O) { if (coolant temperature = 10 + 0.5~C) then -- 109 -count 5 count - 1 else count ~ 120 wait 1 second At this point, the coolant has been at 10~C for 120 seconds and has stabilized. Integrate, over 32 seconds, the power being applied to maintain a cool~nt temperature of 10~C.

{cool init} - coolant temperature {main pwr_sum} - 0 {aux pwr sum} = 0 {delta temp sum} ~ 0 for (count = 1 to 32) { Accumulate the power applied to the main and auxiliary heaters.

The actual code resides in the control task.

{main pwr sum} - {main pwr sum} + actual power {aux pwr sum} ~ {aux pwr sum} + auxl actual + aux2_actual delta temp ~um ~ delta temp sum + (am~ient temp coolant temp) wait 1 second } Compute the ..l ~er of joules of energy added to the coolant mass during the integration interval. "tcoolant temp - cool init)" is the change in coolant temp during the integration interval. 550 is the Cp of the coolant in joules, thus the product is in joules. It represents the extra heat added to the coolant which made it drift -~ 2~5~74~ -- 110 -from setpoint during the integration interval.

This error is subtracted below from the total heat applied before calculating the cooling power.

(22) cool init = (cool~nt temp - cool init) * 550J Add the ~ain power sum to the aux heater sum to get joules dissipated in 32 seconds.

Divide by 32 to get the average joules/sec.

(23) {main pwr sum} - ({main_pwr sum}+{aux pwr sum} cool init) / 32 Compute the chiller power at 10~C by summing all the chiller power components.

(24) Power10.c = main power sum I PUMP PWR ~ ~K AMB * delta temp sum) where: ~ {main pwr sum} = summation of heater power over interval PUMP PWR = 12 Watts, pump that circulates coolant delta temp sum ~ summation of amb - coolant over interval AMB = 20 Watts/K, thermal conductance from cooling to ambient.

RC 10 Test Pseudocode:

This test measures the ramp cooling conductance at 10~C.

Control the coolant temperature at 10~C + 0.5 and allow it to stabilize for 10 seconds.

At this point, the coolant is at setpoint and is being 2 ~ 5 ~ 3 -- 111 -controlled. Integrate, over a 30 second time interval, the power being applied to the heaters to maintain the coolant at 10~C. Sum the difference between the block and coolant temperatures.

{main_pwr_sum} = 0 {aux_pwr_sum} - 0 {dt_sum} ~ 0 for (count ~ 1 to 30) { Accumulate the power applied to the main and auxiliary heaters. The actual code recides in the PID control task.

{main_pwr_sum} = {main_pwr_sum} + actual_power {aux_pwr_sum} ~ {aux_pwr_sum} + auxl_actual + aux2 actual {dt sum} ~ {dt sum} + (block temperature - coolant temp) wait 1 second } Compute the energy in joules delivered to the block over the summation period. Units are in 0.1 watts.

(25) {main pwr_sum} = {main_pwr_sum} + {aux_pwr_sum} Divide the power sum by block - coolant temperature sum to get ramp cooling conductance in 100 mW/K.

25 (26) Xc_10 - {main_pwr_sum} / {dt sum} COO~ PWR 18 Test Pseudocode:

2~7~3 - 112 This test measures the chiller power at 18~C .

Get the sample block and coolant to a known and stable point. Control the coolant temperature at 18~C and stabilize for 128 secs.

count e 128 do while (count != 0) { if (coolant temperature ~ 18~C + 0.5) then count = count - 1 else count = 120 wait 1 second } At this point the coolant has been at 18~C for 120 seconds and has stabilized.

Integrate, over 32 seconds, the power being applied to maintain a coolant temperature of 18~C.

{cool init} = coolant temperature {main pwr sum} - o {aux_pwr sum} - 0 {delta temp_sum} = 0 for (count = 1 to 32) { Accumulate the power applied to the main and auxiliary heaters.

The actual code resides in the control task.

{main_pwr_sum} ~ {main pwr sum} 1 actual power {aux_pwr_sum} ~ {aux pwr_sum} ~ auxl_actual + aux2_actual 2~7~3 - 113 delta temp sum - delta temp ~um + (ambient temp coolant temp) wait 1 second } Compute the number of joules of energy added to the coolant mass during the integration interval. "(coolant temp - cool init)" i5 the change in coolant temp during the integration interval. 550 is the Cp of the coolant in joules, thus the product is in joules.

It represents the extra heat added to the coolant which made it drift setpoint during the integration interval.

This error is subtracted below from the total heat applied before calculating the cooling power.

(27) cool init - (coolant temp - cool init) * 550 Add main power sum to aux heater ~um to get joules dissipated in 32 seconds.

Divide by 32 to get the average joules/sec.

(28) {main pwr sum} = ({main pwr sum}+{aux pwr_sum} cool init) / 32 Compute the chiller power at 18~C by Dl ; ng all the chiller power components.

(29) Power1~.c ~ main power sum + PUMP PWR + (K AMB * delta temp sum) where:

{main pwr sum} ~ summation of heater power over interval PUMP PWR = 12 Watts, pump that circulates coolant delta_temp_sum = summation of amb - coolant over 2~7~3 - 114 interval K AMB = 20 Watts/K, Thermal conductance from cooling to ambient.

5 KÇ 18 Test Pseudocode:

This test measures the ramp cooling conductance at 18~C.

Control the coolant temperature at 18~C + 0.5 and allow it to stabilize for 10 seconds.

At this point, the coolant i6 at setpoint and being lo controlled. Integrate, over a 30 second time interval, the power being applied to the heaters to maintain the coolant at 18~C. Sum the difference between the block and coolant temperature.

{main pwr sum} = o 15 {aux pwr sum} = 0 {dt_sum} = 0 for (count ~ 1 to 30) { Accumulate the power applied to the main and auxiliary heaters.

The actual code resides in the control task.

. .

{main pwr_sum} ~ {main pwr_~um} + actual power {aux pwr sum} ~ {aux pwr sum} + auxl actual + aux2 actual 2S {dt sum} - {dt sum} + (block temperature - coolant temp) wait 1 second } 7 ~ 3 -- 115 -compute the energy in joules delivered to the block over the summation period. Units are in 0.1 watts.

(30) {main_pwr_sum} = {main pwr_sum~ + {aux_pwr sum} Divide power sum by block - coolant temperature sum to get ramp cooling conductance in 100 mW/K.

(31) Kc 18 ~ {main_pwr ~um} / {dt sum} SENLAG Test Pseudocode:

This test measures the sample block ~ensor lag.

Drive the block to 35~C. Hold within + 0.2~C for 20 10 seconds then record temperature of block.

{tempa} = block temperature Deliver 500 watts of power to sample block.

Apply 500 watts of power for the next 2 second6 and count the amount of iterations through the loop for the block temperature to increase 1~C. Each loop iteration executes every 200 ms, therefore actual ~ensor lag is egual to count * 200 ms.

secs ~ 0 count ~ 0 do while (TRUE) { if (secs ~- 2 seconds) then shut heaters off if (block temperature - tempa > 1.0~C) then exit while loop ~. 2 ~ - 116 count = count + 1 } end do while sensor lag = count 5 Coolant CP Test Pseudocode:

This test computes the coolant capacity of the entire system.

Stabilize the coolant temperature at 10~C + 0.5.

Send message to the PID control task to ramp the coolant temperature from its current value (about 10~C) to 18~C.

Wait for the coolant to cross 12~C so that the coolant CP ramp always starts at the same temperature and has clearly started ramping. Note the initial ambient and block temperatures.

do while (coolant temperature < 12~C) { wait 1 second } {blk delta} ~ block temperature {h20 delta} - coolant temperature For the next two minutes, while the coolant temperature i8 ramping to 18~C, sum the coolant temperature and the difference between the ambient and coolant temperatures.

{temp sum} = 0 {cool sum} ~ 0 ~, 2~7~3 - 117 for (count 1 to 128 seconds) { (32) {cool_sum} c cool temp sum I coolant temperature.

5 (33) {temp_sum} ~ ambient - coolant temperature wait 1 second count = count + 1 } Calculate the change in temperatures over the two minute 10 period.

(34) {blk delta} - block temperature - {blk ~delta} (35) {h20_delta} - coolant temperature - {h20 delta} Compute RChill, i.e., the rate of change of chiller power wit,h coolant temperature over the coolant range of 10~C to 20~C. Note that units are in watts/10~C.

(36) Kchill - (Chiller Pwr ~ 18~C - Chiller Pwr ~ 10~C) Compute Sc which is the slope of the ramp cooling conductivity versus the t~mr~rature range of 18~C to 10~C. The units are in watts/~0~C/10~C.

20 (37) Sc ~ (Kc 18 - Kc 10) / 8 r , uLe Kc 0, the ramp cooling conductance extrapolated to 0~C.

(38) Xc O - Kc 10 - (Sc * 10) Compute Cp_Cool, the Cp of the coolant by:

2S (39) Cp_Cool - ( HEATPOWER * 128 + PUMP_PWR * 128 2~5~ - 118 - Power Q 0~C * 128 - Block_Cp * blk delta + K_AMB * temp ~um - Rchill ~ cool_temp_sum 5 h20 delta where:

HEATPOWER ~ 500 W, the heater power applied to warm the block, thus heating the coolant.

It is multiplied by 128, as ~he heating interval was 128 ~ecs.

PUMP PWR ~ 12 W, the power of the pump that circulates the coolant multiplied by 128 seconds.

Pwr_0~C - The chiller power at 0~C multiplied by 128 seconds.

Block_Cp ~ Thermal capacity of sample block.

blk delta ~ Change in block temp over the heating interval.

R_AMB ~ 20 Watts/K, thermal conductance from cooling to a~bient.

temp ~um - The sum once per second of ambient coolant t- ,erature over the interv~l.

h20 delta - Change in coolant temperature over interval of heating (approximately 6~C).

Kchill = Slope of chiller power ver~us coolant 2 ~ -- 119 -temp .

~ool_sum ~ The sum of coolant temp, once per second, over the heating interval.

~5~7~3 P:\~\111l\h~\PWPLl~.~CF F~ru~ry 4, 1991 ~bh) -- 120 - ~F~L TIME OPERATING SYSTEM - CRETIN CRETIN is a stand alone, multitAs~ing kernel that provides system services to other software modules called tasks. Tasks are written in the "C" language with some time 5 critical areas written in Intel 8085 Ass~ hler. Each task has a priority level and provides an independent function.

CRETIN resides in low memory and run6 after the startup diagnostics have successfully been executed.

CRETIN handles the task ccheduling and allows only one 10 task to run at a time. CRETIN receives all hardware interrupts thus enabling waiting tasks to run when the proper interrupt is received. CRETIN provides a real time clock to allow tasks to wait for timed events or pause for known intervals. CRETIN also provides intertask 15 communication through a ~ystem of mes~age nodes.

The fir~ware is composed of nine tasks which are briefly described in priority order below.

Subsequent secticns will describe each task in greater detail.

1.

The control task (PID) is responsible for controlling the sample block t- ,erature.

2.

The keyboard task is responsible for processing keyboard input from the keyp~d.

3.

The timer task waits for a half ~econd hardware interrupt, then sends a wake up message to both the ~equence and the display task.

4. I'he ~eguence task executes the user programs.

5. The pause task handles programmed and keypad pauses ' 2~5~7~ - 121 when a program is running.

6. The display task updates the display in real time.

7. The printer task handles the RS-232 port communication and printing.

5 8. The LED task is responsible for driving the heating LED. It i5 also used to control the coolant temperature while executing Inst211.

9. The link task starts files that are linked together in a method by simulating a keystroke.

~ 2 ~ - 122 ~lock TemPerature Control Program rPID Task) The Proportional Integral Differential (PID) task is responsible for controlling the absolute sample block temperature to 0.1~C, as well as controlling the sample S block temperature non-uniformity (TNU, defined as the temperature of the hottest well minus the temperature of the coldest well) to less than ~ 0.5~C by applying more heating power to the perimeter of the block to compensate for losses through the guard band edges. The PID task is also 10 responsible for controlling the temperature of the heated cover to a less accurate degree. This task runs 5 times per second and has the highest priority.

The amount of heating or cooling power delivered to the 15 sample block is derived from the difference or "error" between the user specified sample temperature stored in memory, called the setpoint, and the current calculated sample temperature. This scheme follows the standard loop control practice. In addition to a power contribution to 20 the film heaters directly proportional to the current error, i.e., the proportional component, (setpoint temperature minus sAmple block temperature), the calculated power also incorporates an integral term that serves to close out any static error (Setpoint temperature - Block temperature less 25 than o 5~C). This component is called the integral component. To avoid integral term accumulation or "windup", contributions to the integral are restricted to a small band around the ~etpoint temperature. The proportional and integral component gains have been carefully selected and 30 tested, ~s the time constants associated with the block sensor and sample tube severely restrict the system's phase margin, thus creating a potential for loop instabilities.

The proportional term gain is P in Equation (46) below and the integral term gain is Ki in Equation (48) below = = = = = ~ 2~67~3 - 123 The PID task uses ~ ~controlled overshoot ~lgorithm" where the block temperature often overshoots its final steady state value in order for the sample temperature to arrive at its desired temperature as rapidly as possible.

5 The use of the overshoot algorithm causes the block temperature to overshoot in a controlled manner but does not cause the sample temperature to overshoot. This ~aves power and is believed to be new in PCR instrumentation.

The total power delivered to all heater of the sample 10 block to achieve a desired ramp rate is given by:

(40) Power = (CP / ramp rate) + bias where:

CP = Thermal mass of block bias = bias or control cooling power ramp_rate = Tfj~, - Tjnjt;., / desired ramp rate This power is clamped to a maximum of 500 watts of heating power for safety.

With every iteration of the task (every 200ms) the system applies heating or ramp cooling power (if necessary) 20 based on the following algorithms.

The control system is driven by the calculated sample temperature. The ~ample temperature is defined as the average temperature of the li~uid in a thin walled plastic sample tube placed in one of the wells of the sample block 25 (herafter the "block"). The time constant of the system (sample tube and its contents) is a function of the tube type and volume. At the start of a run, the user enters the tube type and the amount of reaction volume. The system CA 020~6743 1999-01-1 - 124 computes a resultant time constant (I or tau). For the MicroAmpTM tube and 100 microliters of reaction volume, tau is approximately 9 seconds.

(41) Tblk-new = Tblk + Power * (200ms / CP) (42) Tsamp-new = Tsamp + (TbLk-new - T9amp) * 200 ms / tau where:

Tblknew = Current block temperature Tblk = Block temperature 20Oms ago Power = Power applied to block CP = Thermal mass of block Tsampnew = Current sample temperature Tsamp = Sample temperature 200ms ago Tau = Thermal Time Constant of sample tube, adjusted for sensor lag (approximately 1.5) The error signal or temperature is simply:

(43) error = Setpoint - Tsamp-new As in any closed loop system, a corrective action (heating or cooling power) is applied to close out part of the current error. In Equation (45) below, F is the fraction of the error signal to be closed out in one sample period (200mS).

(44) Tsamp-new = Tsamp + F * (SP - Tsamp) where SP = the user setpoint temperature Due to the large lag in the system (long tube time constant), the fraction F is set low.

Combining formulas (42) and (44) yields:

.. .. . . ...

CA 020~6743 l999-Ol-l~ - 125 (45) Tsamp-new = Tsamp + (Tblk-new-Tsamp) * . 2 / tau = Tsamp + F * (SP-Tsamp) Combining formulas (41) and (45) and adding a term P (the proportional term gain) to limit block temperature oscillations and improve system stability yields:

(46) Pwr = CP * P/T * ((SP - Tsamp) * F * tau/T + Tsamp - Tblk) where P = the proportional term gain and T = the sample period of 0. 2 seconds (200 msec).

and P/T = 1 in the preferred embodiment Equation (46) iS a theoretical equation which gives the power (Pwr) needed to move the block temperature to some desired value without accounting for losses to the ambient through the guardbands, etc.

Once the power needed to drive the block is determined via Equation (46), this power is divided up into the power to be delivered to each of the three heater zones by the areas of these zones. Then the losses to the manifolds are determined and a power term having a magnitude sufficient to compensate for these losses is added to the amount of power to be delivered to the manifold heater zone. Likewise, another power term sufficient to compensate for power lost to the block support pins, the block temperature sensor and the ambient is added to the power to be delivered to the edge heater zones. These additional terms and the division of power by the area of the zones convert Equation (46) to Equations (3), (4) and (5) given above.

Equation (46) iS the formula used by the preferred embodiment of the control system to determine the required , . . .............. ... . ... ... . . . .

CA 020~6743 l999-Ol-l~ - 126 heating or cooling power to the sample block.

When the computed sample temperature is within the "integral band", i.e., _ 0. 5~C around the target temperature (SP), the gain of the proportional term is too small to close out the remaining error. Therefore an integral term is added to the proportional term to close out small errors. The integral term is disabled outside the integral band to prevent a large error signal from accumulating. The algorithm inside the "integral band" is as follows:

(47) Int_sum (new) = Int_sum (old) + (SP - Tsamp) (48) pwr_adj=Ki * Int_sum (new) where, Int_sum = the sum of the sample period of the difference between the SP and Ts~ temperature, and Ki = the integral gain (512) in the preferred embodiment.

Once a heating power has been calculated, the control software distributes the power to the three film heater zones 254, 262 and 256 in Figure 13 based on area in the preferred embodiment. The edge heaters receive additional power based upon the difference between the block temperature and ambient temperature. Similarly, the manifold heaters receive additional power based upon the difference between the block temperature and the coolant temperature.

CA 020~6743 1999-01-1 - 127 PID Pseudocode Upon System Power up or Reset Turn off ramp cooling Turn off all heaters Calculate heater resistances Do Forever - executes every 200 ms If (block temperature ~ 105) then Turn off heaters Turn on ramp valves Display error message Read the line voltage {linevolts} Read the coolant sensor and convert to temperature {h20temp} Read the ambient sensor and convert to temperature {ambtemp} Read the heated cover sensor and convert to temperature {cvrtemp} Read the sample block sensor and convert to temperature {blktemp}. This portion of the code also reads the temperature stable voltage reference and compares the voltage to a reference voltage that was determined during calibration of the instrument.

If there is any discrepancy, the electronics have drifted and the voltage readings from the temperature sensors are adjusted accordingly to obtain .. . . ... . . ....

CA 020~6743 l999-Ol-l~ - 128 accurate temperature readings.

Compute the sample temperature {tubetenths} or the temperature that gets displayed by using a low-pass digital filter.

(49) tubetenths = TTnl + (TBn - TTnl) * T/tau where TTn1 = last sample temp {tubetenths} TBn = current block sensor temp {blktenths} T = sample interval in seconds = 200ms tau = tau tube {cf_tau} - tau sensor {cf-lag} Equation (49) represents the first terms of a Taylor series expansion of the exponential that defines the calculated sample temperature given as Equation (6) above.

Compute the temperature of the foam backing underneath the sample block, {phantenths} known as the phantom mass. The temperature of the phantom mass is used to adjust the power delivered to the block to account for heat flow in and out of the phantom mass. The temperature is computed by using a low pass digital filter implemented in software.

(50) phantenths = TTnl + (TBn - TTn1) * T/tau where TTn-l Last phantom mass temp (phantenths) TBn = Current block sensor temp {blktenths} T = Sample interval in seconds = 2 0 Oms taufOam = Tau of foam block = 30 secs.

Compute the sample temperature error (the difference between the sample temperature and the setpoint ., ~ . ......

~ 3 - 129 temperature) {abs_tube_err}.

Determine ramp direction {fast ramp} ~ UP_RAMP or DN RAMP If (sample temperature is within ERR of cetpoint (SP)) then PID not in fast transition mode. {fast ramp} - OFF where ERR = the temperature width of the "integral band", i.e., the error band ~urrounding the target or setpoint temperature.

Calculate current control cooling power {cool ctrl} to deter~ine how much heat is being lost to the bias cooling channels.

Calculate current ramp cooling power {cool ramp} Calculate {cool brkpt}. {cool brkpt} is a cooling power that is used to detèrmine when to make a transition from ramp to control cooling on downward r~mps.

It i~ a ~unction of block and coolant temperature.

The control cooling power {cool ctrl} and the ramp cooling power {cool ramp} are all factors which the CPU must know to control downward temperature ramps, i.e., to calculate how 25 long to keep the ramp cooling solenoid operated v~lves open.

The control cooling power is egual to a constant plus the temperature Of the cool~nt times the thermal conductance from - the block to the bias cooling channels.

Likewise, the ramp cooling power is equal to the difference between the block 30 temperature and the coolant temperature times the thermal CA 020~6743 1999-01-1 - 130 conductance from the block to the ramp cooling channels.

The cooling breakpoint is equal to a constant times the difference in temperature between the block and the coolant.

Calculate a heating or cooling power (int_pwr) needed to move the block temperature from its current temperature to the desired setpoint (SP) temperature.

(51) (int_pwr) = KP * CP * [(SP - Ts~p) * {cf_kd} + T S ~ TBLK ] where:

KP = Proportional gain = P/T in Equation ( 4 6) = approximately one in the preferred embodiment CP = Thermal mass of block SP = Temperature setpoint Ts~p = Sample temperature TBLK = Block temperature cf_kd = Tau * Kd / Delta_t where tau is the same tau as used in Equation (49) and Kd is a constant and Delta_t is the 200 msec sample period.

If (sample temperature is within {cf_iband} of setpoint) then integrate sample error {i_sum} else (52) clear {i_sum = O}.

Calculate the integral term power.

(53) integral term = {i_sum} * constant {cf_term}.

Add the integral term to the power.

(54) {int_pwr} = {int_pwr} + integral term Adjust power to compensate for heating load due to the effects of the phantom mass (foam backing) by first . , . ~ ., ". ~ .

CA 020~6743 1999-01-1 - 131 finding the phantom mass power then adding it to power (int_pwr).

Calculate phantom mass power ~phant_pwr} by:

(55) phant_pwr = C * (blktenths - phantenths) / 10 where: C = thermal mass of foam backing (1.0 W/K) Adjust heater power {int_pwr} = {int_pwr} + {phant_pwr} Compute power needed in manifold heaters {auxl_power} which will compensate for loss from the sample block into the manifold edges that have coolant flowing through it.

Note that if the system is in a downward ramp, {auxl_power} = O. The manifold zone power required is described below:

(57) {auxl_power} = Kl*( TBLK T~B ) + K2 ( TBLK TCOOL ) + K5*(dT/dt) where:

Kl = Coefficient {cf_lcoeff} K2 = Coefficient {cf_2coeff} K5 = Coefficient {cf_5coeff} dT/dt = Ramp rate TBLK = Block temperature T~B = Ambient temperature TCOOL = Coolant temperature Compute power needed in edge heaters {aux2_power} which will compensate for losses from the edges of the sample block to ambient. Note that if we are in a downward ramp {aux2_power} = O. The edge zone power required is described below:

~i - 132 - 2~ l3 (58) {aux2_power} = K3* (TBLK TAMB) ( BL~ uxL K6* (dT/dt) where:

~3 = Coefficient {cf_3coeff~ K4 = Coefficient {cf_4coeff} K6 - Coefficient {cf_6coeff} dT/dt = Ramp rate T8L~ ~ Block temperature T~p - Ambient temperature 10 Tux~ = Coolant temperature Delete contribution of manifold {auxl_power} ~nd edge heater power {aux2 power} to obtain total power that must _r be supplied by main heaters and coolers.

~5 (59) {int_pwr} = {int power} - {auxl power} {aux2_power} Decide if the ramp cooling should be applied. Note that {cool brkpt} is used as a breakpoint from ramp cooling to control cooling.

If (int_pwr < cool brkpt and performing downward ramp~ to decide whether block temperature is so much higher than the setpoint temperature that ramp cooling is needed then Turn ON ramp valves else Turn OFF ramp valves and depend upon bias cooling At this point, {int_pwr} contains the total heater power and {auxl power} and {aux2 power} contain the loss from the block out to the edges. The power supplied to the auxiliary heaters is composed of two components: aux_power and 30 int_power. The power is distributed {int_pwr~ to the main and auxiliary heaters based on area.

total_pwr = int_pwr 2 ~ - 133 int pwr = total_pwr * 66~ auxl_power = total_pwr * 20%+ auxl_power aux2_power = total_pwr * 14%+ aux2_power Compute the number of half cycles for the triac to conduct S for each end zone and each iteration of the control loop to send the appropriate amount of power to the heaters. ~his loop executes once every 1/5 second, therefore there are 120/5 = 24 half cycles at 60Hz or 100/5 ~ 20 at SOHz. The number of half cycles is a function of requested power 10 {int_pwr}, the current line voltage {linevolt6} and the heater resistance. Since the exact power needed may not be delivered each loop, a remainder is calculated {delta power} to keep track of what to include from the last loop.

~60) int_pwr = int pwr ~ delta power 15 Calculate the number of 1/2 cycles to keep the triac on.

Index is equal to the number of cycles to keep the triac on.

(61) index = power * main heater ohms * t20 or 24] / linevolts squared where Equation (61) is performed once for each heater zone and where Hpower" = int pwr for the main 20 heater zone, auxl_pwr for the manifold heater zone and aux2 pwr for the edge heater zone.

Calculate the amount of actual power delivered.

(62~ actual_power = linevolts squared * index / main heater resistance 25 Calculate the remainder to be added next time.

(63) delta_poWer = int_pwr - actual_power ~ 2~7~3 - 134 Calculate the number of 1/2 cycles for the edge and manifold heaters using the same technique described for the main heaters by substituting {auxl pwr} and {aux2 ~wr} into Eguation (60).

5 Load the calculated counts into the counters that control the main, manifold and edge triacs.

Look at heated cover 6ensor. If heated cover ifi less than 100~C, then load heated cover counter to ~upply 50 watts of power~ 10 Look at sample temperature. If it is greater than SO~C, turn on HOT LED to warn user not to touch block.

END OF FOREVER LOOP ~ 2~7~3 - 135 Keyboard Task The purpose of the keyboard task is to wait for the user to press a key on the keypad, compare the key to a list of valid keystrokes for the current state, execute the command 5 function associated with the valid key and change to a new state. Invalid keystrokes are indicated with a beep and then ignored. This task is the heart of the state driven user interface. It is "state driven" because the action taken depends on the current ~tate of the u6er interface.

10 Reyboard Task Pseudocode:

Initialize keyboard task variables.

Turn off the cursor.

If (install flag not ~et) then Run the in~tall program.

15 Send a message to pid task to turn on the heated cover.

If (the power failed while the user was running a program) then Compute and display the number of minutes the power was off for.

Write a power failure status record to the history file.

Send a message to the sequence ta~k to ~tart ~ 4~C ~oak.

Give the u~er the option of reviewing the history file.

If (the user reguest to review the history fi]e) then Go to the history file display.

25 Display the top level screen.

Do Forever Send a message to the system that this task is waiting for a hardware interrupt from the keypad.

Go to sleep until this interrupt is received.

When awakened, read and decode the key from the keypad.

Get a list of the valid keys for the current state.

Compare the key to the list of valid keys.

If (the key is valid for this state) then Get the "action" and next state information for this ~ 2~67~3 - 136 key.

Execute the "action" (a CG ~n~ function) for this ~tate.

Go to the next tate.

Else Beep the beeper for an invalid key.

End of Forever Loop - 137 ~ 7 ~ 3 Timer Task Overview The purpose of the timer task is to wake up the sequence and the real time display task every half a second. The timer task asks the ~ystem (CRETIN) to wake it up whenever 5 the half second hardware interrupt that is generated by the clock/calendar device is received. The timer task then in turn ~ends 2 wake up messages to the sequence task and the real time display task respectively. This intermediate task is necessary since CRETIN will only service one task per 10 interrupt and thus only the higher priority task (the seguence task) would execute.

Timer Task Pseudocode:

Do Forever Send a message to the system that this task is waiting for a hardware interrupt from the clock/calendar device.

Go to sleep until this interrupt is received.

When awakened, send a message to the sequence and to the real time display task.

End Forever Loop ~ P ~ P~PPL~ jRCF 2 ~ - 138 Seouence Task Overview The purpose of the seguence task is to execute the contents of a user defined program. It sequentially steps through each setpoint in a cycle, consisting of a ramp and 5 a hold segment, and sends out setpoint temperature messages to the pid task which in turn controls the temperature of the &ample block. At the end of each segment, it sends a message to the real time display tAsk to switch the display and a message to the printer task to print the segment's 10 runtime information. The user can pause a running program by pressing the PAUSE key on the keypad then resume the progr~m by pressing the START key. The user can prematurely abort a program by pressing the STOP key. This task executes every half a second when it is awakened by the 15 timer task.

~equence Task Pseudocode:

Do Forever Initialize sequence task variAbles.

Wait for a message from the keyboard task that the user has 20 pressed the START key or ~elected START from the menu or a message from link task that the next program in a method is ready to run.

Go to sleep until this message i~ received.

Whenawakened9update the ADC calibration readings to account 25 for any drift in the Analog circuitry.

If (not starting the 4~C power failure soak sequence) then Send a message to the printer task to print the PE title line, system time and date, program configuration parameters, the program type And its number.

30 If (starting a HOLD program) then Get the temperature to hold at {hold_tp}.

Get the number of seconds to hold for {hold_time}.

If (ramping down more than 3~C and {hold_tp} > 450c) then ~ 2~7~3 - 139 Post an intermediate setpoint.

Else Post the final setpoint {hold_tp}.

While (counting down the hold time {hold time}) Wait for half second wake up message from timer task.

Check block sensor for open or short.

If (keyboard task detected a PAUSE key) then Post a setpoint of current sample temp.

Send a message to wake up the pause task.

Go to sleep until awakened by the pause task.

Post pre-pause setpoint.

If (an intermediate setpoint was posted) then Post the final setpoint.

If (the setpoint temp is below ambient temp and will be there for more than 4 min.) then Set a flag to tell pid task to turn off the heated cover.

Increment the half second hold time counter 20 {store time}.

Post the final setpoint again in case the hold time expired before the intermediate setpoint was reached - this insures the correct setpoint will be written the history file.

Write a data record to the history file.

Send a message to the printer task to print the HOLD info.

End of HOLD program Else if (starting A CYCLE program) then Add up the total number of ~econ~ in a cycle {secs_in_run}, taking into account the ins~rl Ar,t ramp time and the user programmed ramp and hold times.

Get the total number of seconds in the program by multiplying the number of seconds in a cycle by the number - 140 - 2~ of cycles in a program {num_cyc}.

Total {secs in_run} - {secs_in run} per cycle * {num cyc}.

While (counting down the number of cycles {num_cyc}) While (counting down the number of setpoints S {num seg}) Get the ramp time {ramp time}.

Get the final setpoint temp {t final}.

Get the hold time {local_time}.

Send a message to the real time display task to display the ramp segment information.

If (the user programmed a ramp time) then Compute the error {ramp err} between the programmed ramp time ~nd the actual ramp time as ~ollows.

This equation is based on empirical data.

{ramp err} ~ prog ramp rate * 15 + 0.5 (up ramp) {ramp-err} = prog ramp rate * 6 + l.0 (down ramp) where: ~ prog ramp rate G (abs (Tf - Te) - 1) / {ramp time} Tf - setpoint temp {t final} Tc = current block temp {blktemp} abs = absolute value of the expression Note: the '- 1' is there because the clock starts within 1~C of setpoint.

new ramp time ~ old {ramp time} - {ramp err} If (new ramp time > old {ramp time}) then new ramp_time = old {ramp_time}.

Else ~ - 141 - 2~6~ new ramp_time = O.

While (sample temp is not within a user configured temp {cf clk dev} of ~etpoint) Wait for half second wake up message from timer task.

Post a new ramp ~etpoint every second.

Else if (ramping down more than 3~C and {t final} 45~C) then Post an intermediate setpoint.

While (sample temp is not within a user configured temp {cf clk_dev} of setpoint) Wait for half second wake up message from timer task.

Increment the half second ramp time counter.

Check block sensor for open or short.

If (keyboard task detected a PAUSE key) then Post a setpoint of current sample temp.

Send a message to wake up the pause task.

Go to sleep until awakened by the pause task.

Post pre-pause setpoint.

Post the final ~etpoint.

While (sample temp is not within a user configured temp {cf clk dev} of setpoint) Wait for half second wake up message from timer task.

Increment the half second ramp time counter.

Check block sensor for open or short.

2 ~ 3 - 142 If (keyboard task detected a PAUSE key) then Post a setpoint of current sample temp.

Send a message to wake up the pause task.

Go to ~leep until awakened by the pause task.

Post pre-pause setpoint.

Send a message to the printer task to print the ramp information.

Beep beeper to signal end of ramp segment.

lo Send a message to the real time display task to display the ramp segment information.

While (counting down the hold time) Wait for half second wake up message from timer task.

Increment the half second hold time counter.

Check block sensor for open or short.

If (keyboard task detected a PAUSE key) then Post a setpoint of current sample temp.

Send a message to wake up the pause task.

Go to sleep until awakened by the pause task.

Post pre-pause setpoint.

Write a data record to the history file.

Send a message to the printer task to print the hold information.

If (the final setpoint temp has drifted more than the user configurable nmount {cf temp dev}) then Write an error record to the history file.

Check for a programmed pause.

Go to next 6e~r?nt.

Send a message to the printer task to print an end of cycle message.

Go to next cycle.

35 End o~ CYCLE program.

2 ~ 3 - 143 Else if (starting an AUTO-CYCLE program) then Add up the total number of seconds in each program {secs in run} taking into account the instrument ramp time and the user programmed hold times which can be automatically incremented or decremented by a programmed amount each cycle.

While (counting down the number of cycles {num_cyc}) While (counting down the number of setpoints {num ~eg~) Get the final setpoint temp {t final}.

Get the hold time {time hold}.

Check if the user programmed an auto increment or decrement of the setpoint temp and/or the hold time and adjust them accordingly.

If (the auto increment or decrement of the temp causes the setpoint to go below 0~C or a~ove g9.9~C) then An error record is written to the history file.

The setpoint is capped at either OoC or gg.goc.

Send a message to real time display task to display the ramp segment information.

If (ramping down more than 3~C and {t final} >

45~C) then Post an intermediate setpoint.

While (sample temp is not within a user configured temp {cf clk_dev} of setpoint) Wait for half second wake up message from timer task.

Increment the half ~econd ramp time counter.

Check block sensor for open or short.

If (keyboard task detected a PAUSE key) then 2~7~3 - 144 Post a setpoint of current sample temp.

Send a message to wake up the pause task.

Go to sleep until awakened by the pause task.

Post pre-pause setpoint.

Post the final setpoint.

While (sample temp is not within a user configured lo temp {cf_clk_dev} of setpoint) Wait for half second wake up message from timer task.

Increment the half second ramp time counter.

Check block sensor for open or short.

If (keyboard task detected a PAUSE key) then Post a setpoint of current cample temp.

Send a message to wake up the pause task.

Go to sleep until awakened by the pause task.

Post pre-pause ~etpoint.

Send a message to the printer task to print the ramp segment information.

Beep beeper to ~ignal end of ramp portion of segment.

Send a message to the real time display task to display the hold segment information.

While (counting down the hold time) Wait for half second wake up message from timer task.

Increment the half secbnd hold time counter.

Check block sensor for open or ~hort.

If (keyboard task detected a PAUSE key) then Post a setpoint of current sample temp.

Send a message to wake up the pause task.

Go to sleep until awakened by the pause 2~567~3 ~ - 145 task.

Post pre-pause setpoint.

Write a data record to the history file.

Send a message to the printer task to print the hold information.

If (the final setpoint temp has drifted more than the user configurable amount {cf temp dev~) then Write an error record to the history file.

Go to next segment.

Send a message to the printer task to print an end of cycle message.

Go to next cycle.

End of AUTO-CYCLE program.

15 Else if (starting a POWER FAILURE sequence) then Post a setpoint of 4~C.

Set a flag {subamb hold} so that the pid task will shut off the heated cover.

DO FOREVER Wait for a half second wake up message from the timer task.

Increment the half second hold time counter.

END FOREVER LOOP End of power failure sequence 25 Write a run end status record to the history file.

If (running a method) Set a flag {weird flag} so the link task will know to send a message to the sequence task to start the next program running.

30 Else Return user interface to idle state display.

End of Forever Loop ~ 7 ~ 3 - 146 Pause Task Overview The purpose of the pause task is to handle either a pause that the user programs in a CYCLE program or a pause when the u~er presses the PAUSE key on the keypad.

When the sequence task encounters a programmed pause while executing a CYCLE program, it goes to sleep and awakens the pause task. The pause task in turn sends a message to the real time display task to continually display and decrement the time the user asked to pause for. When the pause timer 10 times out, the pause task sends a message to awaken the sequence task and then goes to sleep. The user can prematurely resume the program by pressing the START key on the keypad or can prematurely abort the program by pressing the STOP key.

When the keyboard task detects a PAUSE key while a program is running, it sets a flag {pause flag} then waits for the seguence task to acknowledge it. When the sequence task sees this flag set, it sends an acknowledgment message back to the keyboard task then puts itself to sleep. When the 20 keyboard task receives this message, it awakens the pause task. The pause task sends a mèssage to the real time display task to continually display and increment the amount of time the program is paused for. The timer will time out when it reaches the pause time li~it set by the user in the 25 configuration section. The user can resume the program by pressing the START key on the keypad or abort the program by pressing the STOP key.

Pause Task Pseudocode:

Do Forever Wait for ~ message from the keyboard task indicating a keypad pause, or a message form the sequence task indicating a user programmed pause.

Go to sleep until a message is received.

When awakened, check a flag for the type of pause initiated.

~ 2~7~ - 147 If (it is a programmed pause) then Send a message to the real time display task to display the pause timer counting up.

Else Send a message to the real time display task to display the pause timer counting down.

While (counting down the time out counter) Send a message to the system to suspend this task for half a second.

Send a message to the printer task to print the pause information.

If (it is a programmed pause) then The pause has timed out so send a message to the wake up the sequence task.

Send a message to the real time display task to halt the pause display.

Send a message to the real time display task to resume the running program display.

Else (it is a keypad pause) The pause has timed out and the program must be Aborted so send a message to the system to halt the sequence task And send it back to the top of its FOREVER loop.

If (the program running was a HOLD program) Send a message to the printer task to print the hold information.

Write a ~tatus record to the history file.

Return the user interface to its idle state.

Display an Abort message.

30 End of Forever Loop 2~5~3 - 148 Dis~lay Task Overview The purpose of the real time display task is to display temperatures, timers, sensor readings, ADC channel readings, and other parameters that need to be continually updated 5 every half second.

DisPlaY Task Pseudocode:

Initialize display task variables.

Do Forever Wai~ for a message every half second from the timer task.

Go to sleep until the message is received.

When awakened, check if another task has sent a list of parameters to display or ~ flag to halt the current update.

Toggle the half second flag {half sec}.

If (there's a list of parameters to display) then Set a semaphore so no one else will update the display.

Turn off the cursor.

While (~tepping through the list of parameters) If (it is a time parameter) then Display the time.

If (half second flag {half_sec} is set) then Increment or decrement the time variable.

Else if (it is a decimal number) then Display a decimal number.

Else if (it is an integer number) then Display the integer.

Else if (it is an ADC channel readout) then Read the counts from the ADC channel.

If (need it displayed as mV) then Convert counts to mV.

Display the value.

Else if (it is a power display) then Display the power in terms of watts.

~ - 149 - 2~ Else if (it is the hours left parameter) then Convert seconds to tenths of hourc.

Display the hours left in tenths of hours.

If (half second flag {half_sec} is ~et) then Decrement the seconds variable.

If (the cursor was on) then Turn it back on.

Store the current system time in battery RAM.

Clear the semaphore to release the display.

lo End of Forever Loop - 150 - 2~ 3 Printer Task Overview The purpose of the printer task is to handle the runtime printing. It is a low priority task and should not interfere with other time critical tasks.

S Printer Task Pseudocode:

Do Forever Wait for a message from another task that wishes to print.

Go to sleep until a mess~ge is received.

When awaken, make local copies of the global variables to be printed.

Post a printer acknowledgement message.

If (need to print a status or error message) then Print the information contained in the current history record.

Else if (need to print the page header) then Print the company name, instrument ID, firmware version number and the current system time and date.

Else if (need to print the program header) then Print the type of program and its number.

Else if (need to print the program configuration parameters) then Print the tube type, reaction volume and the sample temperature deviation from setpoint that starts the clock.

Else if (need to print end of cycle information) then Print the ending time and temperature.

Else if (need to print segment information) then Print either the ramp or hold 66', ~nt information.

Else if (need to print a pause ~tatu~ message) then Print the amount of time paused for and at what temp.

End of Forever Loop LED Task Overview The purpose of the LED task is to make the illumination of 2~67~3 - 151 the "Heating" LED reflect the power applied to the main heater. This is a low priority task that runs once a second.

T.~ Task Pseudocode:

5 Initialize LED task variables.

Do Forever Send a message to the system to wake this task every second.

Go to sleep.

When awaken, load counter 2 of PIC timer A with a value that reflects the power applied to the main heater as ~ollows:

load counter with value - {K htled} * {ht led} Where:

{K htled} holds a constant to compute the time to pulse the heating LED and i8 equal to 15200 / 500.

15200 i6 a little greater than the PIC's clock of 14.4KHz and this is the value loaded into the timer to keep the LED constantly on. 500 is the main heater power.

{ht_led} will be a value between 0 and 500 and will be equal to the watts applied to the main heater.

End of Forever Loop Tlink Task overview 25 The purpose of the link task is to simulate the user pressing the START key on the keypad. This task is neces&ary 50 that programs can be executed one right after the other (as in a method) without user intervention. ~he link task wakes up the sequence task and it begins running 2~7~3 - 152 the next program as if the START key were pressed.

T~nk Task Pseudocode:

Initialize link task variables.

Do Forever If ~the flag {weird flag~ i5 set and it is not the first file in the method) then Send a message to the seguence task to wake up.

End of Forever Loop ~ - 153 - 2~ Start Up Sequence ~o~ER-~P 8~Q~ENCE When the power to the instrument is turned on or the software does a RESET, the following ~equence takes place.

5 Note: the numbers below correspond to number~ on the flow chart.

1. Transmit a Ctrl-G (decimal 7) character out the RS 232 printer port. Poll the RS-232 port for at least 1 second and if a Ctrl-G is received, it is assumed that an external computer is attached to the port and all communication during the power-up sequence will be redirected from the keypad to the RS-232 port.

If no Ctrl-G is received, the power-up sequence continues as normal.

15 2. Check if the ~ORE key is depressed.

If so, go straight to the service-only hardware diagnostics.

3. The next 3 tests are an audio/visual check and cannot report an error: l) the beeper beeps 2) the hot, cooling, and heating LEDs on the keypad are flashed 3) each pixel of the display is highlighted.

The copyright and instrument ID screens are displayed as the power-up diagnostics execute.

4.

Should an error occur in one of the power-up diagnostics, the name of the component that failed is displayed and the keypad is locked except for the code 'MORE 999' which will gain access to the service-only hardware diagnostics.

5.

Check channel O of the PPI-B device to ~ee if the automated test bit is pulled low.

If it is, run the UART test.

If the test passes, beep the beeper continuously.

6.

Start the CRETIN operating system which in turn will start up each task by priority level.

~ 2~7~3 - 154 7. Check a flag in battery RAM to cee if the instrument has been calibrated.

If not, display an error message and lock the keypad except for the code 'MORE 999' which will gain accesC to the service-only calibration tests.

8. Run a test that measures the voltage and line frequency and see if both these values match the configuration plug selected while calibrating the instrument.

If not, display an error message and lock the keypad except for the code 'MORE 999' which will gain access to the service-only calibration tests.

9. Perform the heater ping test as described in the Install ~ection.

If the heaters are wired wrong, display an error message and lock the keypad except for the code 'MORE 999' which will gain access to the service-only calibration tests.

1o. Check a flag in battery RAM to see if the instrument has been installed.

If not, display an error message and lock the keypad except for the code 'MORE 999' which will gain access to the install routine.

11. If not in remote mode, check a flag in battery RAM to see if there was a power failure while the instrument was running.

If so, start a 4~C soak and ~isplay the amount of time the power was off for.

Ask the user if they wish to view the history file which will tell them exactly how far along they were in the run when the power went off.

If they select yes, they go straight to the user diagnostics.

30 12. Beep the beeper and clear the remote mode flag 80 all c lnjcation now i8 back through the keypad.

13. Check a flag in battery RAM to see if manufacturing wants their test program ~utomatically started.

If 80, start the program running and reset the instrument after its done.

14. Display the top level user interface screen.

- 2~6~ - 155 Referring to Figure 50, there is shown a crosssectional view of a larger volume, thin walled reaction tube marketed under the trademark M~TAMP.

This tube i~ useful for P~R reactions wherein reagents or other materials need 5 to be added to the reaction mixture which will bring the total volume to greater than 200 microliters. The larger tube shown in Figure 50 made of ~imont PD701 polypropylene or Valtec HH-444 polypropylene and has a thin wall in contact with the sample block. Whatever material is 10 selected should be compatible with the DNA and other components of the PCR reaction mixture 80 as to not impair PCR reaction processing such as by having the target DNA stick to the walls and not replicAte. Glass i8 generally not a good choice because DNA has been known to stick to the 15 walls of glass tubes.

The dimension A in Figure 50 i~ typically 0.012 + .001 inches and the wall angle relative to the longitudinal axis of the tube is typically 17~. The advantage of a 17~ wall angle is that while downward force causes good thermal 20 contact with the sample block, the tubes do not jam in the sample wells. The advantage of thç thin walls is that it ;~; izes the delay between changes in temperature of the sample block and corresponding changes in t. erature of the reaction mixture. This means that if the user wants the 25 reaction mixture to remain within 1~C of 94~C for S seconds in the denaturation segment, and ~G~Lams in these parameters, he or she gets the 5 second denaturation interval with less time lag than with conventional tubes with thicker walls. This performance characteristic of 30 being able to program a short ~oak interval 6uch as a 5 second denaturation soak and get a soak at the programmed temperature for the exact PL G~L ammed time is enabled by use of a calculated ~ample temperature to control the timer. In the system described herein, the timer to time an incubation 35 or ~oak interval is not started until the calculated sample temper~ture reaches the programmed soak temperature.

~ - lS6 - 2~5~ Further, with the thin walled sample tubes, it only takes about one-half to two-thirds as long for the sample mixture to get within 1~C of the target temperature as with prior art thick-walled microcentrifuge tubes and this is 5 true both with the tall MA~TAMP~ tube shown in Figure 50 and the smaller thin walled MICROAMP~ tube shown in Figure 15.

The wall thickness of both the MAXTAMP~ and MI~ROAMP~ tubes is controlled tightly in the manufacturing process to be as thin as possible consistent with adeguate structural 10 strength. Typically, for polypropylene, this will be anywhere from 0.009 to 0.012 inches. If new, ~ore exotic materials which are stroger than polypropylene are uDed to achieve the advantage of speeA;~g up the PCR reaction, the wall thickness cAn be less 80 long as adequate strength is 15 maintained to withstand the downward force to assure good thermal contact, and other stresses of normal u~e. With a height (~;~e~cion B in Figure 50) of 1.12 inches and a ~; -nA-ion C of 0.780 inches and an upper section wall thickness (~; An~ion of D) 0.395 inches, the MA~TAMP tube's 20 time constant is approximately 14 secon~ although this has not been precisely measured as of ~he time of filing. The MICROAMP tube time constant for the shorter tube shown in Figure 15 is typically approximately 9.5 seconds with a tube wall thickness in the conical section of 0.009 inches plus 25 or minus 0.001 inches.

~igure 51 shows the results of use of the thinner walled MICROAMP tube. A similar speeded up att~;- ?nt of target temperatures will result from use of the thin walled M~TAMP tube.

Referring to Figure 51, there is shown a graph of the relative times for the calculated sample temperature in a MICROAMP tube versus the time for a prior ~rt tube to reach a temperature within 1~C of a target denaturation temperature of 94~C from a starting t~ erature of 72~C.

In 35 Figure 51, a 100 microliter sample was present in each tube.

The curve with data points marked by open boxes is the ~ - 157 - 2~67~ calculated sample temperature response for a MICROAMP tube with a 9.5 second response time and a 0.009 inch wall thicknesc. The curve with data points marked by X's represents the calculated sample temperature for a 100 5 microliter sample in a prior art, thick walled microcentrifuge tube with a 0.030 inch wall thickness.

This graph shows that the thin walled MICROAMP tube sample reaches a calculated temperature within 1~C of the 94~C target soak temperature within approximately 36 ~econds 10 while the prior art tubes take about 73 recon~. This is important because in instruments which do not ~tart their timers until the soak temperature i8 substantially achieved, the prior art tubes can ~ubstantially increase overall processing time especially when considered in light of the 15 fact that each PCR cycle will have at least two ramps and 60aks and there are generally very many cycles performed.

Doubling the ramp time for each ramp by using prior art tubes can therefore drastically increase procefising time.

In systems which start ~heir times based upon 20 block/bath/oven t~mr~rature without regard to actual sample t~mp~rature, these long delays~ between changes in block/bath/oven temperature and corresponding changes in sampl~ mixture temperature can have serious negative consequences. The problem is that the long delay can cut 25 into the time that the reaction mixture is actually at the temperature programmed for a soak. For very short soaks as are populAr in the latest PCR proce6ses, the reaction mixture may never actually reach the programmed soak temperature before the heating/cooling sy6tem starts 30 attempting to ch2nge the reaction mixture temperature.

Figure 50 ~hows a polypropylene cap 650 connected to the MAXIAMP sample tube by a plastic web 652. The outside diameter E of the cap and the inside diameter F of the tube upper section are sized for an interference fit of between 35 0.002 and 0.005 inches. The inside surface 654 of the tube should be free of flash, nicks and scratches so that a gas ~ - 158 - 2 ~ ~ ~ 7 4 3 tight seal with the cap can be formed.

Figure 52 shows a plan view of the tube 651, the cap 650 and the web 652. A shoulder 656 prevents the cap from being pushed too far down into the tube and allows 5 sufficient projection of the cap above the top edge of the sample tube for making contact with the heated platen. This also allows sufficient cap deformation such that the minimum acceptable ~orce F in Figure 15 can be applied by deformation of the cap.

In the preferred embodiment, the tube and cap are made of Himont PD701 polypropylene which is autoclavable at temperatures up to 126~C for times up to 15 minutes. This allows the disposable tubes to be ~terilized before use.

Since the caps are permanently deformed in use in machines 15 with heated platens, the tubes are designed for use only once.

caps for the MICROAMP tubes are available in connected strips of 8 or 12 caps with each cap numbered or as individual caps. Single rows of caps may be used and the 20 rows may be easily shortened to as few as desired or individual caps may be cut off the strip. Caps for MAXIAMP tubes are either Attached as shown in Figure 50, or are separate individual caps.

The maximum volume for post-PCR reagent additions to 25 permit iY;ng on the MICROAMP tube is 200 microliters and is up to 500 microliters for the MAXIAMP tube. Temperature limits are -70~C to 126~C.

The response time depends upon the volume of the sample. Response is measured as the time for the sample to 30 come within 37~ of the new temperature when the block suddenly changes temperature. Typical response time for a 50 microliter fill are 7.0 seconds and for a 20 microliter fill are 5.0 seconds.

1 - 159 - 2 ~ ~ ~ 7 ~ 3 APPENDIX A User Interface The objective of the GeneAmp PCR System 9600 user interface is to provide a simple way to develop and run 5 programs that perform PCR.

There are 3 types of ~GylamS available. The ~OLD program consists of a single ~etpoint held for a set ~mount of time or held for an infinite amount of time and terminated by the STOP key. The CYCL~ program adds the features of timed 10 ramps and programmable pauses. This program allows up to 9 setpoints and up to 99 cycles. The AUTO program allows the - user to increment or decrement the setpoint time and/or temperature a fixed amount every cycle. This program also allows up to 9 setpoints and up to 99 cycles. A ~ETHOD 15 program provides a way to link up to 17 hold, cycle or auto programs together.

A total of 150 programs can be stored with nl h~rs ranging from 1 to 150.

Programs can be creàted, stored, protected, printed, or deleted. A directory of the stored programs can 20 be viewed or printed.

~ 2~7~3 - 160 T~E 8Y8TEM 9600 ~EYPAD RUNMORE 7 8 9 Heating BACK STEP 4 5 6 Cooling PAUSE OPTION 1 2 3 Hot STOP CE 0 ENTER 25 RUN starts a program running from the program display or restarts a programmed or keypad pause.

MORE toggles the runtime displays and also accesses the service-only functions (if followed by the code 999 ) ~0 BACK moves to the previous field within the same screen. If currently positioned on the first field, it moves to the previous screen.

STEP moves down to the first field in the next screen.

PAUSE starts a paused time-out for manual interruptions.

~5 OPTION either moves the cursor left-to-right through the menu items (rolling over to the leftmost option) or toggles the YES/NO response.

STOP aborts a running program or moves the user up one level in the user interface.

40 CE clears invalid numeric entries.

ENTER accepts the current numeric entry, accepts a menu item, accepts a YES/NO response, or Ckips to the next field of a display.

If the numeric entry is the last of a display, ENTER steps to the next display.

~ 2~7~3 - 161 CQ~O~ 8Y8TEM 9600 DI8P~AY8 PROGRAX di~play Example:

Prog #~ ~sg Temp CYCL ~17 Done 74.0C Menu ~ UN-STORE-PRINT-HOME Prog is either HOLD, CYCL, AUTO or METH ### is the program ~ (l-lS0) or ??? if it is not stored yet Msg is either Done, Error, Abort or blank Temp is the current sample temperature Menu are the available options R~NTIl~E display Example:

Action TempRamp to 94.0C 29.6C Timer Prog/CyclO:00 Cycle 14 Action is either 'Hold at xx.xC' or 'Ramp to xxOxC' Temp is the current sample temperature Timer counts down the hold or ramp time or counts up a hold time of FOREVER~5 Prog/Cyc for a HOLD file is 'Prog xxx' for a CYCL or AUTO file is 'Cycle xx' - counts up MORE ~isplay Example:

Setpt Tot Cyc Setpt ~3 Tot Cyc 25 Timer Prog Hrs left 2.5 Prog 17 Setpt is the current ~etpoint ~ 9) - counts up Tot Cyc is the total # of cycles (l-99) in the current program Timer is the time left in the program in hrs - counts down Prog is the current program # (l-lS0) ~EYPAD PAU8g ~play Example:

Prog ~#~ Temp AUTO #18 . 55.0C PAUSE Timer PAUSE 9:45 Prog is either HOLD, CYCL, AUTO or METH 2S ### is the program # (l-lSO) or ??? if it is not stored yet Temp is the current sample temperature ~&~ ~3 - 16Z Timer is the configurable pause time - counts down ~ - 163 - 2 ~ ~ ~ 7 ~ 3 TOP LEVEL U8ER INTERFACE Select Option 9600 RUN-CREATE-EDIT-UTIL TOP LEVEL display Run Create program Enter ~oy~am #xxx HOLD-CYCL-AUTO-METH RUN display CREATE display ~dit Select function Enter progrzm ~xxx DIR-CONFIG~DIAG-DEL EDIT display UTIL display Programs are created by selecting a program type in the CREATE display. The user is brought directly to the first display of the program to be edited.

5 Stored programs are retrieved by entering a number 1 to 150 from the RUN, EDIT, or program displays. Entering a valid program number from the RUN display automatically begins the run. Entering a valid program number from the EDIT or program display brings the user to the first displzy of the 10 program to be edited.

Programs are edited by pressing STEP (move down a screen), BACK (move to the previous field) or ENTER (move to the next f ield).

Programs are run by selecting RUN the _UN-STORE-PRINT-HOME menu or by pressing the RUN key on the keypad. The user must first enter 2 parameters required for each run.

The OPTION key toggles the tube 20 Tube type~ ~ICRO type from MICRO (MicroAmp tube) React ~ol. 100uL to THIN (thin-walled GeneAmp tube).

If the user configured a special tube, then the option of OTHER is added. A different reaction volume may be entered.

These parameters are stored with this program. ENTER accepts these values.

If the user configured the 30 Select print mode runtime printer ON and he is OFF-CYCLE-SETPOINT running a cycle, auto or method program, then the following - 164 - 2 ~ ~ ~ 7 ~ 3 printer choices are offered. the program is started. CYCLE prints a message only upon completion of a cycle. SETPOINT prints runtime data for every setpoint (ramp/hold time and temps).

~ - 165 - 2 ~ If the user configured the Select print mode runtime printer ON and he is OFF-ON running a hold program, then the following printer choices are offered.

If the heated cover is below Cover te~p i~ xx~C 100~C, the following ~creen is Run starts at 100~C displayed.

If the u~er i8 on this display when the heated cover reaches 100~C, the run automatically begins. If the user hit STOP to return to the program display, then the run must be manually re-started.

15 Accepting HOME at the RUN-STORE-PRIN$-HOME menu without saving a program displays the screen:

Prog ~xxx not stored Continue? YES _ _ _ _ _ ~ 2~7~3 ~ 166 HOLD PRO~P ~M HOLD ~xxx xx.xC RUN-STORE-PRINT-HOM~ PROGRAM display The user can choose between an ~old at xx.~C infinite soak or A time limited Hold FOREVER-xxx:xx hold.

The beeper will sound once a Beep while Hold? NO second.

UOLD PROGRA~ - Runtim- di~plays Hold at Xx.xC xx.xC None xxx:xx Prog xx RUNTIME display MORE display HOLD ~xx xx.xC None PAUSE xx:xx KEYPAD PAUSE display PROGRAMMED PAUSE ~OLD ppo~ V - Runt~m- printout PE Ce~us GeneAmp PCR System 9600 Ver xx.x Nov 14, 1990 10 xx:xx am Tube type:MICRO Reaction vol:lOOuL Start clock within x.xC of setpt HOLD program #xxx HOLD Program: xx.xC xxx:xx Actual: xx.xC xxx:xx or HOLD Program: xx.xC FOREVER Actual: xx.xC xxx:xx HOLD program #xxx - Run Complete Nov 14, lg90 xx:xx am ~ - 167 - 2~ 3~ CYCLE PROfiP ~M -CYCL ~xxx xx.xC ~UN-STORE-PRINT-HOME PROGRAM display The default is 3. This Temperatu~e PCR determines the number of setpoints in this program. l to 9 setpoints are allowed.

The number of setpoints entered Setpt ~l ~amp x~:xx above determines how many xx.xC ~old xx:xx setpoint edit displays will be offered.

The user can enter a ramp ~nd hold time for each setpoint.

The hold timer will start when the sample temp gets within a user configurable temp of setpoint.

If the user does NOT want to Total cycle~ ~ ~x pause, then the next 3 displays Paus~ during run? NO ar; sk~pped. l to 99 cycles are Entering a 0 for setpoint number Pause a~ter ~etpt ~ also means the user does NOT Beep while pause7YES want to pause therefore the next 2 displays are skipped.

The cycle ,,l h~r is limited to 1st pAuse at cyci Y~ the total number of cycles Pause every xx cycls entered above.

The default pause time is ~et in Pause time x~:xx the user configuration.

2~67~ - 168 CYCLE PROGRAM - Runtim~ displ~ys R~mp to xx.xC xx.xC Setpt ~x Tot Cyc xx xxx:xx Cycle xx Rrs left X.X Progxxx RUNTIME display (ramp) MORE display ~old at xx.xC xx.xC xxx:xx Cycle xx RUNTIME display (hold) CYCL ~xxx xx.xC Setpt #x xx.xC PAUSE XX:XX PAUSE xx:xx Cyc~e xx REYPAD PAUSE display PROGRAMMED PAUSE CYC~g PRO~R~ - Runtim- printout PE Cetus GeneAmp PCR System 9600 Ver xx.x Nov 14, l990 xx:xx am 5 Tube type:MICRO Reaction vol:lOOuL Start clock within x.xC of setpt CYCL program #xxx Cycle #xx Setpt #x RAMP Program: xx.xC xx:xx Actual: xx.xC 1 0 xx: xx HOLD Program: xx.xC xx:xx Actual: xx.xC xx: xx . (up to 9 setpoints) . .

(up to 99 cycles) CYCL program #xxx - Run Complete Nov 14, l990 xx:xx am CYCL program ~xxx - User Aborted Nov 14, l990 xx:xx am (only 20 if aborted) ~ - 169 - 2~ 3 A~O PROG~ AUTO ~xxx xx.xC ~UN-STORE-PRINT-HOME PROGRAM display The default is 3. This Te~perature PCR determines the number of setpoints in this ~- O~r . l to 9 setpoints are allowed.

The number of setpoints entered Setpt ~l xx xC above determines how many ~old for xx xx setpoint edit displays will be offered.

No ramp time is offered thus the instrument ramps as fast as possible.

The hold timer start when the sample temp gets within a user configurable temp of setpoint.

If the user wants to increment Setpt ~l xx~xC or decrement the time and/or Change timeltemp~YES temperature every cycle, then the following display is offered.

The OPTION key toggles the arrow xx.xC delt~ _ x.xC up (increment every cycle) or delt~ xx-xx down (decrement every cycle).

The max time allowed to decrement is limited to the setpoint hold time.

Up to 99 cycles ~re allowed.

Tot~l cyc}e~ ~ x~ . ~.

- 170 - ~5~3 AUTO PROGRAM - Runt~- displays Hold at xx.xC xx.xC Setpt ~x Tot Cyc xx xxx:xx Cycle xx Hrs left X.X ~rogxxx RUNTIME display MORE display AUTO ~xx~c xx.xC ~one PAUSB xx:xx KEYPAD PAUSE display PROGRAMMED PAUSE A~TO ~ROGRAX - Runt~m- printout PE Cetus GeneAmp PCR System 9600 Ver xx.x Nov 14, 1990 xx:xx am 5 Tube type:MICRO Reaction vol:lOOuL Start clock within x.xC of setpt AUTO program #xxx Cycle #xx Setpt #x RAMP Program: xx.xC xx:xx Actual: xx.xC 10 XX:XX HOLD Program: xx.xC, xx:xx Actual: xx.xC xx:xx . (up to 9 setpoints) . (up to 99 cycles) AUTO program ~xxx - Run Complete Nov 12, 1990 xx:xx am AUTO program ~xxx - User Aborted Nov 12, 1990 xx:xx am (only 20 if aborted) ~ - 171 - 2 ~ ~ 6 7 ~ 3 METHOD PROGRAM C METH ~xxx xx.x RUN-STORE-PRINT-HOME PROGRAM display Up to 17 programs can be linked Link ro s~ - - in a method.

If the u~er tries _ P _g ~_ _ _ to enter a non-existant program #, the message "Prog does not exist" is displayed.

If the user tries to link another method, the message "Cannot link a method" is displayed.

MET~OD PROGRAX - Runtimo displays The RUNTIME, MORE and PAUSE displays will be those of the program currently running.

Two additional MORE displays ~re offered when the ~oy~am running is linked in a method.

The number of the program MET~ ~xxx aaa-bbb- currently running will flash.

ccc-ddd-eee-fff-gq~ ADDITIONAL MORE display mmm~nnn-ooo-ppp-qqq ~ ~OD PROGRAM - Runt~m- printout PE Cetus GeneAmp PCR System 9600 Ver xx.x Nov 14, 1990 xx:xx am Tube ~ype:MICRO Reaction vol:lOOuL Start clock within x.xC 20 of setpt METHOD program #xxx - preceeds all linked program data ~ 2~ 7~ - 172 METHOD program #xxx - Meth Complete - follows all linked progra data ~ 2Q~74~ - 173 XET~OD PROGRA~S - Print Select option METHOD-PROGRAM DATA METHOD prints the header of each program linked in the method.

PROGRAM DATA prints the header and contents of each program linked in the method.

~ 2 ~ - 174 8~0RING A PROGRAM When STORE is selected from the RUN-~TORE-PRINT-HOME menu, the routine for storing a program is the same for a file as well ~s a method. Protecting a program insures the user 5 that the program will not be overwritten or deleted without knowledge of the user number. Other users will be able to view, edit, run, and link the protected file in their methods but will not be able to alter the stored version.

xxx is the first available 10 store program number from 1 - 150.

Enter program #xxx The user has entered the # of a Progxxx is protected protected program.

The correct Enter user ~xxxx user # must be entered in order to overwrite this program.

The wrong user # was entered.

Progxxx is protected This display remains for 5 Wrong user number! seconds before reverting to the previous one.

The user is given 3 chances to enter the correct #.

If the user tries to overwrite a Progxxx is linked in program~that is linked in a Methxxx Continue?YES method, the user is warned and given the option of continuing or not.

If the user tries to overwrite a Can't overwrite prog program that is linked in a Linked in method xxx method with another method, an error ~essage lS given.

The user is given the chance to Store protect a ~G~am ~s well as Protect program? NO unprotect a previously protected program.

The user wants to protect the 35 Store program and therefore must enter Enter user ~xxxx a user #.

Ready to store the program in an available slot.

The user ~ appears only if the program is protected.

~ - 175 - 2~ 3 Prog ~xxx User ~xxxx Ready OK to store? YES Prog ~xxx User #xxxx t o OK to overwrite? YES overwr ite an existi n g progra m.

The user # appear s only if the progra m is protec ted.

- - :

- 176 UTILITY ~ud~LION8 Select function DIR-CONFIG-DIAG-~EL UTIL display DIR Allow the user to view or print A directory of the stored programs by either their program number, user number or program type.

5 CONFIG allows the user to tailor the use of the instrument to their specific needs.

DIAG offers the user a means of diagnosing runtime problems and verifying the performance of the instrument.

l0 DEL allows the user to delete stored programs by program number, user number or program type.

~ 2~6~3 - 177 ~TIL - DIRECTORY Directory -PROG-TYPE-USER-PRINT Dir-ctory by PROGram number Programs will be listed in Directory numerical order starting at the 5 Enter program ~xxx given n h~r.

The STEP and BACR keys move through the directory displays.

The beeper ~ounds at the beginning or end of the program list.

lo STOP returns the user to the HOLD ~124 above display.

Dir-ctory by progr~m TYPE The program numbers will be Directory listed for the selected type of 15 HOLD-CYCL-AUTO-METH program.

CYCII ~15 Dir-ctory by U8ER _ ~-r All programs stored under the Directory given user number will be Enter user ~xxxx listed.

~ETH ~150 U~er #1234 20 Directory PRINT ~ 2 ~ - 178 The user can get a hardcopy of Directory Print the directory listing in the PROG-TYPE-USER same manner the directory is viewed above.

~ - 179 - 2 ~ 7~ 3 ~TIL - ~8ER CONFIG~RATION The configuration file can be Conf~guration edited by accepting EDIT from EDIT PRINT the menu or by pressing the STEP key. PRINT prints the contents of this file.

The user can ~et the system time Time~ xx:xx and date.

Datev mm/ddlyy If the runtime printer is ON, 10 Runtime printer OFE the user will be prompted with ~untime beeper ON printer option as the start of each run.

If the runtime beeper is ON, then a beep will sound at the end of each segment (after a ramp or hold portion of a sequence) while running a program.

This time represents the ~imu~ Pause ti ? OU~ limit amount of time a program can 20 xx:xx pause for before it is aborted.

This pertains only to the keypad pause.

This time represents the number Allowed setpt arror of degrees the actual sample 25 x.x~C temp may vary from the setpoint before an error is flagged.

This setpoint is useful for Idle state ~etpoint balancing the control cooling xx~C power which is always present.

The ~ample temp will be maintained at the idle state ~etpoint whenever the instrument is idle.

The clock which times the hold 35 Start clock within segment of a running program can x.x~C of setpoint be configured to be triggered when it gets within this temperature of the sample temp.

The nominal v~lue is 1.0~C.

If the user wishes to use a different type of tube other than the MicroAmp or Thin-walled GeneAmp tubes, they must set this option to YES and enter at least 3 pairs of - 180 - 2~ 3 reaction volume and tube time Special tube? NO constant data.

This curve will be used to extrapolate the correct Tau (tube time constant) for each run using this special tube depending on the reaction volume entered by the user at the tart of a run.

- 181 VTI~ - U8ER CONFIG~RATION ~cont) 2 ~ 5 6 ~ ~ 3 3 sets of this screen will be Rxn vol=xxxuL T-xxxs offered if the user sets Rxn vol~xxxuL ~-xxxs "Special tube?" to YES.

O - 182 ~TI~ - DELETE 2 ~ ~ 6 7 ~ 3 Delete PROGRAM-USER-ALL D-let~ by PRO~ All programs (files and methods) Deletecan be deleted by number.

Enter program #xxx A program cannot be deleted if Can~t delete progxxx it linked in a method.

Linked in methodxxx!

The user has entered the # of a Progxxx is protected protected program.

The correct Enter user ~xxxx user # must be entered in order lo to delete this program.

The wrong user # was entered.

Progxxx i8 protected This display remains for 5 Wrong user number! 6econds before reverting to the previous one.

The user is given 3 chances to enter the correct #.

Ready to delete the program.

The Prog ~xxx U~er ~xxxx user ~ appears only if the Delete proqram? YES program was protected.

20 D-l-t- ~y U8E~ Programs can be deleted under a Delete given user number.

Enter user ffxxxx If no programs exist with the Delete given user #, the following 25 No progs with ~xxxx message is displayed.

~ - 183 - 2~ 3 Programs cannot be deleted if Progs linked in meth they are linked in a method.

The STEP to list progs STEP key will cycle through the list of linked programs.

~ - 184 - 2 ~ ~ ~ '7~ ~ ~TlL - DELETE (cont) The list of the linked programs Can't delete progxxx will show which ~ethod the Linked in methodxxxl ~Lo~Lam is linked to.

This will delete all the User ~xxxx programs under the given user # Delete all progs?YES that are not linked.

D-l-t~ AL~ This will delete every 10 Delete everY unprotected program that is not unprotected prog?YES linked in a protected method.

~ - 185 - 2~7~3 ~IL -- IJ8ER DIAGN08TIC8 While running any diagnostic test, the STOP key always returns the user to the top level diagnostic screen and automatically increments the test number and name to the 5 next test. This facilitates manually cycling through the available diagnostics.

The user can enter the number of - Enter Diag Test t~ the diagnostic to run or can use ~VIEW H~STORY FILE the STEP or BACK keys to cycle through the available tests.

Every time the STEP or BACK key i5 pressed, the test number is incremented or decremented and the associated test name is displayed.

This feature eliminates the need for the user to memorize the number associated with each test.

RE~IElr HI8~0RY FILE The history file is a circular Enter ~iag Test #1 buffer in battery RAM which can R~VIL~N HISTORY ~IL~ store up to 500 records of the latest run.

When the buffer is full, the oldest entries will be overwritten.

The buffer will automatically be cleared before a program is executed.

The history file header displays HISTORY nnn recs the current number of records in 30 AL1-STAT-ERRORS-PRNT the file ('nnn').

AL~ views all the records BTAT views only the status records ~ P~ views only the records with error messages ~R~T prints all or part of the history file The two types of records are 1) status records which give information about the program and 2) data records which give 40 information abount each hold and ramp segment in a program.

A Hold program is treated as one hold segment and the data record will be stored when the file ends.

Since there could be hundreds of entries (50 cycles X 6 setpoints ~= 350 entries), fast, bi-directional movement 45 through the file is required. Note that most PcR programs will be 3 or 6 setpoints and 40 cycles or less. The entries will normally be reviewed in reverse order, thus the first ~ - 186 ~ 7 record ~een will be the last record written.

If the user has chosen a type of record to view, STEP or BACK will move down or up the buffer by one entry of the chosen type.

By preceding STEP or BACK with a number, the 5 second line is replaced with "Skip ~XXX entries". The user enter~ a number and presses ENTER to accept the value and that number of entries is skipped going forward (STEP) or backward (BACK).

By preceding STEP or BAcK with the RUN key, the user can lO quickly move to the largest record ~ (the newest record) or record #l (the oldest record) of the chosen type.

STOP terminates the review mode and displays the file header.

- 187 8TATIJ8 P~CO~2n 'ffff' is either HOLD, CYCL or ffff ~xxx/m~m nnn AUTO ~esc~ge 'xxx' lS the program number '/mmm' is the method number for a linked program, else blank 'nnn' is the record n~ h~r 'message' i~ one of the following:

8tatu- ~-ssag~s 10 Tube Type: xxxxx Type of sample tube used in the run Reaction vol: xxxuL Reaction volume used in the run Clk starts w/in x.xC The hold clock ~tarts within this temp of setpoint Start xx/xx/xx xx:xx Time and date of the start of the run 15 End xx/xx/xx xx:xx Time and date of the end of the run Meth Complete All programs linked in the method are complete Pause xx:xx at xx.xC The program paused for this time at this temp 20 Fatal ~tatus ~essagos Sensor Error A sensor had a bad reading 10 times in a row Power fail xxx.x hrs The power was off for this amount of time 25 User Abort The user pressed the STOP key during the run Pause Timeout xx:xx The keypad pause has reached its configurable time limit.

Fatal Setpoint Error Is the requirement to abort a program if the setpoint is not reached within a calculated amount of time. A lO X lO lookup table of starting ramp temperature (0~C - 100~C in 10~C increments) vs. ending ramp temperature (same axis labeling) will hold the average time the TC2 should take to ramp up or down any given amount of degrees.

The file will be aborted if the setpoint is not reached in the amount of time calculated as follows:

programmed r~mp time + (2 * lookup table value) + 10 minutes DATA P~Co~n 'f' is either HOLD, CYCL or auTo 'xxx' is the program number - 188 - 2~ '/mmm' is the method number for f~xxx/mmm ddd.dC nnn a linked program else blank Cycyy Setpt z mmm: 55 'ddd.d' is the ending setpoint temp 'nnn' is the record number 'yy' is the cycle number ~zl is the setpoint number 'mmm:ss' is the setpoint time The cycle 2nd setpoint number fields will be omitted for a Hold program.

~ 2 ~ 3 - 189 DATA ERROR RECORD 'ddd.d' is the ending setpoint message ddd.dC nnn temp Cycw Setpt z mmm:ss 'nnn' is the record number 'yy' is the cycle number 'z' is the setpoint number 'mmm:ss' is the setpoint time 'message' indicates a non-fatal error as follows:

10 Non-fat-l Error n-ssag-s Setp Error The setpoint was not reached in the calculated time:

programmed ramp time + (2 * lookup table value)~ 15 Prog Error An Auto program auto increment/decrement of the setpoint temp or time caused the hold time to go negative or the temp to go out of the range 0.1~C to 100~C.

- Temp Error At the end of the segment, the setpoint temp has drifted +/- a user configurable amount.

For the Hold program, the cycle and setpoint fields will be omitted.

~ lgo- 2~5~7~3 PRIN~ING q!~E HI8q!0RY FILE Access to the history file print routines is through the history file header menu. The OPTION key cycles the cursor through the options:

HISTORY nnn recs ALL-STAT-ERRORS-PRNT 5 Pressing the ENTER key when the cursor is positioned under PRNT displays the print screen:

Print History ALL-STAT-ERRORS ALL prints all the records in the file STAT prints only the status records ERRORS prints on}y the records with error messages 10 When one of print options is selected, the following screen is displayed:

Print History Print from prog #xx The first (most recent) program number will be the default program. The user can change the program number from which to begin printing. While printing, the following screen is 15 displayed:

Print H~story ...printing At the end of printing, the Print History menu i5 again displayed.

~ 2~6~ -- 191 -~EATER TB8 -Enter.Diag Test HEATER TEST The heater test calculates the heat rate of the sample block as its temperature rises from 35~C to 65~C. The following screen is displayed as it forces the block temperature to 5 35~C.

Heater Test Blk=XX.X going to 35C...

When the temperature stabilizes, all heaters are turned on full power. The display now reads Hgoing to 65C" and the block temperature is monitored for 20 seconds after it passes 50~C. After 20 seconds, a pass or fail ~essage is 10 displayed.

Heater Test PASSES .

2 ~ 7 ~ 3 -- 192 - CuTT-T-~ TEST Enter Diag Test ~HILLER TEST The chiller test calculates the cool rate of the sample block as its temperature drops from 35~C to 15~C. The following screen is displayed AS it forces the block s temperature to 35~C.

Chillr Test Blk~XX.X qoing to 35C...

When the temperature stabilizes, the chiller is on. The display now reads "going to 15C" and the ~lock temperature is monitored for 20 seconds after it passes 25~C. After 20 seconds, a pass or fail message is displayed.

Chiller test PASSES