AU2053300A - Method and apparatus for controlling the ratio of an output signal and an input signal - Google Patents

Description

WO 00/38318 PCT/US99/29679 METHOD AND APPARATUS FOR CONTROLLING THE RATIO OF AN OUTPUT SIGNAL AND AN INPUT SIGNAL This application claims the benefit of U.S. Provisional Patent Application 5 Number 60/113,083, filed December 21, 1998. Background of the Invention This invention relates to a method and apparatus for controlling a ratio of an output signal and an input signal of a device. The invention can be implemented using relatively simple control electronics and is suitable for a 10 wide variety of applications including, for example, electronic control of electrical and optical amplifiers and attenuators. Conventional signal ratio control techniques typically involve monitoring an output signal or an input signal of a device relative to an external reference signal that is set in accordance with a predetermined target ratio. The device is 15 controlled to operate at the target ratio based on an error signal representing a difference between the monitored signal and the reference signal. Another conventional technique involves monitoring both the output signal and the input signal of a device in order to determine a ratio of the two signals. The ratio thus determined is compared with an external reference 20 signal, corresponding to the target ratio, in order to generate an error signal representing the difference between the determined ratio and the target ratio. The device is controlled to operate at the target ratio based on the error signal. Fig. 1 illustrates the conventional technique just described. In Fig. 1, a divider 12 receives signal levels x and y proportional to the input and output WO 00/38318 PCTIUS99/29679 2 signal levels of a controlled device 10, respectively. That is, x = C 1 X and y =

C

2 Y, where X and Y are the input and output signal levels, respectively, and C1 and C2 are proportionality constants depending upon the particular system design (often C1 = C2 ). The signal levels x and y may, for example, represent 5 voltage, current, or power levels of the input and output signals. The ratio G of the output signal level to the input signal level may be represented as follows: G = (Y/X) = K(y/x) . . . (1) where K = C2/C1. Thus, the signal levels x and y are related as follows: 10 y = Gx/K . .. (2) The divider 12 performs a division operation based on the inputs x and y, and outputs a ratio signal proportional to the ratio G: y/x = G/K ... (3) The ratio signal is supplied to a subtractor 14. Also supplied to the 15 subtractor is a specified reference signal level Gp = GT/K, where GT is the target ratio. The subtractor 14 subtracts the measured ratio y/x from the specified ratio Gsp to generate an error signal level E: E = Gsp - y/x = (GT/K) - (G/K) = (GT - G)/K . . . (4) The error E is supplied to a controller 16 which adjusts the device 20 operation in order to bring the error E to zero. In this state, G becomes equal to GT as will be recognized from Equation (4). Disclosure of the Invention The present invention reflects a new and unique approach to the ratio 25 control process that does not require an external reference signal or a division operation as in conventional techniques. Briefly speaking, the invention utilizes first and second signals having levels corresponding to the input signal level and the output signal level of the device, respectively. As will be seen hereinafter, by subtracting one of the signals from the other, an error signal is 30 generated for supply to a controller which, in turn, controls the device operation WO 00/38318 PCTIUS99/29679 3 to bring the error signal level to zero. In this state, the ratio of the input and output signal levels of the device will match a predetermined target ratio. As shown in the illustrative embodiments hereinafter, the first and second signals may be produced, respectively, using first and second amplifier 5 units of known gains. When the error signal level is zero, the ratio of the input and output signal levels matches a predetermined target ratio based on the known gains of the first and second amplifier units. Fig. 2 is a diagram illustrating basic principles of implementation using first and second amplifier units as just described. In Fig. 2, an input-side 10 amplifier unit 23 having a known gain Gi and an output-side amplifier unit 25 having a known gain Go are connected to the input and output, respectively, of a device 20 to be controlled (the connections having been omitted from the drawing for simplicity). The input-side amplifier unit 23 produces a signal having a level Si = Gix, where x is proportional to the input signal level of the 15 device 20. The output-side amplifier unit 25 produces a signal having a level So = Goy, where y is proportional to the output signal level of the device 20. The signals from the input-side and output-side amplifier units are supplied to a subtractor 24, which produces an error signal having a level Serr: Serr = So - Si . . . (5) 20 The error Serr is supplied to a controller 26, which adjusts the operation of device 20 so as to bring the error Serr to zero and to maintain this condition. Substituting for So and Si, Equation (5) can be rewritten as follows: Serr = Goy - Gix . . . (6) Substituting Gx/K for y (see Equation (2)) in Equation (6) yields: 25 Serr = (GoGx/K) - Gix = x{(GoG/K) - Gi} ... (7) Multiplying the right-hand side of Equation (7) by Gi/Gi yields: Serr = Gix{(Go/KGi)G - 1} ... (8) As is evident from Equation (8), Serr will be zero when G = KGi/Go. In 30 other words, the selection of the input-side and output-side amplifier units or, more precisely, their gains establishes a target ratio WO 00/38318 PCT/US99/29679 4 GT = KGi/Go. And controlling device 10 to maintain Serr at zero will thus result in a constant ratio G = GT = KGi/Go. The same results are obtained if the minuend and the subtrahend in Equation (5) are reversed: 5 Serr = Si - So .. . (5') Serr may thus be expressed more generally as follows: Serr =± (So - S) -. (9) As will be appreciated from the foregoing discussion, the present invention does not require an external reference signal representing the target 10 input-output signal ratio and does not require a division to determine the actual operating ratio G. Instead, the control operation is based on a simple subtraction of signals having levels corresponding to the input and output signal levels. By avoiding the need for an external reference signal and a division operation, as are used in conventional techniques, the present 15 invention offers the advantages of simplified control electronics and high control speeds at low cost. Furthermore, the principles of the invention are not limited to application to any particular device or class of devices, or to the use of particular components to produce the signals to be subtracted. It will thus be understood 20 that the applications mentioned herein, as well as the illustrative embodiments, are merely exemplary. The use of amplifier units to produce the signals to be subtracted may be preferred, as in the illustrative embodiments, due to advantages in accommodating devices having low input (and possibly also low output) signal 25 levels, as well as devices that utilize non-electrical signals. For example, transimpedance amplifier units may be employed in applications involving devices that utilize optical signals.

WO 00/38318 PCT/US99/29679 5 Summary of the Invention As mentioned above, the present invention enables ratio control of an input signal and an output signal of a device without the need for an external reference signal or a division operation. 5 According to one of its principal aspects, the invention provides a method of controlling a ratio of an output signal level from a device and an input signal level to the device, comprising: providing a first component connected to an input of the device to produce a first signal having a level corresponding to the input signal level; 10 providing a second component connected to an output of the device to produce a second signal having a level corresponding to the output signal level; and adjusting the ratio based on a difference between the levels of the first and second signals. 15 According to one preferred mode, the device itself is an amplifier unit, the first and second components are first and second amplifier units, and the controlled ratio is a gain of the amplifier unit. In a specific implementation, the device is a pumped fiber-optic amplifier unit, and the controlled ratio is an optical gain of the fiber-optic amplifier unit. 20 The optical gain is adjusted by adjusting pump power of a pump laser of the fiber-optic amplifier unit. The first and second amplifier units are transimpedance amplifier units which are connected by corresponding photodetectors to the input and output of the fiber-optic amplifier unit. (A transimpedance amplifier provides an output voltage signal proportional to an 25 input current signal.) Each of the input signal and the output signal may be a composite signal constituted by a plurality of signals. In such case, the signal levels of the aforementioned first and second signals may correspond to the RMS signal levels of the composite input and output signals, respectively, and the 30 controlled ratio may be an RMS optical gain.

WO 00/38318 PCT/US99/29679 6 According to another of its principal aspects, the present invention provides apparatus for implementing the above method. Yet another aspect of the invention provides a method of controlling a ratio of an output signal level from a device and an input signal level to the 5 device, comprising, producing a first signal having a level corresponding to the input signal level, producing a second signal having a level corresponding to the output signal level, and adjusting the ratio based on a difference between the levels of the first and second signals The aforementioned and additional aspects of the present invention, as 10 well as its various features and advantages, will be more fully understood upon consideration of the following detailed description with reference to the accompanying drawings. Brief Description of the Drawings 15 Fig. 1 is a diagram for explaining a conventional ratio control technique. Fig. 2 is a diagram for explaining the control technique according to the present invention. Fig. 3 is a diagram illustrating a first apparatus according to the invention. 20 Fig. 4 is a diagram illustrating a second apparatus according to the invention. Fig. 5 is a diagram illustrating a third apparatus according to the invention. Fig. 6 is a diagram illustrating a fourth apparatus according to the 25 invention. Fig. 7 is a diagram of the control electronics employed in a test apparatus constructed according to Fig. 4. Fig. 8 is a flow diagram of the control operation in the test apparatus. Fig. 9 is an oscillograph showing temporal dynamics of an optical 30 amplifier unit of the test apparatus without the ratio control circuitry in operation.

WO 00/38318 PCT/US99/29679 74 Figs. 10 and 11 are oscillographs showing temporal dynamics of the optical amplifier with the ratio control circuitry in operation. Detailed Description of the Invention 5 Fig. 3 is a diagram illustrating a first apparatus I according to the present invention. Reference number 30 in the figure represents an electrical device operable with a controllable ratio G of an input signal and an output signal. In the form shown, device 30 is an adjustable-gain amplifier unit including a single amplifier. But, in practice, device 30 may be any electrical 10 device that has a controllable ratio of its input and output signals (e.g., a multi stage amplifier having cascaded amplifiers, a single or multi-stage attenuator, etc.). The input and output signals may be voltage or current signals, for example. A first amplifier unit 33 of known gain Gi, and preferably including a 15 single amplifier, is connected to the input side of the controlled amplifier unit 30. A second amplifier unit 35 of known gain Go, also preferably including a single amplifier, is connected to the output side of the controlled amplifier unit 30. The connections of amplifier units 33 and 35 to the input and output sides of amplifier unit 30 may be direct, as shown, or they may be indirect (e.g., via a 20 current sensor) depending on the requirements of a given implementation. The output signals of the amplifier units 33 and 35 are supplied to a controller 36, which operates, as discussed in more detail below, to adjust the operation of the controlled amplifier unit 30 so as to control the ratio G. In the illustrative implementation, G is of course a gain of the amplifier unit 30. 25 The input-side amplifier unit 33 outputs a signal level Si = Gix, where x is an input signal level to the input-side amplifier unit and is proportional to the input signal level X of the controlled amplifier unit 30. The output-side amplifier unit 35 outputs a signal level So = Goy, where y is an input signal level to the output-side amplifier unit and is proportional to the output signal level Y of the 30 controlled amplifier unit 30.

WO 00/38318 PCT/US99/29679 8 The controller 36 performs a subtraction as previously explained to determine an error value Serr = ± (So - Si), and it outputs a control signal St to the controlled amplifier unit 30 to adjust its gain such that Serr is substantially maintained at zero. Thus, the amplifier unit 30 is maintained at a substantially 5 constant gain G = GTKGi/Go, that is, a target gain proportional to a ratio of the known gains of the input-side and output-side amplifier units 33 and 35. To give a more concrete example, assume that the input and output signals to the controlled amplifier unit are voltage signals having levels Vi and Vo ( = GVi), respectively. Assuming sufficiently high input impedances of the 10 amplifiers 33 and 35, the respective input voltage levels thereof will be Vi and Vo (corresponding to the case C 1 = C2, K=1). The output voltages of the input side and output-side amplifier units may be expressed as follows: Si = GiVi So= GoVo = GoGVi 15 In this case, Gi and Go represent the respective voltage gains of the input-side and output-side amplifier units 33 and 35, and G represents the voltage gain of the controlled amplifier unit 30. The controller 36 subtracts the voltage Si from the voltage So to determine the error Serr: 20 Serr = So - Si = GoGVi - GiVi = Vi(GoG - Gi) = GiVi{(Go/Gi)G - 1} . . . (10) Based on the error thus determined, the controller 36 adjusts the gain of 25 amplifier unit 30 to bring the error Serr to zero and to maintain that state. In that state, the gain G will be equal to the target gain GT = Gi/Go, as will be evident from Equation (10). The controller 36 may utilize any suitable control algorithm for controlling the ratio G based on the error Serr. Proportional-integral (PI) or proportional 30 integral-derivative (PID) control algorithms may be preferred for optimal performance. Both digital and analog controllers may be used. Suitable WO 00/38318 PCTIUS99/29679 9 algorithms for specific applications may be determined by conventional techniques-for example, empirically and/or by computer simulation. For a more complete discussion of P1, PID, and other control techniques, see Koenig, D., Control and Analysis of Noisy Processes, Prentice Hall, 1991 5 (incorporated herein by reference). Fig. 4 illustrates a second apparatus 2 according to the present invention for controlling the optical gain (optical power gain) of an optical amplifier unit 40. In this embodiment, the optical amplifier unit 40 is a single stage (single-coil) pumped fiber-optic amplifier. Such amplifiers are well known 10 in the art and so will not be discussed in detail herein. Briefly, the amplifier unit includes a fiber-optic coil 41 doped with ions of a rare-earth element (e.g., erbium or praseodymium) and a wavelength division-multiplex (WDM) optical coupler 42 that couples an input optical signal with "pump" light from a controller laser source 44. Note that thin lines 15 connecting components in Fig. 4 represent optical fiber connections, whereas bold lines connecting components represent electrical connections. The laser source or pump laser unit 44 operates at an optical wavelength that is outside the wavelength band of the optical input signal to the amplifier but effective to excite the doped ions in the optical fiber coil 41. The optical input signal may 20 be composed of a single optical signal at a predetermined wavelength, or it may be a composite signal composed of a plurality of optical signals at different predetermined wavelengths, as is typical in WDM fiber-optic communication networks. Light of the input signal stimulates the excited ions in the fiber coil 41 to emit additional light of the same wavelength(s), effectively 25 amplifying the input optical signal. The optical power gain G of the amplifier unit 40 depends on the output power of the pump laser unit 44, and therefore can be controlled by adjusting the output power of the pump laser unit. When the input signal is to be a composite signal, the optical amplifier unit 40 may incorporate a gain-flattening 30 filter so that the individual wavelength components of the composite signal experience equal gain. Otherwise, the optical power gain will be the RMS gain WO 00/38318 PCTIUS99/29679 10 for the collective wavelength components. The optical amplifier unit 40 may also include optical isolators (not shown) at its input and output sides. In the arrangement of Fig. 4, a first transimpedance amplifier (TIA) unit 43, preferably including a single amplification stage, has its input connected to 5 an input side of the optical amplifier by way of a photodetector 47 (e.g., photodiode) and an optical tap 48. The optical tap functions to couple a small portion of the incoming optical signal from an input fiber I of the apparatus to a monitor output 48b which is connected to photodetector 47. The remainder of the incoming optical signal propagates via a main output 48a of the tap to the 10 input of the optical amplifier unit 40. A second TIA unit 45 has its input connected to an output side of the optical amplifier unit 40 via a corresponding photodetector 47' and associated optical tap 48'. This optical tap operates to couple a small portion of the output light from the optical amplifier unit to a monitor output 48b' which is connected 15 to photodetector 47'. The remainder of the output light propagates via a main output 48a' of the tap 48' to an output fiber 0 of the apparatus. The coupling ratios of the optical taps 48, 48' may be the same or different. Also, they are not restricted to any particular limit. But, it is generally preferred to use coupling ratios which substantially preserve the input and 20 output signal powers of the apparatus-for example, a coupling ratio of at least 90/10 (meaning that 10% of the light is coupled to the monitor output with 90% of the light propagating to the main output). Returning to Fig. 4, the photodetectors 47, 47' convert the light received from taps 48, 48' into electrical current signals having levels proportional to the 25 amounts of light received via monitor outputs 48b, 48b'. The current signals are thus proportional to the optical power levels of the input and output optical signals of the optical amplifier 40. The TIA units 43 and 45, in turn, produce output voltage signals proportional to their input current signals, and these output signals are supplied to a controller 46. The respective output voltage 30 signals Si and So of the TIA units 43 and 45 may represented as follows: Si = Gi 1 Pi WO 00/38318 PCT/US99/29679 11 So = GoC 2 Po = GoC 2 GPi In the foregoing, Gi and Go respectively represent the transimpedance gains of the input-side TIA unit 43 and the output-side TIA unit 45, Pi and Po respectively represent the input optical signal power level and the output 5 optical signal power level, G represents the optical power gain of the optical amplifier unit, and C1 and C2 represent proportionality constants dependent upon the coupling ratios of the taps 48, 48' and the responsivities of the photodetectors 47, 47'. The controller subtracts one of the two voltage signal levels from the 10 other to obtain an error voltage level Serr: Serr = (So - Si) = (GoC 2 Po - GiC1Pi) = (GoC 2 GPi - GiC 1 Pi) = Pi(GoC 2 G - GiC 1 ) 15 = PiGiC 1 {(GoC2/GiCl) G - 1} . .. (11) Based on the error voltage, the controller 46 generates a control signal to adjust the operation of the pump laser unit 44, and thereby adjust the optical gain of the amplifier unit 40, to bring Serr to zero and maintain that condition. For example, the pump laser unit may incorporate a conventional voltage 20 controlled pump current controller, and the control signal from controller 46 may be a voltage established based on a PI or PID control algorithm using the error signal level Serr. Again, the control algorithm may be implemented by digital or analog circuitry as desired in a particular application. As will be appreciated from Equation (11), when the error Serr is brought 25 to zero, the optical gain of amplifier unit 40 will be G = G- = (Gi/Go)(C 1

/C

2 ). When the selected optical taps 48, 48' have the same coupling ratio and the selected photodetectors 47, 47' have the same responsivity, C, and C 2 will be equal, so that the target gain is simply GT = Gi/Go. By controlling the pump laser unit to maintain Serr at zero, the amplifier 30 unit 40 operates at a constant gain G = GT. And the target gain is simply a predetermined value based on the known gains of the input-side and output- WO 00/38318 PCT/US99/29679 12 side amplifier units 43, 45. No external reference signal representing the target gain is required. Nor is a division operation to determine the actual operating gain G. Fig. 5 illustrates another embodiment 3 of the invention, in this case for 5 controlling a multi-stage pumped fiber-optic amplifier unit 40'. The arrangement in Fig. 5 is generally similar to that in Fig. 4, except that the amplifier unit includes a plurality of amplification stages, each including a doped fiber coil 41, WDM coupler 42, and pump laser unit 44 as previously described in connection with Fig. 4. Here, the number of amplification stages 10 is two, but a greater number of stages may be used. The amplifier coils of the different stages (which may be pumped the same but provide the same or different gains depending upon the particular application) are connected in series, as shown. Each stage may also include a gain-flattening filter (not shown). 15 As in the embodiment of Fig. 4, the controller in Fig. 5 subtracts one of the TIA output voltages Si, So from the other to obtain an error signal level Serr. Based on the error thus obtained, the controller generates control signals for adjusting the outputs of the pump laser units and thereby adjusts the optical gain of the optical amplifier unit 40' to bring the error Serr to zero and to 20 maintain that state. In that state, the optical gain of the amplifier unit 40' will equal the target gain, which is based on the known gains of the input-side and output-side TIA units, as previously explained. The pump laser units may receive identical control signals, even though the two amplification stages may or may not be identical. Of course, in a case where the amplification stages 25 differ and do not provide equal contributions to the overall gain, they can be controlled differently, based on their respective gain contributions. Fig. 6 shows another embodiment 4 in which two individually controlled optical amplifier units 40 are connected in series to form a multi-stage amplifier unit. This arrangement allows for more precise control of the overall optical 30 gain than the arrangement of Fig. 5, since each amplification stage is individually monitored and controlled.

WO 00/38318 PCT/US99/29679 13 As shown in Fig. 6, the output-side TIA unit 45 for the first (left) amplification stage and the input-side TIA unit 43 for the second (right) amplification stage may have their inputs connected in common to a photodetector 47' (47). Each amplification stage is separately controlled in the 5 manner described in connection with Fig. 4. Of course, although the two amplification stages are the same in Fig. 6, as are their respective control systems, this need not be the case in practice. In practice, the two amplification stages may differ, as may their control systems (e.g., the TIA unit pairs may be selected to provide different target gains). 10 An apparatus as described in connection with Fig. 4 was constructed and tested to examine its temporal dynamics. The single coil amplifier unit was composed of 13.7 m of conventional erbium-doped fiber, two optical isolators at the input and output, a 1550/980 nm WDM optical coupler, and a standard 976 nm grating stabilized pump laser unit including a voltage-controlled pump 15 current source capable of sub-microsecond (ps) response time. The input signal to the controlled amplifier unit consisted of two signals: a -10 dBm continuous wave (CW) 1555 nm signal, and a 0 dBm, 500 Hz on/off modulated square-wave signal at 1553 nm. The TIA units were constructed from respective operational amplifiers 20 (frequency response at least 10 MHz), resistors, and capacitors to have transimpedance gains of 10,000 for the input-side TIA unit and 681 for the output-side TIA unit. These values were determined to be suitable based on the particular physical parameters of the system, such as splice losses, characteristics of the pump laser, characteristics of the photodetectors, etc. 25 The TIA units were connected to the input and output sides of the optical amplifier unit via 90/10 optical couplers and substantially identical photodetectors (InGaAs PIN photodiodes with at least 10 MHz frequency response). Thus, the proportionality constants C1 and C2 were related as C2 = (9/10) C1. This is readily derived from the fact that, by virtue of the 90/10 30 coupling ratios, the amount of light coupled to the input-side photodetector is 1/9 of the light input to the optical amplifier unit, and the amount of light WO 00/38318 PCT/US99/29679 14 coupled to the output-side photodetector is 1/10 of the light output from the optical amplifier unit. The input and output fibers of the apparatus were standard SMF-28 optical fiber. Optical connections between components of the apparatus were 5 also made with SMF-28 optical fiber. The signal on the output fiber was measured with a 125 MHz photoreceiver and a 500 MHz digital oscilloscope. In order to provide a wide range of control flexibility, we used a digital, microprocessor-based control system including: - two 8-bit, analog-to-digital (A/D) converters (sampling at 2.5 Mhz) 10 - a Mach XL Programmable Array Logic (PAL) - an IBM Power PC microprocessor running at 16 MHz - an 8-bit digital-to-analog (D/A) converter Fig. 7 is a block diagram showing the arrangement of the control system 100 and its connections to the TIA units and laser pump current controller 15 (pump drive circuitry). The two A/D converters 101 and 102 are used to digitize the electrical output voltages from the input-side and output-side TIA units 43 and 45, respectively. The PAL 103 performs a fast hardware subtraction of the digitized amplifier outputs from the A/D converters. The subtraction result, 20 representing the error Serr, is used by the microprocessor 104 to execute a PI control algorithm, and the resulting calculated control signal is then output via the D/A converter 105 to the pump drive circuitry of the laser unit 44. Fig. 8 is a flow diagram of the control process. In step S1, the PAL reads the digitized TIA outputs (Si, So) from A/D converters 101, 102. Next, in 25 step S2, PAL 103 calculates the error Serr(n) = Si - So, where n denotes the present (nth) control iteration. In step S3, the microprocessor 104 calculates a control move M(n) for adjusting the pump laser output to bring the error Serr to 0. The calculation is performed in accordance with the following PI control 30 algorithm: M(n) = M(n-1) + I Serr(n) + P (Serr(n) - Serr(n-1)) WO 00/38318 PCT/US99/29679 15 where coefficients I and P are, respectively, the integral and proportional control gains and may be selected and optimized empirically. For a more complete discussion of PI control, see the aforementioned Koenig text. The optimum values of I and P for the test apparatus were determined empirically 5 to be 1 and 500, respectively. Next, in step S4, the value Serr(n-1) for use in the next iteration is set to the current error Serr(n). Finally, the control move M(n) is output to the pump drive circuitry, and the flow returns to step S1 to perform the next control iteration. 10 The temporal dynamics of the single-coil amplifier described above, but without the automatic gain control system of the invention in operation, are shown in Fig. 9. More particularly, Fig. 9 shows the traces of the square-wave input signal and the resulting output signal. The relative scales of the two signal traces were set so that the traces could be superposed, as shown, for 15 easier comparison. A comparison of the signal traces reveals that the amplifier output suffers severe distortion due to the slow gain dynamics associated with erbium ions in glass. Fig. 10 shows the temporal dynamics with the gain control system turned on. As seen in Fig. 10, the optical output signal has a square-wave 20 shape with low distortion. The third trace in Fig. 10, above those of the input and output signals, represents the pump control signal and is proportional to the pump power. The details of the turn-on transient are shown in Fig. 11. The initial rapid increase in the output signal is provided by the initial inversion (inversion 25 refers to the population of erbium ions in the excited state). However, the pumping rate is insufficient to accommodate the increased signal power. Hence, after a processing delay of approximately 5 ps, the gain control circuit increases the pump power to the maximum allowed value. The amplifier responds with a temporally increasing gain until the desired gain is achieved. 30 The pump power is then decreased to the value required for steady-state operation. The total time that is required to correct the gain is approximately WO 00/38318 PCT/US99/29679 16 25 ps. The response speed of the gain control system can, of course, be increased if more pump power is available during the turn-on transient. Additionally, the response time can be reduced by using faster electronics, e.g., an analog control circuit instead of a microprocessor based circuit. 5 It bears repeating that the particular applications of the invention which are mentioned herein, as well as the illustrative embodiments, are merely exemplary. Numerous other implementations are possible in keeping with the basic principles of the invention, the scope of the invention being set forth in the appended claims.

Claims (45)

1. A method of controlling a ratio of an output signal level from a device and an input signal level to the device, comprising: providing a first component connected to an input of the device to 5 produce a first signal having a level corresponding to said input signal level; providing a second component connected to an output of the device to produce a second signal having a level corresponding to said output signal level; and adjusting said ratio based on a difference between the levels of said first 10 and second signals.

2. A method according to Claim 1, wherein said device is an amplifier unit and said ratio is a gain of that amplifier unit. 15

3. A method according to Claim 2, wherein each of said output signal and said input signal is a composite signal.

4. A method according to Claim 1, wherein said device is an optical device, said input and output signals are optical signals, and said first and 20 second amplifier units are transimpedance amplifier units.

5. A method according to Claim 4, wherein each of said output signal and said input signal is a composite signal. 25

6. A method according to Claim 5, wherein said output and input signal levels are RMS levels of the composite signals.

7. A method according to Claim 1, wherein said adjusting includes producing an error signal level Serr, where Serr = ± (So - S), So is a signal level 30 of said second signal, and Si is a signal level of said first signal, and using said error signal level as a basis to adjust said ratio. WO 00/38318 PCT/US99/29679 18

8. A method according to Claim 7, wherein said first signal and said second signal are composite signals and Si and So are RMS signal levels. 5

9. A method according to Claim 7, wherein said error signal level and said first and second signal levels are voltage levels.

10. A method according to Claim 1, wherein said adjusting includes adjusting a component of said device that affects said ratio. 10

11. A method of controlling an optical gain of an optical amplifier unit, comprising: providing a first transimpedance amplifier unit connected to an input of said optical amplifier unit; 15 providing a second transimpedance amplifier unit connected to an output of said optical amplifier unit; and adjusting the optical gain based on a difference between an output signal of said first transimpedance amplifier unit and an output signal of said second transimpedance amplifier unit. 20

12. A method according to Claim 11, wherein said optical amplifier unit is a pumped fiber-optic amplifier unit and said adjusting includes adjusting pump power of a pump laser of said fiber-optic amplifier unit. 25

13. A method according to Claim 12, wherein the optical gain is RMS optical gain for a composite plurality of wavelengths input to said fiber-optic amplifier unit.

14. A method according to Claim 12, wherein the pump power is 30 adjusted by adjusting a drive control signal for said pump laser. WO 00/38318 PCT/US99/29679 19

15. A method according to Claim 12, wherein said adjusting includes producing an error voltage Serr, where Serr = ± (So - Si), So is an output voltage of said second transimpedance amplifier unit and Si is an output voltage of said first transimpedance amplifier unit, and using said error voltage as a basis to 5 adjust the pump power of said pump laser.

16. A method according to Claim 15, wherein said first and second transimpedance amplifier units are respectively connected to said input and said output through corresponding photodetectors, and So and Si satisfy the 10 following relationships: So= GoC 2 GPi Si = GiC 1 Pi where Go and Gi are the respective transimpedance gains of said second transimpedance amplifier unit and said first transimpedance amplifier unit, Pi is 15 input optical signal power, G is the optical gain, and C2 and C1 are constants.

17. A method according to Claim 12, wherein said fiber-optic amplifier unit has a single fiber-optic amplification stage. 20

18. A method according to Claim 12, wherein said fiber-optic amplifier unit has a plurality of fiber-optic amplification stages connected in series.

19. An apparatus for controlling a ratio of an output signal level from a device and an input signal level to the device, comprising: 25 a first component connected to an input of the device to produce a first signal having a level corresponding to said input signal level; a second component connected to an output of the device to produce a second signal having a level corresponding to said output signal level; and a controller connected to said first and second amplifier units and said 30 device, said controller operating to adjust said ratio based on a difference between the levels of said first and second signals. WO 00/38318 PCT/US99/29679 20

20. An apparatus according to Claim 19, wherein said device is an amplifier unit and said ratio is a gain of that amplifier unit. 5

21. An apparatus according to Claim 20, wherein each of said output signal and said input signal is a composite signal.

22. An apparatus according to Claim 19, wherein said device is an optical device and said first and second amplifier units are transimpedance 10 amplifier units.

23. A method according to Claim 22, wherein each of said output signal and said input signal is a composite signal. 15

24. A method according to Claim 23, wherein said output and input signal levels are RMS levels of the composite signals.

25. An apparatus according to Claim 19, wherein said controller produces an error signal level Serr, where 20 Serr = ± (So - S), So is a signal level of said second signal and Si is a signal level of said first signal, and said controller uses said error signal level as a basis to adjust said ratio.

26. A method according to Claim 19, wherein said controller adjusts a 25 component of said device that affects said ratio.

27. A gain-controlled optical amplifier apparatus, comprising: an optical amplifier unit; a first transimpedance amplifier unit connected to an input of said optical 30 amplifier unit; WO 00/38318 PCT/US99/29679 21 a second transimpedance amplifier unit connected to an output of said optical amplifier unit; and a controller connected to said first and second transimpedance amplifier units and to said optical amplifier unit, said controller operating to adjust an 5 optical gain of said optical amplifier unit based on a difference between an output signal of said first transimpedance amplifier unit and an output signal of said second transimpedance amplifier unit.

28. An apparatus according to Claim 27, wherein said optical amplifier 10 unit is a pumped fiber-optic amplifier unit including a pump laser, and said controller adjusts pump power of said pump laser.

29. A method according to Claim 27, wherein the optical gain is RMS optical gain for a composite plurality of wavelengths input to said fiber-optic 15 amplifier unit.

30. An apparatus according to Claim 27, wherein said first and second transimpedance amplifier units are respectively connected to said input and said output through corresponding photodetectors. 20

31. An apparatus according to Claim 27, wherein said controller produces an error voltage Serr, where Serr = ± (So - Si), So is an output voltage of said second transimpedance amplifier and Si is an output voltage of said first transimpedance amplifier, and 25 said controller uses said error voltage as a basis to adjust the optical gain.

32. An apparatus according to Claim 31, wherein said first and second transimpedance amplifier units are respectively connected to said input and 30 WO 00/38318 PCT/US99/29679 22 said output through corresponding photodetectors, and So and Si satisfy the following relationships: so= GoC 2 GPi Si = GiC 1 Pi 5 where Go and Gi are the respective transimpedance gains of said second transimpedance amplifier unit and said first transimpedance amplifier unit, Pi is input optical signal power, G is the optical gain, and C 2 and C 1 are constants.

33. An apparatus according to Claim 32, wherein said optical amplifier 10 unit is a pumped fiber-optic amplifier unit including a pump laser, and said controller adjusts pump power of said pump laser.

34. An apparatus according to Claim 33, wherein said fiber-optic amplifier unit has a single fiber-optic amplification stage. 15

35. An apparatus according to Claim 33, wherein said fiber-optic amplifier unit has a plurality of fiber-optic amplification stages connected in series. 20

36. A method of controlling a ratio of an output signal level from a device and an input signal level to the device, comprising: providing a first component connected to an input of the device to produce a first signal having a level corresponding to said input signal level; providing a second component connected to an output of the device to 25 produce a second signal having a level corresponding to said output signal level; and adjusting said ratio based on a difference between the levels of said first and second signals. 30

37. A method according to Claim 36, wherein said device is an amplifier unit and said ratio is a gain of that amplifier unit. WO 00/38318 PCT/US99/29679 23

38. A method according to Claim 36, wherein said device is an optical device, said input and output signals are optical signals, and said first and second amplifier units are transimpedance amplifier units. 5

39. A method according to Claim 36, wherein said adjusting includes producing an error signal level Serr, where Serr = ± (So - S), So is a signal level of said second signal, and Si is a signal level of said first signal, and using said error signal level as a basis to adjust said ratio. 10

40. An apparatus for controlling a ratio of an output signal level from a device and an input signal level to the device, comprising: a first component connected to an input of the device to produce a first signal having a level corresponding to said input signal level; 15 a second component connected to an output of the device to produce a second signal having a level corresponding to said output signal level; and a controller connected to said first and second amplifier units and said device, said controller operating to adjust said ratio based on a difference between the levels of said first and second signals. 20

41. An apparatus according to Claim 40, wherein said device is an amplifier unit and said ratio is a gain of said amplifier unit.

42. An apparatus according to Claim 40, wherein 25 said controller produces an error signal level Serr, where Serr = ± (So - Si), So is a signal level of said second signal and Si is a signal level of said first signal, and said controller uses said error signal level as a basis to adjust said ratio. 30

43. A method of controlling a ratio of an output signal level from a device and an input signal level to the device, comprising: WO 00/38318 PCT/US99/29679 24 producing a first signal having a level corresponding to said input signal level; producing a second signal having a level corresponding to said output signal level; and 5 adjusting said ratio based on a difference between the levels of said first and second signals.

44. An apparatus according to Claim 43, wherein said ratio is adjusted so as to substantially maintain a target ratio of said input and output signal 10 levels, said target ratio being based on proportionality factors relating said first and second signal levels to said input and output signal levels, respectively.

45. An apparatus according to Claim 44, wherein said first signal and said second signal are produced, respectively, by a first amplifier unit and a 15 second amplifier unit, said first and second amplifier units have predetermined gains, and said proportionality factors are said predetermined gains.