WO1997049029A1

WO1997049029A1 - Optical analog signal transmission system

Info

Publication number: WO1997049029A1
Application number: PCT/US1997/009311
Authority: WO
Inventors: Edward F. Breya
Original assignee: Breya Edward F
Priority date: 1996-06-21
Filing date: 1997-06-16
Publication date: 1997-12-24
Also published as: US5724363A

Abstract

A signal transmission system using two optical transmission channels (112, 113, 114, 211, 212 and 122, 123, 124, 221, 222) with switches (111, 213, 121, 223) and calibration circuits (104, 215, 225) operated by a control system (105, 106, 206, 205, 204) such that one channel carries the signal (133) while the other is calibrated. The process makes the responses of the channels more accurate and substantially identical to each other, so that either of the outputs (275, 285) alone, or their average, selected at appropriate times, provides a continuous output (202) representative of the signal. The process does not limit transmission bandwidth, and the calibration cycle repetition rate needs only to be high enough to reduce errors such as gain and offset drift caused by temperature changes and mechanical stress. It is used advantageously as an input device for an oscilloscope, to make galvanically-isolated, wide-bandwidth measurements.

Description

OPTICAL ANALOG SIGNAL TRANSMISSION SYSTEM

Technical Field

The present invention relates to a method and system for transmission of wide-bandwidth analog electrical signals by optical radiation, specifically to an optical signal transmission system using a plurality of optical transmission channels and an automatic calibration system. Background Art

Optical fiber transmission systems are used for high-speed, long-distance signal transmission, for galvanic isolation, and for high immunity to electromagnetic interference. Amplitude modulation (AM) is commonly used for optical transmission of analog and digital signals, and generally provides the highest transmission bandwidth, or utilization of channel information capacity, compared to other modulation techniques. In an optical AM system, a transmitter converts an electrical input signal to an optical signal with optical power proportional to the amplitude of the electrical input signal. An optical fiber carries the optical signal to a receiver, which converts the optical signal to an electrical output signal proportional to the optical power and hence, proportional to the electrical input signal. Although the electronic portions of such systems can be designed to be very accurate, the electro-optic elements and optical portions can cause significant errors in transmission, such as DC offset error, gain error, and non linearity, all of which can change or drift with time, temperature, and other factors. In many applications these errors can limit performance, so it is desirable to reduce them or compensate for them.

The prior art provides a number of solutions to these problems. For digital transmission the electrical signals can be AC coupled to eliminate DC offset errors, and an automatic gain control (AGC) can adjust the receiver gain by measuring the received peak-to-peak digital signal levels, as long as a signal is being received. This type of AGC is exemplified by Muoi U.S. Pat. No. 4,415,803.

Alternatively, the transmitted signal itself may contain encoded calibration information. Summerhayes U.S. Pat. No. 4,070,572 teaches the use of an added DC reference signal to correct AC gain in an AC coupled optical fiber transmission system. Crochet et al U.S. Pat. No. 4,249,264 shows a wide-bandwidth optical fiber transmission system that uses a low-frequency gain reference signal which is filtered out at the receiver and used to adjust the system gain. Arita et al U.S. Pat. No. 4,742,575 discloses a low-speed optical fiber transmission system that uses relatively narrow reference pulses to replace the transmitted signal periodically. At the receiver the pulses are detected and sampled to determine the required gain correction, and are also filtered out of the output signal. Little et al U.S. Pat. No. 5,267,071 uses an added RF pilot signal as an amplitude reference in an optical fiber RF distribution system.

System linearity can be improved electronically, as taught by Toms U.S. Pat. No. 5,077,619, or optically as in Jeffers U.S. Pat. No. 5,126,871, which reduces high-order distortion by transmitting differential optical signals. Laser transmitters with stabilizing feedback arc sometimes used for higher accuracy, stability, and linearity, but fiber coupling efficiency variations and detector characteristics still cause errors at the receiver. Adolfsson et al U.S. Pat. No. 4,290,146, Nelson U.S. Pat. No. 5,162,935, and Gross U.S. Pat. No. 5,453,866 teach embodiments that reduce overall DC offset error by adjusting the DC operating point of the electrical to optical transmitter by using feedback from the receiver back to the transmitter via a second optical fiber transmission channel.

Adolfsson et al U.S. Pat. No. 4,316,141 teaches the use of multiple optical fibers to transmit the signal with a reference signal encoded in a combination of the optical signals, such as their difference or ratio, which is decoded at the receiver during a calibration mode to regulate the system gain and offset. Gross U.S. Pat. No. 5,453,866 also shows automatic gain control with gain adjusted according to the measured response of the optical transmission channel to a calibration signal applied to the input of the transmitter.

Similar calibration systems for other devices are known in the art of electronic instrument design, exemplified by Bohler U.S. Pat. No. 3,711,774, Rode et al U.S. Pat. No. 4,162,531, Nickel et al U.S. Pat. No. 4,200,933, Murooka U.S. Pat. No. 4,364,027, Bristol U.S. Pat. No. 4,553,091, Kannari U.S. Pat. No. 4,799,008, and Eccleston U.S. Pat. No. 4,859,936.

Also in the art of electronic instrumentation is a "Self-Calibrating Signal-Conditioning Amplifier" (NASA Tech Briefs, February 1997, page 48; Reference no. KSC-11750, related to CIP of U.S. patent application 08/233,583, Medelius et al) for transducers, which uses two multiplexers and two programmable-gain amplifiers feeding an analog-to-digital (A/D) converter, in conjunction with digital signal processing, arranged so that one amplifier carries the transducer signal while the other is being calibrated. The resulting system digital output is converted back into an analog signal and also fed back to the input multiplexers in order to calibrate the output stages. The amplifiers are intentionally made very slow in response in order to minimize noise, but the system avoids gaps in the data that would occur during calibration by switching the A/D converter between the two amplifiers because there are always valid data samples available from one amplifier or the other.

Some wide-bandwidth optical fiber analog signal transmission systems that have been on the market for some time (ca. 1989 or earlier), such as the EMC4-1G from Opcom Research (UK), the FOL S series from Electro Optic Developments, Ltd. (UK), and the OP- 300-2 from NanoFast (Chicago, IL) appear to have a calibration mode wherein a built-in calibration generator that produces a signal of known amplitude is switched into the input, the resulting output signal is measured or displayed, and the transmission characteristics are adjusted to the desired accuracy. As the transmission characteristics drift it becomes necessary to calibrate the system again. A trait of many of the optical fiber signal transmission systems is that there is an operate mode in which measurements can be made, and a calibrate mode, automatically or manually invoked, wherein the system is occupied with calibration and is thus unable to make a measurement or transmit a signal. For DC and relatively low-speed signals an alternative is to use some form of modulation that 80 allows the signal to be represented digitally. The resulting digital data can then be transmitted through an optical transmission channel relatively easily. The data may then be used by the receiving equipment directly or converted back to an analog signal. Optical signal transmission systems using frequency modulation (FM), pulse-width modulation (PWM), sigma-delta modulation, voltage-to-frequency (V/F) conversion, and A/D conversion with serial data 85 transmission are known in the art and can provide very high resolution and accuracy at low bandwidth.

It is apparent that the existing art provides various transmission performance capabilities for different applications, but none appear to provide the combination of high accuracy, high bandwidth, and DC coupling, without the need to frequently interrupt measurements to adjust or 90 verify the transmission characteristics. This combination is desirable for general purpose use such as a galvanically-isolated input probe for a wide-bandwidth oscilloscope or other instrumentation.

It is therefore an object of the present invention to reduce transmission errors in an amplitude modulated optical signal transmission system by using an automatic calibration process. It is 95 another object of the invention to provide wide- bandwidth, DC coupled, continuous, real-time signal transmission. Disclosure of Invention

An amplitude-modulated optical analog signal transmission system with an automatic calibration process is disclosed. The system provides accurate, continuous, real-time signal 100 transmission by using two or more substantially identical optical transmission channels to carry the transmission signal, with their inputs and outputs automatically switched such that one channel carries the signal while the other is being calibrated. At least one of the channels is carrying the signal at any time, and the switching is performed such that there is no apparent perturbation of the signal caused by the calibration process. The process makes the responses of 105 the channels more accurate and substantially identical to each other, so that the output of either channel alone, or the average of their outputs, selected at appropriate times, provides a continuous output signal representative of the transmission signal. The calibration cycle repetition rate needs only to be high enough compared to the rate of change of the system errors such as gain and offset drift caused by temperature changes and mechanical stress, to provide the desired 110 error-correction bandwidth, which can be much lower than the system transmission bandwidth. Calibration is performed off-line for each channel, so it does not limit the bandwidth of the channel that is carrying the transmission signal at any given time. Brief Description of Drawings

Fig. 1 is the overall block diagram of the best mode for carrying out the invention to make an 115 optical analog signal transmission system using optical fibers.

Fig. 2 shows some of the signals and activities of the calibration cycle for offset and gain correction associated with the description of Fig. 1. Fig. 3 shows more detail of the offset and gain correction block for the first optical transmission channel.

120 Fig. 4 shows details of a linearity correction block for the first optical transmission channel.

Fig. 5 shows some of the signals related to the calibration cycle for linearity correction associated with the description of Fig. 4.

Fig. 6 shows some of the signals associated with auxiliary data transmission on the first optical transmission channel.

125 Fig. 7 shows the use of wavelength division multiplexing (WDM) to convey the optical signals of both optical transmission channels through a single optical fiber.

Fig. 8 shows the use of wavelength division multiplexing (WDM) to convey the optical signals of both optical transmission channels and the control channel through a single optical fiber.

130 Best Mode for Carrying Out the Invention

In the preferred embodiment of Fig. 1 an input signal 1 with respect to input common terminal 102 is applied to input terminal 101 of transmitter 100 and amplified by an input amplifier 103 to provide a transmission signal 133 which is passed by input switches 111 and 121 to two parallel, substantially identical first and second optical transmission channels,

135 respectively. The first channel comprises driver 112, light emitting diode (LED) 113, optical fiber 114, optical detector 211, and amplifier system 212. The second channel comprises driver 122, light emitting diode 123, optical fiber 124, optical detector 221, and amplifier system 222.

Switches 111 and 121 are separately controlled by control logic 105, and can connect the

140 driver input of the first channel and the driver input of the second channel, respectively, to the transmission signal 133, or to a calibration signal 134 from a calibration generator 104, so that each of the driver inputs can be proportional to either the transmission signal or the calibration signal.

In the receiver 200 a first received output signal 275 from amplifier system 212, and a

145 second received output signal 285 from amplifier system 222, are connected to an output switching circuit comprising switches 213 and 223, resistors 214 and 224, and an output amplifier 201, arranged so that there is an output signal 202 that can be proportional to the received output signal of either one of the channels, or proportional to the average of the received output signals of the channels. The output amplifier 201 produces the output signal 202 with

150 respect to receiver common 203. The output signal 202 is representative of the input signal 1 but may have different scaling or polarity, for example, depending on the overall system requirements, and a time delay due to the optical fiber and amplifier delays. There is no limit to the common-mode voltage between the transmitter input signal and the receiver output signal because the transmitter and receiver are galvanically isolated by the optical fibers, which are

155 electrically non-conductive. Offset and gain errors in the channels are reduced by a first error correction circuit 215 thai measures the response of the first channel to the calibration signal, and adjusts its characteristics to reduce its transmission errors, and a second error correction circuit 225 that does the same for the second channel.

160 A control system comprising calibration cycle control logic 204, a light emitting diode 205, an optical fiber 206, an optical detector 106, and control logic 105 controls the timing and operation of the calibration generator, the input switches, the error correction circuits, and the output switching circuit.

The calibration cycle control logic 204 initiates system calibration cycles by sending serial

165 digital control signals al regular intervals to control logic 105, which decodes the serial digital control signals. Because the calibration cycle is predetermined, and the receiver initiates it, the logical activities of the transmitter and receiver are inherently synchronized.

When a calibration cycle starts, the calibration cycle control logic 204 via first select signal 273 opens switch 213, disconnecting the received output of the first channel. The output

170 amplifier 201 is driven only by the signal from resistor 224, so the output signal is proportional to the received output of the second channel. Control logic 105 then changes the state of switch 111, disconnecting the input of driver 112 from the transmission signal 133 from input amplifier 103, and connecting it to the calibration generator 104, which produces a number of discrete levels in a prescribed sequence. Each calibration level is transmitted through the first

175 channel. During calibration the first error correction circuit 215 measures the response the of first channel to the calibration signal in order to adjust its transmission characteristics to reduce its transmission errors. The drive input of the second channel is still connected to the transmission signal from amplifier 103, and the second received output signal 285 is still connected through switch 223 and resistor 224 to the output amplifier 201. After calibration of the first channel is

180 completed, the control logic 105 in the transmitter disconnects driver 112 from the calibration generator 104 and connects it back to the transmission signal from input amplifier 103 via switch 111. The calibration cycle control logic 204 then turns on switch 213 to again connect the first channel to the output amplifier. For a period of time the output amplifier 201 is driven by both channels via switches 213 and 223 and resistors 214 and 224, so the output signal

185 202 is proportional to the average of the received output signals.

The calibration cycle control logic 204 via second select signal 283 then opens switch 223, disconnecting the received output of the second channel. The output amplifier 201 is driven only by the signal from resistor 214, so the output signal is proportional to the received output of the first channel. Control logic 105 then changes the state of switch 121, disconnecting the input of

190 driver 122 from the transmission signal, and connecting it to the calibration generator 104. Each calibration level is then transmitted through the second channel while the first channel carries the transmission signal. After calibration of the second channel is completed, the control logic 105 disconnects driver 122 from the calibration generator 104 and connects it back to the transmission signal via switch 121. The calibration cycle control logic 204 then turns on switch 195 223 to again connect the second channel to the output amplifier, so the output signal is again proportional to the average of the received output signals.

Fig. 2 shows the salient signals associated with the calibration process of the preferred embodiment of the invention. The input signal 1 can be anything but is shown as a simple sine wave for this example. There is no need for any time correlation between the input signal and the

200 calibration process activity. The calibration signal 134 levels comprise zero, plus full span and minus full span, or they could be any appropriate levels, depending on the overall requirements of the system. The signals transmitted through the channels comprise the transmission signal multiplexed with the calibration signal, with the channels alternating so that at least one of them is carrying the transmission signal at any time. In the receiver the received output of each channel is

205 sampled at limes corresponding to the levels of the calibration signal, according to strobe signals from the calibration cycle control logic 204. The sampled values are used to adjust the DC offset and gain of each channel. The first and second received outputs 275 and 285 arc selected according to first and second select signals 273 and 283, respectively, to provide the output signal 202 that is representative (except for scaling, polarity and time delay) of the transmission

210 signal, and thus representative of the input signal, with no apparent perturbation from the calibration process.

Fig. 3 shows the amplifier system and error correction circuit 212 and 215 for the first channel in more detail. During the time that the calibration generator signal is at zero, strobe signal 270 actuates sample and hold circuit 234 which samples and stores the offset error at the

215 output of amplifier 233 and applies it to integrator 235, which integrates the error to reduce the DC offset by adding a correction signal to the detector signal at summing circuit 230. Likewise, sample and hold circuits 236 and 237 sample in response to strobe signals 271 and 272 and store the plus full span and minus full span values, respectively, at the appropriate times. Summing circuit 238 takes their difference, which represents the peak-to-peak full span

220 response of the channel, and compares it to a reference voltage 240. The resulting error is integrated by integrator 239 which adjusts a gain control circuit 232 to correct the channel gain by making the peak-to-peak full span response equal to the reference voltage. The second channel amplifier system and error correction circuit are substantially the same as those for the first channel. Fig. 2 further shows the timing of the corresponding second channel strobe signals

225 280, 281 and 282 with respect to the second received output signal and the first channel circuit activity.

The control system provides a plurality of states comprising: a state wherein ihe driver input of the first channel is proportional to the transmission signal, the driver input of the second channel is proportional to the calibration signal, the second error correction circuit is measuring

230 the response of the second channel to the calibration signal, in order to adjust the characteristics of the second channel to reduce its transmission errors, while the output signal is proportional to the received output signal of the first channel; a state wherein the driver input of the second channel is proportional to the transmission signal, the driver input of the first channel is proportional to the calibration signal, the first error correction circuit is measuring the response of

235 the first channel to the calibration signal, in order to adjust the characteristics of the first channel to reduce its transmission errors, while the output signal is proportional to the received output signal of the second channel; and a state wherein the driver input of each of the channels is proportional to the transmission signal, while the output signal is proportional to the average of the received output signals of the channels.

240 The control system repetitively controls the plurality of states in a prescribed sequence, which is repeated indefinitely, allowing one channel to carry the transmission signal while the other is being calibrated. In the steady state the calibration process reduces the DC offset and gain errors of both channels, making their individual responses more accurate and substantially identical, so that the received output of either channel alone or the average of their received outputs, selected at

245 the appropriate times, provides accurate, continuous representation of the transmission signal. The outputs and inputs of the channels are switched in such a way that there is no apparent interruption of the transmission signal through the overall system. During the calibration cycle the received output of each channel is disconnected before its driver input is transferred from the transmission signal to the calibration signal, and its driver input is reconnected to the

250 transmission signal before its received output is reconnected to the output amplifier, with enough time allowed for digital control time delays, switching time delays and device settling, so that switching transients do not appear in the output signal. The period of overlap when both channels are connected to the output amplifier may be reduced to zero in theory, but it would require essentially perfect switches with zero time delay in their control. The overlap period with

255 averaging of the received output signals allows practical switching devices to make smooth transitions between the channels with minimal effect on the output signal.

The averaging of the received output signals is preferably accomplished by adding them through resistors of equal value. The resistance is sufficiently low in value to provide fast system response, but not so low that the received output signal amplifiers will be overloaded or become

260 unstable due to driving against each other. Although the received output signals are made to be substantially identical by the calibration process, there may still be slight differences which could cause large currents to flow between the amplifiers during the time that both output switches are closed. Also, when error effects introduced into the system are beyond the bandwidth or dynamic range of the error correction system, the channel outputs may be quite different from

265 each other. For example, if the optical fibers are moved suddenly, the transient change in transmission characteristics will affect the output signal until the responses of the optical transmission channels are corrected by the calibration process. The same occurs when the system is powered up until all of the error correction circuit control loops have stabilized.

The calibration cycle repetition rate needs only to be high enough compared to the rate of

270 change of the system errors such as gain and offset drift caused by temperature changes and mechanical stress, to provide the desired error-correction bandwidth, which can be much lower than the system transmission bandwidth. Calibration is performed off-line for each channel, so it does not limit the bandwidth of the channel that is carrying the transmission signal at any given time. For example, the calibration cycle rate may be several hundred hertz, while the

275 transmission bandwidth may be several hundred megahertz.

It does not matter in what order the channels are calibrated nor does it matter in what order the various calibration levels are transmitted, as long as the receiver and transmitter circuits process the information properly. It is also possible to transmit the different calibration levels in separate calibration cycles rather than to transmit the entire plurality of calibration levels in one

280 calibration cycle. It is also possible to instead put the calibration cycle control inside the transmitter, and have it send digital control signals to the receiver circuits via an additional optical transmission channel, to accomplish the same function.

The calibration system can be adapted to improve the linearity of the optical transmission channels as in Fig. 4, which shows a linearity correction block for the first channel. The linearity

285 correction block for the second channel is substantially the same. In this embodiment the calibration signal 134 comprises a plurality of discrete levels. These levels are sampled at the receiver by a plurality of sample and hold circuits such as 251 through 254, and compared to prescribed corresponding levels that would be transmitted by a perfectly linear optical transmission channel, and the resulting error information is used to adjust the gain and linearize

290 the transfer function with the adjustable gain and distortion circuit 250, which may be implemented with techniques known in the art. Fig. 5 shows some of the signals associated with the embodiment of Fig. 4, using an example with eight levels strobed at appropriate times by eight strobe signals 299.

In the preferred embodiment of Fig. 1 the calibration control logic 204 sends serial digital

295 control data to the control logic 105 primarily to synchronize the calibration cycle activities of the receiver and transmitter. Additional data can be sent and decoded to control other functions of the transmitter such as input ranging and signal coupling (not shown). It is also useful to send data from the transmitter to the receiver, for example, to confirm function settings, indicate battery levels, or diagnose problems in the transmitter. This data can be sent from the transmitter to the

300 receiver during the calibration periods, thereby avoiding the need for an additional optical transmission channel for this purpose. To do this, the calibration signal 134 further comprises additional time slots to convey information, during which the signal levels vaiy according to the data to be sent to the receiver, and can represent analog or digital data because each channel is calibrated. During each calibration cycle, additional strobe signals in the receiver corresponding

305 to the data time slots operate sample and hold circuits to collect the data from each of the channels. This is a type of time division multiplexing (TDM), so the time allowed for each calibration level and data level becomes shorter as more are used, for a given calibration duty factor and repetition rate. Conversely, the time per level could be lengthened by reducing the repetition rate, but the error-correction bandwidth would be reduced. Fig. 6 shows an example

310 wherein the first received output signal 275 includes three calibration reference levels followed by auxiliary data consisting of one analog level and four binary data bits, for a total of eight time slots, sampled according to the strobe signals 299. The channels can transmit different auxiliary data if necessary, by appropriately controlling the calibration generator.

Each of the channels does not necessarily have to use a separate optical fiber. Wavelength

315 division multiplexing (WDM) is known in the art and can be used to provide a plurality of channels on a single optical fiber. In Fig. 7 light emitting diodes 113 and 123 emit light at different wavelengths λa and λb, respectively. The light from the LEDs is combined in optical coupler 301 and transmitted through optical fiber 302, to optical coupler 303, which splits the optical signal to optical detectors 211 and 221, which are substantially responsive only to λa

320 and λb, respectively. The wavelength discrimination means of the optical detectors can be provided by known art and may be intrinsic to the detectors or to the optical couplers, or provided by additional optical elements not shown. Fig. 8 further shows the elimination of the separate control optical fiber (part 206 in Fig. 1) by using a third wavelength λc to carry the serial digital control signal from the receiver to the transmitter through optical fiber 302. LED

325 205 emits light at wavelength λc which is coupled into optical fiber 302 by optical coupler 303 and separated by optical coupler 301 to optical detector 106 which is substantially responsive only to λc. It is also possible to convey the optical radiation from the LEDs to the optical detectors in other ways, such as transmission through space, using wavelength division multiplexing or physical barriers to prevent crosstalk between the optical transmission channels.

330 The light emitting diodes, optical fibers, optical detectors, and related apparatus used for the foregoing description of the various embodiments may be replaced by other known apparatus that emit, convey, or detect optical radiation. The embodiments described herein can also be extended to a system where more than two optical transmission channels are used to carry the transmission signal, with each channel carrying the signal for a time and being calibrated for a

335 time, and being switched at appropriate times. The means for calibraϋon cycle control and measurement and adjustment of the optical transmission channel characteristics can be provided in many ways, and located in the transmitter or the receiver, or distributed between them in many ways. However, it is generally preferable to minimize the size, complexity, and power dissipation of the transmitter by placing as much of the circuitry as possible in the receiver.

340 The foregoing description of the preferred embodiment of the invention has been presented for the purposes of illustration and description. It is not intended to be exhaustive or to limit the invention to the precise form disclosed. Many modifications and variations are possible in light of the above teaching. It is intended that the scope of the invention be limited not by this detailed description, but rather by the claims appended hereto.

345 Industrial Applicability

The system is advantageously used as an input probe device for oscilloscopes or other instrumentation, to make galvanically-isolated, wide-bandwidth measurements, particularly in power conversion applications.

Claims

Claims 350 1. An apparatus of the type used to transmit an electrical input signal optically from a transmitter to a receiver, which provides an electrical output signal representative of said electrical input signal, said apparatus characterized in that it comprises: a) a first optical transmission channel comprising means to convert an electrical signal applied to a drive input into optical radiation, means to convey the optical radiation to a receiver

355 comprising a detector to convert the received optical radiation into an electrical detector signal, and an amplifier system to amplify the detector signal to produce a received output signal representative of the electrical drive signal; b) a second optical transmission channel comprising means to convert an electrical signal applied to a drive input into optical radiation, means to convey the optical radiation to a

360 receiver comprising a detector to convert the received optical radiation into an electrical detector signal, and an amplifier system to amplify the detector signal to produce a received output signal representative of the electrical drive signal; c) input circuit means to accept said electrical input signal d) output circuit means to provide said electrical output signal

365 c) first calibration means to calibrate said first optical transmission channel; f) second calibration means lo calibrate said second optical transmission channel; g) control means to control and connect the optical transmission channels, said input circuit means, said output circuit means, and first and second calibration means in a plurality of arrangements comprising: an arrangement wherein the drive inputs of the optical

370 transmission channels are proportional to said electrical input signal, and said electrical output signal is proportional to the average of the received output signals of the optical transmission channels; an arrangement wherein the drive input of said first optical transmission channel is proportional to said electrical input signal, and said electrical output signal is proportional to the received output signal of said first optical transmission 375 channel, while said second optical transmission channel is calibrated by said means to calibrate said second optical transmission channel; and an arrangement wherein the drive input of said second optical transmission channel is proportional to said electrical input signal, and said electrical output signal is proportional to the received output signal of said second optical transmission channel, while said first optical transmission channel is 380 calibrated by said means to calibrate said first optical transmission channel; and h) means to repetitively operate said control means, the object being to make said electrical output signal an accurate, continuous representation of said electrical input signal.

2. An apparatus of the type used to convey a transmission signal optically and provide an electrical output signal representative of said transmission signal, said apparatus characterized in 385 that it comprises: a) a first optical transmission channel comprising means to convert an electrical signal applied to a drive input into optical radiation, means to convey the optical radiation to a receiver comprising a detector to convert the received optical radiation into an elecuical detector signal, and an amplifier system to amplify the detector signal to produce a received output 390 signal representative of the electrical drive signal; b) a second optical transmission channel comprising means to convert an electrical signal applied to a drive input into optical radiation, means to convey the optical radiation to a receiver comprising a detector to convert the received optical radiation into an electrical detector signal, and an amplifier system to amplify the detector signal to produce a

395 received output signal representative of the electrical drive signal; c) a calibration generator to generate a calibration signal; d) a first and a second input switching means arranged so that the drive input of each of the optical transmission channels can be proportional to either said transmission signal or said calibration signal;

400 e) a first error correction circuit comprising means to measure the response of said first optical transmission channel to said calibration signal, in order to adjust its characteristics to reduce its transmission errors, and means to adjust its transmission characteristics; and a second error correction circuit comprising means to measure the response of said second optical transmission channel to said calibration signal, in order to adjust its characteristics

405 to reduce its transmission errors, and means to adjust its transmission characteristics; f) an output switching circuit means arranged so that there is an electrical output signal that can be proportional to any one of a plurality of signals comprising: the received output signal of said first optical transmission channel, the received output signal of said second optical transmission channel, and the average of the received output signals of the optical

410 transmission channels; and g) a control system means which controls the operation of said calibration generator, the first and second input switching means, the first and second error correction circuits, and the output switching circuit means in such a way that there can exist a plurality of states comprising:

415 i) a state wherein the drive input of said first optical transmission channel is proportional to said transmission signal, the drive input of said second optical transmission channel is proportional to said calibration signal, said second error correction circuit is measuring the response of said second optical transmission channel to said calibration signal, in order to adjust the characteristics of said second optical

420 transmission channel to reduce its transmission errors, while said electrical output signal is proportional to the received output signal of said first optical transmission channel; ii) a state wherein the drive input of said second optical transmission channel is proportional to said transmission signal, the drive input of said first optical

425 transmission channel is proportional to said calibration signal, said first error correction circuit is measuring the response of said first optical transmission channel to said calibration signal, in order to adjust the characteristics of said first optical transmission channel to reduce its transmission errors, while said electrical output signal is proportional to the received output signal of said second optical transmission 430 channel, and; iii) a state wherein the drive input of each of the optical transmission channels is proportional to said transmission signal, while said output signal is proportional to the average of the received output signals of the optical transmission channels.

3. The apparatus of claim 2 wherein said control system repetitively controls said plurality of 435 states in a prescribed sequence.

4. The apparatus of claim 2 wherein the first and second optical transmission channels operate at separate optical wavelengths sufficiently different from each other that the optical transmission channels can utilize a single optical transmission medium as the means to convey their optical radiation to their receivers.

440 5. A method of automatically reducing errors in an apparatus of the type used for optical transmission of an electrical signal, said apparatus comprising: a first optical transmission channel comprising means to convert an electrical drive signal applied to a drive input into optical radiation, means to convey the optical radiation to a receiver comprising a detector to convert the received optical radiation into an electrical detector signal, and an amplifier system to amplify the

445 detector signal to produce a received output signal represenlative of the electrical drive signal; a second optical transmission channel comprising means lo convert an electrical drive signal applied to a drive input into optical radiation, means to convey the optical radiation to a receiver comprising a detector to convert the received optical radiation into an electrical detector signal, and an amplifier system to amplify the detector signal to produce a received output signal

450 representative of the electrical drive signal; a calibraϋon generator to generate a calibration signal; a first and a second input switching means arranged so that the drive input of each of the optical transmission channels can be proportional to either a transmission signal or said calibration signal; a first error correcϋon circuit comprising means to measure the response of said first optical transmission channel to said calibration signal, in order to adjust its characterisϋcs to

455 reduce its transmission errors, and means to adjust its transmission characteristics; a second error correction circuit comprising means to measure the response of said second optical transmission channel to said calibration signal, in order to adjust its characteristics to reduce its transmission errors, and means to adjust its transmission characteristics; an output switching circuit means arranged so that there is an output signal that can be proportional to any one of a plurality of

460 signals comprising the received output signal of said first optical transmission channel, the received output signal of said second optical transmission channel, and die average of the received output signals of the first and second optical transmission channels; and a control system which controls the operation of said calibration generator, the first and second input switching means, the first and second error correction circuits, and said output switching circuit; 465 the object being to make said output signal an accurate, continuous representation of said transmission signal; and said method characterized in that it comprises the steps of: a) disconnecting the received output of said first optical transmission channel while die received output of said second optical transmission channel remains connected, so that said output signal is proportional to the received output of said second optical transmission

470 channel; b) disconnecting the drive input of said first optical transmission channel from said transmission signal, and connecting it to said calibration signal; c) measuring the response of said first optical transmission channel to said calibration signal, and adjusting die transmission characteristics of said first optical transmission channel in

475 order to reduce its transmission errors; d) disconnecting the drive input of said first optical transmission channel from said calibration signal, and reconnecting it to said transmission signal; e) reconnecting the received output of said first optical transmission channel so that said output signal is proportional to the average of the received outputs of the optical

480 transmission channels;

0 disconnecting the received output of said second optical transmission channel while the received output of said first optical transmission channel remains connected, so that said output signal is proportional to the received output of said first optical transmission channel; 485 g) disconnecting the drive input of said second optical transmission channel from said transmission signal, and connecting it to said calibration signal; h) measuring the response of said second optical transmission channel to said calibration signal, and adjusting the transmission characteristics of said second opϋcal transmission channel in order to reduce its transmission errors; 490 i) disconnecϋng the drive input of said second optical transmission channel from said calibration signal, and reconnecϋng it to said transmission signal; and j) reconnecting the received output of said second opϋcal transmission channel so that said output signal is proportional to the average of the received outputs of die optical transmission channels. 495 6. The method of claim 5, wherein the steps (a) through (j) are repeated indefinitely.