Sample-and-hold circuit.

BACKGROUND OF THE INVENTION

1. Field of the Invention

0001The present invention relates to a sample-and-hold circuit adapted for a video signal or the like having a blanking period per predetermined cycle, and more particularly to an improved circuit so designed as to produce an output signal without any difference relative to the gate-source offset voltage induced by a buffer of a source follower or the like.

2. Description of the Prior Art

0002In the conventional circuit for sampling and holding an input such as a video signal, there is known an exemplary configuration of Fig. 12. In this example, a signal received at an input terminal 201 is fed to a gating MOS transistor 208 which is turned on by a sampling pulse fed thereto from a terminal 203. The output signal of the MOS transistor 208 is fed to a holding capacitor 211, which then produces a signal fed to the gate of a buffer MOS transistor 209. A line voltage is applied through a power terminal 205 to the drain of the MOS transistor 209, whose source is grounded via a constant current source 213. And the signal produced at the source of the MOS transistor 209 is fed to an output terminal 206.

0003In the circuit configuration mentioned above, a buffer circuit composed of a source follower is provided in the output stage. Accordingly, when sampling pulses of Fig; 13B are fed from the terminal 203, signals obtained in the individual stages become such as shown in Fig. 13A. In this chart, a solid line represents the waveform of an input voltage applied to the input terminal 201 point ); a one-dot chain line represents the waveform of a voltage at the gate (point ) of the MOS transistor 209; and a broken line represents the waveform of an output voltage produced at the output terminal 206 (point ). As is clear from Fig. 13A, there occurs in the output a drop of V_GS (offset voltage) which is expressed as<maths id="math0001" num=""><img file="EP0190973A2_D0001.tif" /></maths>where V_th is a threshold voltage of the MOS transistor 209; _P is a mobility of a carrier; C_ox is a gate capacity per unit area; W is a channel width; L is a channel length; and I_o is a current value of the constant current supply 213.

0004It follows, therefore, that some problems are unavoidable including DC potential variation between the input and the output as well as so-called 1/f noise derived from nonuniformity of the threshold voltage V_th, drift of the current I_o and so forth.

0005In an arrangement where a plurality of sample-and-hold circuits are connected in series with one another, a voltage drop of V_GS occurs in each stage, so that a great DC variation is induced in total between the input and the final output.

0006In an attempt to eliminate such disadvantages, an improved circuit configuration of Fig. 14 has already been proposed by the present applicant. In the circuit of Fig. 14, there are shown an input terminal 221, a terminal 223 to be fed with a blanking pulse ØBLK, and a terminal 224 to be fed with a sampling pulse Ws.

0007A drain of a gating MOS transistor 227 is connected to the input terminal 221, and a gate thereof is connected to the terminal 224. Meanwhile a source of the MOS transistor 227 is connected via a capacitor 231 to a gate of an enhancement-type output MOS transistor 229 while being grounded via a capacitor 232 and is further connected to a drain of a gating MOS transistor 230. A drain of another gating MOS transistor 228 is connected to the input terminal 221, while a gate and a source thereof are connected respectively to the terminal 223 and the gate of the transistor 229. The gate of the gating MOS transistor 230 is connected to the terminal 223, and its source is connected to the source of the enhancement-type output MOS transistor 229, whose drain is connected to a power terminal 225. And a constant current supply 233 is connected to the source of the MOS transistor 229, which is led out to an output terminal 226.

0008In the sample-and-hold circuit of Fig. 14 mentioned above, when a blanking pulse ØBLK of Fig. 15B and a sampling pulse ØS of Fig. 15C are fed respectively to the terminal 223 and the terminal 224, then signals of Fig. 15A are obtained in the individual stages, where a solid line represents an input voltage applied to the input terminal 221, and a broken line represents an output voltage produced at the output terminal 226.

0009In case a blanking pulse ØBLK is fed when the input terminal 221 has a voltage V_SO, the transistors 228 and 230 are turned on so that the gate potential of the MOS transistor 229 becomes V_SO while the source potential (at the output terminal 226) is :educed by V_GS through the MOS transistor 229 as V_SO- V_GS, whereby the potential at the junction of the MOS transistor 230 and the capacitor 232 is changed as V_{SO -} V_GS. Consequently the difference between the gate potential and the source potential of the MOS transistor 229 becomes as follows, and a charge corresponding to such potential difference is stored in the capacitor 231.<maths id="math0002" num=""><img file="EP0190973A2_D0002.tif" /></maths>

0010The MOS transistors 228 and 230 are turned off in the state described above, and a sampling pulse ØS is fed. In case the potential at the input terminal 121 is V_S1 when a first sampling pulse ØS1 is fed, the potential V_S1 is obtained also at the junction of the transistor 227 and the capacitor 232. Since the charge corresponding to the aforesaid potential difference V_GS is previously stored in the capacitor 231 in response to the blanking pulse ØBLR, the potential at the junction of the capacitor 231 and the transistor 229 is increased to be V_S1 + V_GS. As a result, the potential at the output terminal 126 becomes<maths id="math0003" num=""><img file="EP0190973A2_D0003.tif" /></maths>which is equal to the potential at the input terminal 221.

0011The above operation is performed at each sampling time and therefore enables the sampled potential to appear at the output without any variation despite the drop of V_GS induced by the buffer.

0012In the sample-and-hold circuit of Fig. 14 where the enhancement-type MOS transistor 229 is employed in its output stage, the input signal level needs to be higher than at least the threshold voltage Vt_h of the MOS transistor 229. Accordingly the dynamic range of the circuit is required to be D_l which is greater than the essential signal dynamic range (e.g. D₀) by a value corresponding to the offset voltage V_GS.

0013Furthermore, due to the Early effect, there occurs a variation in the offset voltage V_GS of the MOS transistor 229 with a shift of the operation point. Consequently a problem arises with regard to a resultant difference between the sample voltage V_GS held in the capacitor 231 during the blanking period and the actual voltage V_GS of the MOS transistor 229 at any time other than the blanking period.

0014Furthermore, in the circuit configuration mentioned above, an input capacity C_IN is generated alternatingly in the output-stage MOS transistor 229 as viewed from the input signal source. Consequently the output from the source of the MOS transistor 229 is produced with a delay, and there exists another problem that the power consumption increases with charge and discharge of the input capacity C_IN.

SUMMARY OF THE INVENTION

0015It is therefore an object of the present invention to provide an improved sample-and-hold circuit which is capable of accurately cancelling the gate-source offset voltage V_GSin the output stage and thereby producing an output voltage without any voltage difference of V_GS relative to the input voltage with another advantage of minimizing the required dynamic range.

0016Another object of the present invention is to provide an improved sample-and-hold circuit which is capable of reducing the input capacity of the buffer MOS transistor as viewed from the input signal source.

0017According to an aspect of the present invention, an input signal is fed via a first gate element to one end of a first capacitor whose other end is grounded alternatingly, the one end of the first capacitor being connected via a second capacitor to a gate (or base) of a source (or emitter) follower transistor to obtain an output from the source (or emitter) of the source follower transistor which is connected via a second gate element to one end of the first capacitor, while the gate (or base) of the source follower transistor is connected via a third gate element to a DC voltage supply having a predetermined voltage value, and the second and third gate elements are turned on during a first period of the input signal so that a voltage corresponding to the gate-source (or base-emitter) offset voltage of the source follower transistor is stored in the second capacitor, while the first gate element is turned on during a second period of the input signal to reduce an output signal equivalent in level to the input signal.

BRIEF DESCRIPTION OF THE DRAWINGS

0018<ul id="ul0001" list-style="none"><li>Fig. 1 is a connection diagram of an exemplary embodiment of the present invention;</li><li>Fig. 2 is a waveform chart for explaining the operation in the exemplary embodiment of Fig. 1;</li><li>Fig. 3 is a waveform chart of signals in individual stages of a buffer circuit to which the present invention is applied;</li><li>Fig. 4 is a connection diagram of another embodiment of the invention;</li><li>Fig. 5 is a waveform chart for explaining the operation in the embodiment of Fig. 4;</li><li>Fig. 6 is a connection diagram of a further embodiment of the invention;</li><li>Fig. 7 is a waveform chart for explaining the operation in the embodiment of Fig. 6;</li><li>Fig. 8 is a connection diagram of a further embodiment of the present invention;</li><li>Fig. 9 is a waveform chart of signals for explaining the operation in the embodiment of Fig. 8;</li><li>Fig. 10 is a connection diagram of a still further embodiment of the present invention;</li><li>Fig. 11 is a waveform chart of signals for explaining the operation in the embodiment of Fig. 10;</li><li>Figs. 12 and 14 are connection diagrams of conventional sample-and-hold circuits;</li><li>and Figs. 13 and 15 are waveform charts for explaining the operations in the conventional sample-and-hold circuits.</li></ul>

DESCRIPTION OF THE PREFERRED EMBODIMENT

0019Hereinafter an exemplary embodiment of the present invention will be described with reference to the accompanying drawings. In Fig. 1 shown at 1 is an input terminal to which a video signal or the like is fed.

0020The drain of a gating MOS transistor 7 is connected to the input terminal 1, and the gate thereof is connected to a terminal 4 to be fed with a sampling pulse Ø5. The source of the MOS transistor 7 is connected via a capacitor 11 to the gate of an enhancement-type output MOS transistor 9, and the junction between the source of the MOS transistor 7 and the capacitor 11 is grounded via a capacitor 12 while being connected to the drain of a gating MOS transistor 10.

0021The drain of a gating MOS transistor 8 is connected to a terminal 2 to which the bias voltage V_B is applied, and the gate thereof is connected to a terminal 3 to be fed with a blanking pulse $_BLK. Meanwhile the source of the MOS transistor 8 is connected to the gate of the output MOS transistor 9. The gate of the gating MOS transistor 10 is connected to the terminal 3 to be fed with a blanking pulse Ø_BLK, and the source thereof is connected to the source of the MOS transistor 9.

0022The drain of the output MOS transistor 9 is connected to a power terminal 5 while a constant current supply 13 is connected to the source thereof, from which an output terminal 6 is led out.

0023The terminal 2 is fed with a predetermined bias voltage V_B substantially equivalent to the mean level of the input signal, and the terminal 3 is fed with a blanking pulse Ø_BLK of Fig. 2B while the terminal 4 is fed with a sampling pulse Ø_S of Fig. 2C.

0024In the circuit configuration described above, signals of Fig. 2A are produced in the individual stages.

0025When a blanking pulse Ø_BLK is fed, the transistors 8 and 10 are turned on so that the potential at the gate (point ②) of the MOS transistor 9 becomes V_B as shown by a one-dot chain line in Fig. 2A while the potential at the source (point ) is reduced by V_GS through the MOS transistor 9 to become V_B- V_GS as shown by a broken line in Fig. 2A, whereby the potential at the junction (point ) of the MOS transistor 10 and the capacitor 12 is changed as V_B- V_GS. Consequently the potential difference between the points and becomes as follows, and a charge corresponding to such potential difference is stored in the capacitor 11.<maths id="math0004" num=""><img file="EP0190973A2_D0004.tif" /></maths>

0026The MOS transistors 8 and 10 are turned off in the state described above, and then a sampling pulse ØS is fed. In case the potential at the input terminal 1 (point ) is V_S1 when a first sampling pulse Ø_S1 is fed, the potential V_S1 is obtained also at the point . Since the charge corresponding to the aforesaid potential difference V_GS is previously stored in the capacitor 11 between the points ④ and ② in response to the blanking pulse Ø_BLK, the potential at the point is increased to be V_S1 + V_GS. As a result, the potential at the point ③ becomes<maths id="math0005" num=""><img file="EP0190973A2_D0005.tif" /></maths>which is equal to the potential at the input terminal 1 (point ).

0027The above operation is performed at each sampling time and therefore enables the sampled potential to appear at the output without any variation despite the drop of V_GS induced by the buffer.

0028In case the above embodiment of the present invention is used as a buffer circuit, a pulse held at a high level as shown in Fig. 3B is fed to the terminal 4 in relation to an input video signal of Fig. 3A. And the configuration may be so designed that the offset voltage V_GS of the MOS transistor 9 is stored in the capacitor 11 in response to a blanking pulse of Fig. 3C fed per blanking period of the video signal.

0029Fig. 4 shows another embodiment of the present invention applied to an analog delay circuit, in which two of the foregoing embodiment of Fig. 1 are arrayed in cascade connection.

0030A first sample-and-hold circuit comprises gating MOS transistors 27, 28 and 30; capacitors 31 and 32; an output MOS transistor 29 and a constant current supply 33 similarly to the aforesaid exemplary embodiment. Meanwhile a second sample-and-hold circuit comprises gating MOS transistors 37, 38 and 40; capacitors 41 and 42; an output MOS transistor 39 and a constant current supply 43 similarly to the above.

0031In Fig. 4, an input terminal 21 is connected to the drain of the MOS transistor 27, whose gate is connected to a terminal 24 to be fed with a first sampling pulse Ø_S1. A terminal 22, to which a bias voltage V_B is applied, is connected to the respective drains of the MOS transistors 28 and 38, while a terminal 23 to be fed with a blanking pulse 4_BLK is connected to the respective gates of the mos transistors 28, 30 and 40. The source of the output MOS transistor 29 is connected to the drain of the MOS transistor 37, whose gate is connected to a terminal 34 to be fed with a second sampling pulse Ø_S2. A power terminal 25 is connected to the respective drains of the output MOS transistors 29 and 39, and an output terminal 36 is led out from the source of the output MOS transistor 39.

0032A blanking pulse 6_BLK of Fig. 5A is fed to the terminal 23, and offset voltages V_GS of the output MOS transistors 29 and 39 are stored respectively in the capacitors 31 and 41. In response to a first sampling pulse Ø_S1 of Fig. 5B, an input signal is sampled and held to appear as an output at the source of the MOS transistor 29. The output thus produced is then fed to the drain of the MOS transistor 37. And when a second sampling pulse Ø_S2 of Fig. 5C having a predetermined phase lag with respect to the first sampling pulse Ø_S1 is fed to a terminal 34, the output of the preceding stage is sampled in response to the second sampling pulse Ø_S2. Consequently an output delayed for the phase lag between the second sampling pulse Ø_S2 and the first sampling pulse Ø_S1 is obtained at an output terminal 36.

0033Fig. 6 shows another embodiment of the present invention applied to an analog serial-parallel conversion circuit, in which two of the foregoing embodiment shown in Fig. 4 are connected in parallel with each other.

0034A first delay circuit comprises gating MOS transistors 57, 58, 60, 67, 68 and 70; capacitors 61, 62, 71 and 72; output MOS transistors 59 and 69; and constant current supplies 63 and 73 in the same configuration as that of Fig. 4. And a second delay circuit comprises gating MOS transistors 77, 78, 80, 87, 88 and 90; capacitors 81, 82, 91 and 92; output MOS transistors 79 and 89; and constant current supplies 83 and 93 in the same configuration as that of Fig. 4.

0035In Fig. 6, an input terminal 51 is connected to the respective drains of the MOS transistors 57 and 77, and the gate of the MOS transistor 57 is connected to a terminal 54 to be fed with a first sampling pulse Ø_S1. A terminal 52, to which a bias voltage V_B is applied, is connected to the respective drains of the MOS transistors 58 and 68, while a terminal 53 to be fed with a blanking pulse Ø_BLK is connected to the respective gates of the MOS transistors 58, 60, 68 and 70. A power terminal 55 is connected to the respective drains of the output MOS transistors 59 and 69. And a first output terminal 66 is led out from the source of the output MOS transistor 69.

0036A terminal 74 to be fed with a second sampling pulse d_S2 is connected to the respective gates of the MOS transistors 67 77 and 87. And a terminal 76, to which a bias voltage V_B is applied, is connected to the respective drains of the MOS transistors 78 and 88, while a terminal 53 to be fed with a blanking pulse ¢BLK is connected to the respective gates of the MOS transistors 78, 80, 88 and 90. A power terminal 75 is connected to the respective drains of the output MOS transistors 79 and 89. And a second output terminal 86 is led out from the source of the output MOS transistor 89.

0037A blanking pulse ¢_BLK of Fig. 7A is fed to the terminal 53, and offset voltages V_GS of the output MOS transistors 59, 69, 79 and 89 are stored respectively in the capacitors 61, 71, 81 and 91. In response to a first sampling pulse Ø_S1 of Fig. 7B, an input signal (Fig. 7D) is sampled and held to appear as an output at the source (point in Fig. 6) of the MOS transistor 59. The output (Fig. 7E) thus produced is then fed to the drain of the MOS transistor 67. When a second sampling pulse Ø_S2 of Fig. 7C having a predetermined phase lag with respect to the first sampling pulse Ø_S1 is fed to a terminal 74, the output (Fig. 7E) of the preceding stage is sampled in response to the pulse Ø_S2. Consequently an output (Fig. 7G) delayed for the phase lag between the second sampling pulse Ø_S2 and the first sampling pulse O_S1 is obtained at an output terminal 66.

0038The input signal at this moment is sampled in response to the sampling pulse Ø_S2, so that an output (Fig. 7F) comes to appear at the source (point in Fig. 6) of the MOS transistor 79. The output (Fig. 7F) thus produced is sent via the MOS transistors 87 and 89 to the output terminal 86. That is, the input signal rendered parallel as shown in Figs. 7G and 7H is obtained from the output terminals 66 and 86.

0039A further embodiment of the present invention in which the input capacity of the output transistor can be reduced will be described with reference to the Fig. 8 through Fig. 11. In Fig. 8, shown at 101 is an input terminal to which a video signal or the like is fed.

0040The drain of a gating MOS transistor 107 is connected to the input terminal 101, and the gate thereof is connected to a terminal 104 to be fed with a sampling pulse Ø_S.

0041The source of the MOS transistor 107 is connected via a capacitor 115 to the gate of an enhancement-type output MOS transistor 111. And the junction between the source of the MOS transistor 107 and the capacitor 115 is grounded via a capacitor 114 while being connected via a capacitor 113 to the gate of a MOS transistor 110.

0042The respective gates of MOS transistors 108, 109 and l12 are connected to a terminal 103 to be fed with a blanking pulse Ø_BLK. The drain of the MOS transistor 108 is connected to the input terminal 101, and the source thereof is connected to the gate of the output MOS transistor 111. The drain of the MCS transistor 109 is connected to a terminal 102 to which a DC voltage V_DC1 is applied, and its source is connected to the gate of the MOS transistor 110. The drain of the MOS transistor 112 is connected to the source of the MOS transistor 107, and the source of the MOS transistor 112 is connected to the source of the output MOS transistor 111.

0043The drain of the MOS transistor 110 is connected to a power terminal 105, and its source is connected to the drain of the output MOS transistor 111. Meanwhile a constant current supply 116 is connected to the source of the MOS transistor 111, from which an output terminal 106 is led out.

0044An input such as a video signal of Fig. 9A is fed to the input terminal 101. The minimum level of the video signal or the sync tip level V_BLK thereof during the blanking period is so set as to exceed the threshold voltage Vt_h of the output MOS transistor 111. To the terminal 102, there is applied from a DC voltage supply a voltage V_DC1 which is higher than the sync tip level V_BLK and hence further higher than the threshold voltage Vt_h of the MOS transistor 110. And the terminal 103 is fed with a blanking pulse Ø_SLK of Fig. 9B, while the terminal 104 is fed with a sampling pulse Ø_S of Fig. 9C.

0045When a blanking pulse Ø_BLK is fed to the terminal 103, the MOS transistors 108, 109 and 112 are turned on. Upon conduction of the MOS transistor 109, the potential at the gate (point in Fig. 8) of the MOS transistor 110 is equalized to the DC voltage V_DC1 applied via the terminal 102. Consequently, the potential at the source (point in Fig. 8) of the MOS transistor 110 is reduced by the gate-source offset voltage V_GS110 as _VDC1- V_GS110.

0046Upon conduction of the MOS transistor 108 in response to the blanking pulse Ø_BLK, the potential at the gate (point in Fig. 8) of the output MOS transistor 111 is equalized to the instantaneous potential at the input terminal 101 (point in Fig. 8) to become V_BLK. As a result, the potential at the source (point in Fig. 8) of the output MOS transistor 111 is reduced by the gate-source offset voltage V_GS111 as V_BLK- V_GS111. And due to conduction of the _MO_S transistor 112 in response to the blanking pulse Ø_BLK, the potential at the junction (point in Fig. 8) between the MOS transistor 112 and the capacitor 114 is also reduced as _VBLK- V_GS111.

0047Therefore the potential difference between the points ③ and ② becomes as follows, and a charge corresponding to such potential difference is stored in the capacitor 113.<maths id="math0006" num=""><img file="EP0190973A2_D0006.tif" /></maths>Meanwhile, the potential difference between the points and becomes as follows, and a charge corresponding thereto is stored in the capacitor 115.<maths id="math0007" num=""><img file="EP0190973A2_D0007.tif" /></maths>

0048The MOS transistors 108, 109 and 112 are turned off in the state described above, and then a sampling pulse Ø_S is fed to the terminal 104. The MOS transistor 107 is turned on by the sampling pulse Ø_S, and the potential at the point is equalized to the instantaneous potential at the input terminal 101 (point in Fig. 8) to become V_S, for example.

0049Since the charge corresponding to _VDCl - (VBLK - V_GS111) is stored in the capacitor 113 between the points and in response to the blanking pulse Ø_BLK as described previously, the potential at the point is altered to V_S + V_DCl- (V_BLK- V_GS111). Accordingly, the potential at the point 4 is changed to

0050<maths id="math0008" num=""><img file="EP0190973A2_D0008.tif" /></maths>

0051Furthermore, as also described already, the charge corresponding to V_GS111 is stored in the capacitor 115 between the points and in response to the blanking pulse Ø_BLK, so that the potential at the point is altered to V_S + V_GS111. Accordingly, the potential at the point ⑥ is changed to<maths id="math0009" num=""><img file="EP0190973A2_D0009.tif" /></maths>Thus, the output voltage equalized to the input voltage applied to the input terminal 101 is obtained from the output terminal 106. The above operation is performed at each sampling time and therefore enables the sampled potential to appear at the output without any variation despite the drop of V_GS induced by the buffer.

0052Furthermore, each of the alternating potential variations at the drain and the source of the output MOS transistor 111 is Vs so that the apparent input capacity C_IN as viewed from the input signal source is reduced to nil.

0053When it is necessary in the above-described embodiment of the present invention to take into consideration the dynamic range of the circuit or any V_GS variation caused by the Early effect, a positive pilot signal represented by a broken line Pt in Fig. 9A may additionally be provided during the blanking period of the video signal fed to the input terminal 101. In such a case, the pilot signal Pt is inserted in the high-level duration of the blanking pulse Ø_BLK, and the peak level of the pilot signal Pt is set to be substantially equal to the mean level of the input signal.

0054By the use of such pilot signal Pt, the offset voltage V_GS111 at the same operation point as the signal level can be stored in the capacitor 115 in response to the blanking pulse Ø_BLK. Moreover, since the peak level of the pilot signal Pt is substantially equalized to the mean level of the input signal, it is not necessary to set the sync tip level V_BLK of the input signal to be higher than the threshold voltage Vt_h of the output MOS transistor, hence minimizing the required dynamic range of the circuit.

0055In a modification, an additional terminal may be provided separately from the input terminal 101 to receive a DC voltage equivalent to the mean level of the input signal, and the drain of the MOS transistor 108 may be connected to such terminal instead of the input terminal 101.

0056The output MOS transistor 111 of enhancement type employed in the above embodiment of the invention may be of depression type as well, and a bipolar transistor may be used in place of the MOS transistor.

0057Fig. 10 shows still further embodiment of the present invention contrived to constitute an analog delay circuit, wherein two of the foregoing embodiment of Fig. 8 are arrayed in cascade connection with a modification that the drain of the MOS transistor 108 is connected not to the input terminal but to an additional terminal provided separately to receive a DC voltage.

0058In this example, a first sample-and-hold circuit 120 comprises MOS transistors 127, 128, 129, 130 and 132; capacitors 133, 134 and 135; an output MOS transistor 131 and a constant current supply 136. Meanwhile a second sample-and-hold circuit 140 comprises MOS transistors 147, 148, 150 and 152; capacitors 153, 154 and 155; an output MOS transistor 151 and a constant current supply 156.

0059In Fig. 10, an input terminal 121 is connected to the drain of the MOS transistor 127, whose gate is connected to a terminal 124 to be fed with a first sampling pulse Ø_S1. A terminal 122, to which a DC voltage V_DC1 is applied, is connected to the respective drains of the MOS transistors 129 and 149, while a terminal 123 to be fed with a blanking pulse ¢_BLK is connected to the respective gates of the MOS transistors 128, 129, 148 and 149. A terminal 137, to which a DC voltage V_DC2 is applied, is connected to the respective drains of the MOS transistors 128 and 148. A power terminal 125 is connected to the drain of the MOS transistor 130, and the source of the output MOS transistor 131 in the first sample-and-hold circuit 120 is connected to the drain of the MOS transistor 147 in the second sample-and-hold circuit 140. The gate of the MOS transistor 147 is connected to a terminal 144 to be fed with a second sampling pulse Ø_S2. Furthermore, a power terminal 145 is connected to the drain of the MOS transistor 150, and an output terminal 146 is led out from the source of the output MOS transistor 151.

0060A video signal is fed to the input terminal 121, and a DC voltage V_DC2 substantially equalized to the mean level of the video signal is applied to the terminal 137. A blanking pulse Ø_BLK of Fig. 11A is fed to the terminal 123, so that the offset voltage V_DC2-(^VDC2 - ^VGS111) = V_GS111 of the output MOS transistors 131 and 151 is stored in each of the capacitors 135 and 155. Meanwhile the voltage V_DCl- (V_DC2- V_GS111) is stored in each of the capacitors 133 and 135.

0061The voltage at the drain of the output MOS transistor 131 is altered to V_S + V_DC1- (V_DC2- V_GS111) - V_GS110 in response to the first sampling pulse Ø_S1 of Fig. 11B, whereby the input capacity of the output MOS transistor 131 is reduced to nil alternatingly. At the source of the MOS transistor 131 is obtained a sample holding output V_S in which the offset voltage V_GS111 has already been canceled relative to the input voltage. This output is fed to the drain of the MOS transistor 147 in the second sample-and-hold circuit 140, and a second sampling pulse Ø_S2 of Fig. 11C having a predetermined phase lag with respect to the first sampling pulse Ø_S1 is fed to the terminal 144. Then the first output V_S of the preceding stage is sampled in response to the second sampling pulse Ø_S2. This time the voltage at the drain of the output MOS transistor 151 is altered to _{VS + VDCl}- (V_DC2- VGS111) - V_GS110, whereby the input capacaty of the output MOS transistor 151 is reduced to nil alternatingly.

0062Consequently an output delayed for the phase lag between the second sampling pulse Ø_S2 and the first sampling pulse Ø_S1 is obtained at the output terminal 146.

0063In addition to the above embodiment where two states of sample-and-hold circuit are arrayed in cascade connection, the present invention is applicable also to another arrangement employing three or more stages of sample-and-hold circuits in cascade connection. And it is to be further understood that the latter embodiment is applicable to a serial-parallel conversion circuit as well.

0064According to the present invention, the offset voltage V_GS of the MOS transistor in the output stage is held in the capacitor during the blanking period, and a composite signal obtained by adding such offset voltage V_GS in the capacitor and the input voltage sampled and held in response to a sampling pulse is fed to the gate of the output-stage MOS transistor. It becomes possible, therefore, to provide an output voltage without inducing any offset voltage difference relative to the input voltage.

0065Furthermore, due to application of the bias voltage V_B which is substantially equivalent to the mean level of the input signal, complete coincidence is attainable between the offset voltage sampled and held in response to a blanking pulse and the offset voltage at the time of sampling and holding the input signal, and there is achievable another advantage of minimizing the required dynamic range.