High speed free-space optical communications

(Cross-reference to related applications) This application claims the interests of US Provisional Patent Application No. 61 / 528,119 filed on August 26, 2011 and US Provisional Patent Application No. 61 / 671,036 filed on July 12, 2012. And each of these documents is incorporated herein by reference as is.

The bandwidth of the transmitted data and the extent to which this data can be transmitted outdoors is such that it is equipped with a slow single or multiple high power semiconductor laser transmitters, or optical modulators and / or optical amplifiers. It has relied on technology that depends on the combination of transmitters, or technology that uses multiple wavelengths in combination with the components described above to achieve high bandwidth of free-space optical communication over distances of more than a few meters. It was. To date, the complexity involved in implementing these techniques has been exorbitantly costly for localized systems, especially for short distances in meters. The power available from the available link budgets, or emitters, should be cost-considered, as are the complex and expensive alignment and detection issues. A cost-effective wireless optical transmitter with many link budgets is desirable. Vertical cavity light emitting ("VCSEL") arrays can produce the required optical power for the above distances and are much more cost effective, while existing VCSEL arrays require a very high band ("VCSEL"). It is not possible to create (generally associated with a single VCSEL device).

For short-range optical communications, the use of a fiber structure between adjacent transmitters and receivers on the circuit board limits the alignment of the fiber to the laser opening. This alignment is generally achieved with mechanically assembled parts that have increased in size and manufacturing costs, and this problem is exacerbated by a large number of fibers. The low-power-based free-space optical design in the link budget means that achievable tolerances require extremely mechanical board-to-board alignment, which is elaborate mechanical. It is costly to use a flexible connector structure. Again, a single VCSEL device is best suited for bandwidth and cost structure, but lacks the required power and limits alignment to nearly unattainable tolerances.

<p num="0004"> The relevant application illustrates how a VCSEL array is manufactured and tested, and the results are superior to state-of-the-art technology when considering power and speed. Due to its structural flexibility in terms of design and implementation possible with VCSEL arrays, unique configurations such as sub-array arrays, array multi-wavelength arrays, and patterned shapes can be easily achieved. It allows the optical path to be easily and quickly scanned without mechanical means, and to shape the optical path according to an array configuration, or to increase the performance of the optical path with a large number of possible links. The ability to use these technologies to create cost-effective, high-speed, high-power arrays creates unique opportunities for cost-effective, high-speed optical wireless communications.</p><p num="0005"> Similarly, high-speed optical communication between adjacent circuit boards has traditionally used wavelength division multiplexing into fiber optics or multiple fiber optic transmitters and receivers, or multi-channel single fiber optic configurations or semiconductor laser free space optical transmitter configurations of data. Has been achieved. In any of these configurations, sufficient energy needs to be delivered from the emitter to the detector to achieve the minimum signal-to-noise ratio in the desired bandwidth. A successful design begins with the available power or link budget, then calculates all the losses incurred in the system and ends with sufficient output density at the detector. Fast detectors are small and therefore require more output density to maintain a sufficient signal-to-noise level. One of the serious losses in link budget calculations is the loss of detector alignment with respect to the emitter. System reliability relies heavily on the amount of power available from the emitter to overcome alignment problems and other system losses.</p><p num="0006"> Optical communication between adjacent circuit boards offers many advantages, including higher bandwidth than is available for copper connections. Free space optics will usually be preferred over fiber systems due to the simplicity of the components. However, in free space systems, adjustments must be made taking into account translations and angles between the transmitter and receiver and both misalignments. In addition, laser safety matters must be considered. In general, free space configurations are limited by expensive or precision mechanical connectors or low power devices or link budgets that require expensive mounting configurations. The embodiments disclosed herein are techniques disclosed in related applications, including US Pat. No. 7,949,024 by the same applicant, to enable high power arrays of VCSEL devices operating in very high bandwidth. Is adopted. With the additional power or link budget available from these laser sources, loose alignment tolerances increase overall bandwidth to levels previously unattainable in a cost-effective manner, small, fast, and cost-effective. It can bring about dramatic design changes that allow for good, free space simplex or double single or parallel channels.</p><p num="0007"> Optical communications can also provide benefits to data centers. Many of the obstacles to line-of-sight communications, such as particulate matter in the atmosphere, are minimal in a data center environment. An increase in beam alignment accuracy can also be achieved. The embodiments of the invention described herein can utilize these factors to achieve unprecedented bandwidth values at a reasonable cost. In addition, using optical communications dramatically reduces the amount of cables required in a data center, reducing complexity and maintenance costs.</p>

<figref num="1">A single, channel depicts a pair of transmitters and receivers.</figref><figref num="2">It depicts a pair of matching multi-channel transmitters and receivers.</figref><figref num="3">It depicts a 32-bit emitter chip that includes both an emitter and a detector.</figref><figref num="4">It depicts the placement of subclusters that are useful for intensity modulation.</figref><figref num="5">Draw a free space optical switch in a data center rack.</figref><figref num="6">Draw a structure in which an optical switch or transmitter / receiver in free space is mounted on the surface.</figref><figref num="7">An embodiment that employs frequency, angle, and spatial multiplexing is depicted.</figref><figref num="8">An embodiment of an optical switch in a high-speed free space is drawn.</figref>

(Array cluster) High power and high speed freedom on the corresponding array of receiver detectors in a symmetrical way to facilitate a pair of inwardly facing devices to provide full bidirectional communication. Embodiments are described that include a single, dual, one-dimensional or two-dimensional array of small clusters of spatial laser beams (such as those described in related applications). The clusters may be connected in parallel or divided into binary weighted subgroups that are driven individually and in parallel. In either case, the beams from each individual cluster may be mixed with a holographic light diffusing element that extends the beam flux to the collimator lens for transfer to the condensing lens of the respective corresponding detector. Good. This also facilitates a board-to-board "daisy chain" scheme to enable bus-like data structures shared by all boards.

Figure 1 depicts a single, one-channel transmitter / receiver pair from Figure 2 or any similarly designed system. The emitter may be on a chip from an epitaxially grown GaAs wafer processed according to the concept described in US Pat. No. 7,949,024, in which clusters of two or more VCSEL elements (100) are electrically parallel. It may be connected to or driven by a single high speed driver. In one embodiment, the beam from each of the elements affects the surface of the light diffusing element (102), such as a holographic light diffuser. Such diffusers have the advantage of providing excellent optical power uniformity that outweighs the resulting beam diffusion and high transmission efficiency. The light cones diffused from each of the VCSEL elements affect the rear surface of the lens (104), the focal length of which is equal to that of the distance from the diffuse surface to the main surface of the lens. The beam that emerges from the lens is a combination of beams from individual elements that is not coherent and reduces the speckle effect of the laser. The bundle appears in front of the receiver as a single, semi-collimator disc of light (106). If the disc has a sufficient diameter compared to its optical power, treating the bundle as a "dispersed light source" may easily meet laser safety standards.

The size of this bundle at the receiver's condenser lens (108) may underfill, match, or match the size of the condenser lens, depending on the amount of translational tolerance desired. , May be overfilled.

As this bundle affects the condenser lens of the receiver, it focuses on a small area behind the surface of the detector (110) so as to form a circle of confusion (112) on the surface of the detector. This allows a certain amount of tilt or translational tolerance of the receiver with respect to the optical axis of the transmitter while still providing sufficient optical power to the surface of the detector to satisfy the link budget. Thereby, a trade-off can be made between the angular tolerance and the optical power to provide sufficient link budget energy within the circle of confusion.

Figure 2 depicts a pair of aligned multi-channel transmitters and receivers in a configuration, thereby between two or more detector / amplifier pair elements (202) on a common board (204). A single GaAs emitter chip (200) is located. The same emitter array and detector / amplifier array (206) with a second common substrate (208) are on the optical axis but facing each other at a distance from the first substrate (204). doing. A symmetric system of transmitters and receivers may be constructed in this way to provide simultaneous multi-channel communication between two symmetric facing substrates.

The emitter chip (200) consists of two or more high speed VCSEL clusters, each cluster driven by its own high speed current control circuit. As already illustrated in Figure 1, the lens of the transmitter is located behind the diffuser so that each different cluster is itself semi-parallel with different angles from the beam bundles from the other clusters. You will be able to create beam bundles. Thus, each bundle may be directed to a matching receiver condenser lens around the corresponding receiver substrate. The receiver lens around the board is offset towards the center of the board to accept the fact that the transmitter beam arrives near the center of the corresponding board and is not parallel to the optical axis between the boards. ing. The framework may be extended to two dimensions, whereby the emitter chip is a two-dimensional array of m × n clusters and the detector / amplifier array is a ring of elements surrounding the emitter chip.

FIG. 3 represents one plan view of a matching pair with a 32-bit configuration. This figure is a pattern that looks like a square donut with clusters formed around a square area in the center (310), with 32 clustered emitter sources arranged on a chip (300). Illustrates the density possible with a transmitter / receiver. Many other configurations can be used as described herein. The illustrated design of the embodiment allows for a small emitter chip at a high density. The emitter chip is located in a square area (310), allowing many detectors and support chips (306) to be placed around the perimeter of the structure. Like a transverse impedance amplifier, the support chips can be aligned on one chip and sent to the detector in many ways known to those skilled in the art of chip layout and integrated hybrid chip layout. Chips can also be connected through holes in the board. Passive or active chip cooling technology can be used with this configuration.

FIG. 3 shows an embodiment of a matching pair of transmitters and receivers, where the array of parallel driven cluster elements summarized in configuration (300) is as described in US Pat. No. 7,949,024. Using laser light source technology, it is easily optically transmitted to the corresponding detector array (304). Each cluster of elements is itself an array of single elements (308), each cluster driven by its own source provides a powerful laser light source, which extends over the location of the detector. The abundant power that allows imaging large "blurred" areas (312) overcomes design problems with link budgets and related tolerances, allowing loose alignment tolerances from substrate to substrate. This may improve the plug-and-play structure for optical communication between circuit boards.

(Intensity modulation) Binary weighted arrays from the techniques described in US Patent Gazette 2011/0148328 A1 by the same applicant, formed within each of the clusters (308) in Figure 3, are imaged with a detector for one cluster. It may be used to encode additional data into each channel by using Amplitude Shift Keying on any reasonable number of subclusters that can reflect the image. Therefore, binary data may be encoded by associating an intensity level with a bit position in the binary data. In one embodiment, the least significant bit ("LSB") is associated with the lowest intensity level. Therefore, the number of bits that can be effectively encoded depends mainly on the receiving end having the LSB signal larger than the smallest link budget and the system of the level identification circuit. The diffuser described above is ideal for evenly distributing the different power levels of each emitter subcluster over the unclear area of the detector.

A VCSEL that has a high frequency response or becomes faster as the current bias increases until it approaches a power rollover may improve the intensity modulation. The VCSEL is preferably fully operational for high speed performance. Intensity modulation is usually achieved by using analog signal levels to achieve different intensity levels than lasers, so lower levels will delay the entire data transmission system. Very high bandwidth with intensity modulation by turning on different arrays up to their highest bias (to reduce distortion) and allowing the selected group to be used to change the intensity level. It will be possible to achieve.

Figure 4 shows individual cluster groups divided into three separate binary weighted subclusters. The laser power of each subcluster is increased by the binary method by doubling the power of a single element or by adjusting the number of elements to increase the power of the subcluster by the binary method. .. Each subcluster is controlled by its own driver source, which is independent of the other subclusters and their drivers. In this way, any combination of bits (subclusters) can be controlled, thereby encoding during each clock pulse. Turning on different subclusters or combinations thereof during the same pulse timing means increasing the intensity level, which can be identified as different encoding levels by the level identification circuit. In this case, group (400) has two VCSEL emitters, group (402) has four VCSEL emitters, and group (404) has eight VCSEL emitters. Each of the emitters in this example is assumed to be close to the same output as the adjacent emitter. The holographic optics (406) distribute the hot region into a uniform mixed beam of intensity, and the lens (408) makes the outputs of any group of arrays operating in a single pulse semi-parallel.

To operate various groups, provided that the detector has a dynamic range to detect the difference between the minimum significant bit (LSB) level and the most significant bit (MSB) intensity level. The following data information can be oriented by beam intensity level. 000-No groups during clock pulses 001-During the clock pulse, only group (400) operates. 010-During the clock pulse, only group (402) operates. 011-Groups (400) and (402) operate during the same clock pulse. Only group (404) operates during 100-clock pulses. 101-Groups (400) and (402) operate during the same clock pulse. 110-Groups (402) and (404) operate during the same clock pulse. 111-Groups (400), (402), and (404) operate during the same clock pulse.

These configurations allow eight separate data codes to be relayed in a single pulse. A preferred embodiment can use this intensity modulation technique, but it is not always necessary. Intensity modulation may also be achieved, for example, without defining a group or by having a one-to-one relationship between the group and the emitter.

Additional embodiments employ wavelength division multiplexing or high density wavelength division multiplexing, where each particular wavelength is often used to transmit data, with bitstring information encoded at each of the pulses of that wavelength. It has other optical multiplexing methods such as orthogonal frequency division multiplexing in which adjacent orthogonal subcarrier signals are used. Devices of this nature have very high data transfer rates. Moreover, frequency doubling is not required in this embodiment. Frequency additions may be employed to obtain longer and eye-safe wavelengths, which is an important factor in the adoption of high power laser propagation.

In another embodiment, any number of wavelengths or beams made by any number of light sources and / or emitter chips are combined by one or more optics to form a highly decomposed data pulse. May be done. Such devices have potentially unlimited number of subgroups that can be defined according to the size, focal length, bit information capacity affected only by the coupling lens, and the limits of the dynamic range of detection.

(Beam steering) In embodiments, an array or cluster of photonic elements behind the lens provides directivity so that the beam emitted from the photonic element hits the lens at various locations. Beam steering may be achieved by selectively activating the elements of the array. This allows for non-mechanical optical alignment, thereby reducing or eliminating excessive gimbal movement, conserving energy and increasing reliability. Beam steering may be achieved by making full use of this technique. One-dimensional and two-dimensional directivity may be achieved by using a one-dimensional array or a two-dimensional array, respectively. In addition, 360 ° coverage may be achieved by adopting multi-sector, eg, hexagonal, spherical, or other three-dimensional arrangements of externally facing arrays.

(Optical switch in data center environment) Additional embodiments take into account optical switch applications that employ a photonic element behind the lens. A narrow beam may be preferred in this embodiment. Optical switches that employ this technique may be placed in the data center rack, so that the narrow beam emitted by the optical switch can reach a defined position within the rack. For example, as seen in FIG. 5, the switch may employ a multi-lens array or may be placed at the top of the rack. In one embodiment, equipment rack (equipment) The rack) (500) may include an optical switch (502) mounted on a surface (504), which may be the ceiling of the rack. The detector (508) may be mounted on the surface (506), which may be inside the rack or on the wall of the rack. The optical switch (502) and the detector (508) allow optical communication between the optical switch (502) and the detector (508), for example, a line of sight between the optical switch (502) and the detector (508). It is placed so that it will not be disturbed by maintaining it. This arrangement allows data shower beams to reach the server connection at defined locations within the rack, allowing a rack without cables. Although the ceiling arrangement is illustrated in Figure 5, the optical switch (502) can be mounted on the floor of the rack, or the rack configuration and which part of the rack provides the clearest line of sight to the beam. It can be mounted anywhere between the floor and ceiling, depending on how you do it.

Other embodiments employ different link geometries such as start, daisy chain, ring, mesh to enable different network topologies. Various factors such as rack placement, airflow, and electrical cables can form line-of-sight obstructions, which can be overcome by choosing the right link geometry. Suitable link geometry may include accurate positioning in both horizontal and vertical dimensions as well as network topology. For example, the device depicted in FIG. 6 may be employed to position the photon array on the rack at an appropriate distance. The optical switch or transmitter / receiver (600) may be mounted in a multi-faceted structure (602). The structure is shown to have six faces, but can be configured to have many different structures with many different faces. The mounting system may support alignment or reorientation of the optical switch or transmitter / receiver (600). For example, as shown at the bottom of FIG. 6, the multi-faceted structure (602) shown from a lateral perspective is racked (604) by a support structure (604) that allows rotation or height adjustment. It may be connected to 606). In addition, the optical switch (600) is formed from an array of subarrays with lenses positioned in front of a group of subarrays to guide the beam from the switch (600) to a specific position where detection is seen. May be done. This may allow for another degree of freedom and allow automatic beam alignment. These subgroups of arrays combined with accurate lens design may cover a small area where automated beam alignment is performed by scanning and receiving methods.

Link budgets are improved by using embodiments within the data center. More precise alignment of the beam narrows the field of view, relaxes power requirements and allows for higher bandwidth. When used in a data center, embodiments can achieve a throughput of at least gigabit / sec over a free space optical communication network.

On the macro scale, not all frequencies propagate in the atmosphere. Fog, dust, rain, or suspended particulate matter such as snow can also interfere with the propagation of light waves. However, in data centers, most frequencies propagate well enough that particulate matter is usually not a problem.

High total bandwidth of 40 Gb / s and above may be achieved via multiplexing. Wavelength division multiplexing allows multiple light wavelengths to be placed in the same optical path. Larger output densities occur when the source configuration is covered. The composition of the source may be tiled to allow angle separation. One embodiment may employ frequency multiplexing, or spatial and angular multiplexing with a single wavelength, as depicted in FIG. For example, the stacked emitters (700) may emit laser power at multiple frequencies, hit the surface of the lens (706), and follow the optical path (708). The unstacked emitters (702) and (704) may emit frequencies of the same or different wavelengths, the output of which hits the lens (706) and follows the optical paths (710) and (712), respectively. Beam size determines both spatial channel density and angular resolution.

(High-speed, high-power array for optical switches) Arrays of VCSEL devices may be used in high speed switch matrices. Information may be supplied to the switch by a single source, such as a 100 Gb / s fiber connection. Each packet of information that needs to be routed is separated by a standard routing chip that routes the packet to the appropriate output channel.

The signal from the output channel may be amplified by a single VCSEL device or a fast current driver connected to an array of VCSEL devices. Each of these is part of a larger array, the size of which defines how many channels are available in the overall switch network.

The output of a laser device is divided by a controlled distance that is imaged through the optics on a pattern of optics such as a lens, an array of lenses, or an optical fiber. The signal of each small element is then put into its single optical or fiber channel. Additional optics that can be deployed within the path of the laser beam for a variety of purposes include diffusers, mirrors, and MEMS devices, to name a few.

Optical channels or Fiber Channel further form an array. The outputs of this array may be bundled in at least two ways. First, the output of the array may be directed to the array of detectors, and the signal of each detector is reconverted into an optical signal, which is injected into the fiber. The fiber may travel to an optical plug, such as a single fiber plug, or to other types of optical terminations. The output of the optical switch array may then be injected directly into the optics or fibers.

The link budget of the configuration must be analyzed to end with a detectable signal. In many cases, there is a lack of power at the start of the system. The extra power will improve the quality of the signal and improve the bit error rate. Therefore, it may be desirable to operate each small element at the highest possible power level.

One embodiment is depicted in FIG. Many VCSEL arrays (802), (804) may be mounted on the surface (800). The output from the VCSEL array (802) passes through the lens (806) and hits the ball lens (808), which may be connected to a path (812) that may be an optical fiber. Similarly, the output from the VCSEL array (804) may pass through the lens (806) and hit the ball lens (808), which may be connected to a different channel (810). The line drawn in FIG. 8 indicates that the output does not mean ray tracing, but rather merely the imaging relationship on position.

(Eye-safe optical network for home and work) Many uses may utilize the formats disclosed herein for free space optical communication. For example, a counter with one or more emitters / receivers embedded in a tabletop and an optical signal with a transparent protective cover, using a transmitter to increase the output density over a large area where the output density is safe for the eyes. Or reach the surface, and as a result, mobile devices with emitters / receivers on or embedded in the device will be placed on the tabletop and automatically connected to 1G to 10G data sources. .. Many other physical configurations are possible, including emitters / receivers that are placed around other locations or surfaces in the room for the same purpose. For example, a laser emitter cluster emits a laser beam onto an optical element formed as a transparent plane with a first surface and a second surface, and the first surface receives the laser beam and the laser beam. Spread the power of the laser beam over the area of the second plane sufficient to reduce the power density of the laser beam to a level that is safe for the eyes.

The embodiment facilitates this main operation because the signal application area is wide and no alignment is required. In addition, there are no wiring connections that can be lost, crushed, or stolen. This type of public access network is much easier without the use of cables. The transmitter can take advantage of THUNDERBOLT technology from INTEL or other protocols. These transmitters require more power for a wider transfer area.

Free-space optical communication over a meter distance requires limited optical power density for eye-safe operation in combination with a wide range of applications to facilitate reception positioning. These requirements limit devices that can provide the high power and speed required in various applications.

A single link transfers data or transfers data over multiple channels, as sufficient bandwidth and output density are possible with high-power, high-speed devices that provide a variety of uses for optical signals. At the same time, it is useful for many users. The array of subgroups behind the lens may be switched so that different beams can be positioned with respect to different regions. A detector array, also known as a Free-Space Optical MIMO (multi-input-multi-output) detector, sends a signal to identify where it needs to be sent and to its specific area. They may be lined up for sending. Multiple subgroup arrays may be functional in any one time series, allowing simultaneous communication links to multiple users. Another embodiment for multiple users may employ wavelength division multiplexing with the same MIMO detector scheme as above, but with multi-wavelength sensitive filtered inputs, both inputs and outputs are different. Sensing and reacting with multiple channels at wavelength Ability may be adopted.

One or more transmitters may be connected to a data source such as fiber optic cable, high speed Ethernet cable, or video source.

The transmitter may consist of a signal input interface, a packaged VCSEL array, a VCSEL driver, and control and amplification electronics, receiver components and electronics that make the transmitter and receiver, and optics that shape the appropriate beam. The systems may all be contained within one common housing. Other components may include optics that distribute the light power density to eye-safe levels and lenses that control the beam diameter and the dispersion of the beam in the free space region.

The transmitter and receiver housings may face each other using a simple plastic molded eyeball socket scheme. The transmitter or transmitter may be directed towards the receiver or transmitter with an adjustable gooseneck lamp type configuration. The transmitter and receiver may each operate from a simple low voltage DC power supply or even from a battery. Transmitters / transmitters may be installed in conjunction with other ceiling-mounted equipment such as lighting, safety / security sensors, camcorders, and security cleats.

A VCSEL array is an addressable array that can be grouped (interlocked) or one or more subarrays, that is, each of one or more elements is independently signal driven and moves across the array. May operate as.

The transmitter may operate in "broadcast" mode, which sends a wide beam of data to one or more receivers.

The transmitter may operate in "beaming" mode, which sends a narrow beam to one or more receivers.

The transmitter may operate in a panning mode in which the subarrays operate sequentially across the VCSEL array, which translates into angular motion in free space.

The system may be configured as a single-way (one-way) link or a full-duplex (two-way) link. In the latter case, there is a VCSEL array and one or more detectors at each end of the optical link.

When a detector array is used, the detector array or sub-array with the strongest signal is the light receiving optics away from the desired transmitter for "handoff" to adjacent transmitters for priority attention. It may be selected to indicate the closest transmitter to accommodate minor misalignment of the angle of.

Since the transmitters may be placed on the ceiling of the workplace and are spaced apart from each other, the transmitters do not have excessive signal overlap and have sufficient coverage in the receiver.

The transmitter may be centered on a vertical support such as a lighting pole, and each transmitter covers an area of space that is itself angular.

The receiver may consist of a condensing device, a detector, an electronic device for amplification, and a suitable output interface.

The condensing device may be an imaging lens or a non-imaging device such as a cylindrical parabolic concentrator.

In the receive detector, the data may be transmitted to a wireless local data transmission method such as optical fiber, Ethernet®, digital video cable, or Wi-Fi.

The receiver may be integrated directly into a digital switch or router that distributes ultra-high bandwidth to many local users over cable or wireless communication.

The receiver may be integrated directly into a component that requires a high bandwidth link.

Due to the high likelihood of bandwidth asymmetry required for downlink-to-uplink, alternative cables or wireless communication uplinks may be used (Wi-Fi, BPL, Ethernet®, etc.).

An optical filter for the wavelength of the transmitter may be used at the receiving end to suppress all wavelengths other than the required wavelength of the transmitter. Multi-wavelength links in a single transmitter may be used.

Polarized VCSEL arrays may be used in conjunction with the receiver's polarizing filter to facilitate elimination of stray reflected signal interference.

The blindfold of the light limiting path in the receiver may remove the stray reflected signal coming in at a different angle than the incident signal.

Many transmitters or transmitters and receivers may be used from different locations, allowing a triangulation positioning grid for best signal reception, or preventing signal blackouts from moving objects.

Window filters may be used to keep the data out of the building. The transmitter or transmitter / receiver may be mounted on a wall, floor, ceiling, or object. The transmitter or transmitter / receiver may be mounted in a very narrow space in the attic, or it may be difficult to access the area between the two points. The transmitter or transmitter / receiver may be attached to a pipe. The transmitter or transmitter / receiver may be attached to the chimney.

(Additional Embodiment) Many additional embodiments are possible. For example, each matching pair of 1 to 1000 pairs of transmitters and receivers may make full use of this technique in any number of layout configurations.

The cluster of elements may be driven as a single channel or may be configured for intensity modulation of each transmitter / receiver channel.

The bottom emitting array is disclosed in US Pat. No. 7,949,024. However, a group of bottom emission arrays, or top emission arrays, or top emission arrays in a flip chip configuration may be used in a multi-element emission device.

A single dual transceiver may be used for substrate-to-board free space optical communication, making full use of the array VCSEL technology mentioned in US Pat. No. 7,949,024.

Instead of just square or rectangular patterns, two-dimensional patterns of most shapes may be formed, and circular optics may work better.

A one-dimensional array with any number of clusters or emitters can be used in configurations other than a single or two-dimensional array.

Any number of patterns can be used to form a cluster and image into an array of the same pattern.

Using a one-dimensional or two-dimensional array as a backplane to select which channel receives the data pulse by selecting different emitters or multiple emitters and sending data to that selected channel. Can be done. If data needs to be sent to a particular channel, a suitable emitter that is already aligned with that channel through the front imaging or projection lens of the entire array is selected. The embodiments can be used over longer distances for high speed data communication applications with suitable imaging optics.

The emitter array (300) mentioned above in FIG. 3 can be formed from the same high speed and high power technology, or, as described above, is flip-chipped onto the substrate instead of being designed on the substrate.

Flip-chip emitters can be of different wavelengths, allowing wavelength division multiplexing capability for the emitter and proper filtering of wavelengths for the detector.

The high power beam can spread over a larger receiving area or overlap with other emitter beams on the area where many detectors are located. A filter covering the detector can be used in conjunction with the multi-wavelength emitter chip to distinguish these overlapping beam signals from emitter clusters of different wavelengths.

The transmitter or transmitter / receiver may be used for high speed long distance applications.

Transmitters or transmitters and receivers may be used between satellites.

The transmitter or transmitter / receiver may be mounted on a utility pole or on the roof of a building.

The transmitter or transmitter / receiver may be mounted on a vertical pole for a better transmission angle.

The receiver portion of the transmitter / receiver is all possible optical power when the bundle of laser beams is configured to overfill the condenser lens and detector in one region to facilitate alignment. Or at least a photocell or "solar cell" may be provided around the detector so that excess laser energy can be recovered.

The photovoltaic device mentioned above may be a power source optically transmitted from a base station or device to transmit output and / or data communications.

In one embodiment, a linear array with any number of columns may be used for data transmission. Any number of columns or all columns may be turned on at the same time, even if the pulses have the same or similar wavelengths. The output position of each column adds a dimensional element to wavelength division multiplexing or high density division multiplexing in relation to the other columns.

In another embodiment, a one-dimensional array is used to generate a single wavelength pulse, which is combined with other sources of the same or different wavelengths. The output pulse intensities of the combined wavelengths are scanned vertically and horizontally, allowing the transmission of data.

In another embodiment, the speed and data transfer rate of the VCSEL array is for flip-chip technology and high speed arrays with waveguides formed around each subarray or device, as described in US Pat. No. 7,949,024. It can be increased by adopting the design of.