1、 TDA Progress Report 42-128 February 15, 1997Characterization of the Avalanche Photo DiodeDetector Used for Optical CommunicationsExperiments With the Japanese ETS-VIC. PasqualinoCommunications Systems and Research SectionThe avalanche photo diode (APD) was used as the front-end light collector onth
2、e receiving telescope in a series of optical communications experiments. Theseexperiments were conducted between October 1995 and May 1996 in two phasesat Table Mountain, California, and utilized the Japanese Engineering Test Satellite(ETS-VI). Following the Phase 1 experiments, extensive laboratory
3、 testing of theAPD module was carried out and the results documented here. Essentially, it isfound that the APD provides a linear response proflle between input light powerand output voltage level over the expected operational range of from 2 to 20 nW.In addition, it was found that the APD used in t
4、he Phase 1 experiments yields a1.75-MHz, 3-dB bandwidth. It was also found that a better gain distribution couldbe obtained by cascading the APD with a linear power amplifler. This extendedthe APD bandwidth to 3.75 MHz for the Phase 2 experiments. This improved gaindistribution also results in impro
5、ved optical receiver performance as characterizedby the bit-error rate.I. IntroductionFrom October 1995 through May 1996, a series of optical communications experiments was performedutilizing the Japanese Engineering Test Satellite (ETS-VI). The ground portion of these experimentsemployed the 0.61-
6、and 1.22-m telescopes located at Table Mountain, California. A bit-error rate tester(BERT) provided the data stream that was used to modulate the transmit laser located in the 0.61-mtelescope. The 1.22-m telescope acted as the Earth-based optical receiver. The light gathered at thistelescope was foc
7、used on an avalanche photo diode (APD). The signal produced by the APD was amplifledand fed into the receive BERT.This article describes the results of a series of tests that was performed with the APD module followingthe completion of the Phase 1 satellite experiments. Two APD conflgurations are co
8、nsidered in this article.The flrst conflguration consists of the APD module that was used during the Phase 1 experiment. Thesecond conflguration is a combination of the APD and a new amplifler stage that was used in the Phase 2experiment.1II. Testbed Equipment ConfigurationThree testbed conflguratio
9、ns were used in this experiment. Diagrams of the difierent conflgurationsare presented in Figs. 1 through 3. The setups were quite similar and so, in order to make it easier tonote the difierences, the components that were replaced remain in the diagram but are crossed out. Keyelements common to all
10、 three conflgurations include the power meter probe 2 and the monitor APD.The power meter probe 2 was used for real-time monitoring of the light level reaching the APD underevaluation. The monitor APD was used to establish the bandwidth of the laser modulator.Figure 1 shows the conflguration used to
11、 calibrate the ratio of light levels between probes 1 and 2of the light power meter. The conflguration used to determine the bandwidth of the received signal ispresented in Fig. 2. Figure 3 is a diagram of the conflguration used to test the amplitude response of thereceived signal for varying light
12、levels. Finally, Figure 4 shows the conflguration used for bit-error rate(BER) testing. Details about the equipment used in the laboratory testing can be found in Table 1.HP 3325AFUNCTIONGENERATORCURRENTSOURCELASERDIODECOLIMATORMODULATORVARIABLETRANSMISSIONBEAMSPLITTERMONITORAPDPOWERMETERPROBE 2ANGL
13、EDNEUTRALDENSITYFILTERNEUTRALDENSITYFILTERFOCUSINGLENSPOWERMETERPROBE 1SHROUDFig. 1. The power meter probe and modulator frequency response calibration configuration.IRISFOCUSINGLENSILX LIGHTWAVELDX 3620POWERMETERPROBE 1CURRENTSOURCELASERDIODEILX LIGHTWAVELDX 3620COLIMATORMODULATORVARIABLETRANSMISSI
14、ONBEAMSPLITTERMONITORAPDPOWERMETERPROBE 2ANGLEDNEUTRALDENSITYFILTERNEUTRALDENSITYFILTERFOCUSINGLENSAPDUNDERTESTSHROUDIRISFOCUSINGLENSFig 2. The APD frequency response configuration.HP 3325AFUNCTIONGENERATORIII. Testbed CalibrationBefore any experiments were performed, it was necessary to calibrate t
15、he test setup being used. In thisregard, there were two primary areas of interest. First, it was important to determine if any bandwidthlimitations existed in the transmitted signal. Second, it was necessary to calibrate the laser light powerreadings so that a relatively precise estimate of the lase
16、r light collected by the APD could be provided.2POWERMETERPROBE 1HP 3325AFUNCTIONGENERATORCURRENTSOURCELASERDIODEILX LIGHTWAVELDX 3620COLIMATORMODULATORVARIABLETRANSMISSIONBEAMSPLITTERMONITORAPDPOWERMETERPROBE 2ANGLEDNEUTRALDENSITYFILTERNEUTRALDENSITYFILTERFOCUSINGLENSAPDUNDERTESTSHROUDIRISFOCUSINGL
17、ENSFIREBERDTESTERETS-VIMODULATORFig. 3. The APD amplitude response configuration.CURRENTSOURCELASERDIODEILX LIGHTWAVELDX 3620COLIMATORMODULATORFIREBERDTESTERFIREBERDTESTERETS-VIMODULATORETS-VIDEMODULATORAPDUNDERTESTOPTICALTRANSMISSIONCHAIN(SEE PREVIOUSFIGS.)Fig. 4. The APD bit-error rate test config
18、uration.A. Light Power Meter CalibrationIn order to obtain the most accurate light power meter readings, the responsitivity of the light powermeter probes must be set prior to operation. For probe 1, the factory-provided responsitivity calibrationfactor is 1:593101, and for probe 2 it is 1:556101.Us
19、ing the test setup of Fig. 1, several readings of the light levels at probes 1 and 2 were taken. Theroom was darkened during the measurements to ensure that stray light did not cause false readings.Figure 5 shows a plot of the resulting measurements. The response appears to be largely linear, witha
20、slope of approximately 2:518103W/nW over the measurement region. This information is usedwhen experimental measurements are taken. By multiplying the light power impinging on probe 2 by theslope, the power reaching the APD under test can be determined.B. Transmitter Characterization With the Monitor
21、 APDThe flnal phase of the calibration was to characterize the modulation/transmitter portion of the testsetup. The modulator was the same one used during fleld experimentation, a Conoptics Model 10. Forthis test, the variable transmission beam splitter was adjusted to maximize the light impinging o
22、n themonitor APD. The output voltage of the monitor APD was terminated with 50 s and observed onan oscilloscope. Then, the frequency was changed, and the resultant peak-to-peak voltage was noted.The results of this test, provided in Fig. 6, show the 3-dB roll-ofi occurring at approximately 17 MHz.Si
23、nce it is possible that the measurement equipment could be limiting the bandwidth of the signal, thisvalue serves only as a lower limit for the bandwidth of the signal being transmitted. This bandwidth isadequate to determine the bandwidth of the receive APD.3Table 1. Test configuration equipment de
24、tails.Item (block label) Manufacturer Model no. Serial no. NotesCurrent source ILX Lightwave LDX 3620 36201207 High bandwidth settingLaser diode Spectra Diode Labs SDL2300 N/A 810 nm, 200 mWColimator Melles Griot 4X N/A |Function generator Hewlett Packard 3325A 1748A04830 Delivers a sine wave from 0
25、 to 1Fireberd test set Telecommunications Fireberd 17402 Diphase card installedTechniques Corp. MC6000Uplink modulator JPL Satellite Com- N/A N/A Bias set to 150 Vmunications GroupModulator Conoptics 10 765 10-MHz bandwidthMonitor APD N/A ET200 N/A |Angled neutral N/A N/A N/A 3.0density fllterLight
26、power meter United Detector S380 17874 Measures watts at 633 nmTechnologies (UDT) Accuracy: from 10 pW10 mWis 5% of readingSee probe 1 and probe 2 settingsfor further informationLight power meter UDT 268R 16249 Calibration 20169probe 1 Responsitivity set to 1:593101Light power meter UDT 268R 16249 C
27、alibration 20169probe 2 Responsitivity set to 1:556101Neutral density fllter N/A N/A N/A #2 0D2.0 1.0 bmmAPD under test EG&G C30872E 1366 Bias voltage set to 270 VDCOptoelectronicsCanada5010015020020 40 60 80Slope = 2.58500PROBE 2 POWER, WFig. 5. Relative light power at probe 1 (nW) and probe 2 (W).
28、40.00.51.01.52.02.53.03.54.04.5104105106107108FREQUENCY, Hz3 dBFig. 6. The frequency response of the modulator section inthe testbed.IV. APD Module CharacterizationThe purpose of this section is to provide a characterization of the APD module that was used in thefleld during the Phase 1 experiments
29、at Table Mountain. No attempt has been made to optimize eithermagnitude or frequency response. After the laboratory tests were completed, the feedback resistance forthe APD amplifler was determined by direct measurement to be 2.51 M.1A. Frequency ResponseFigure 2 shows the test setup used to determi
30、ne the frequency response of the APD. Initially, a 10-kHzsine wave was output from the modulator. The variable transmission beam splitter was adjusted untilthe power level at probe 2 measured 10.1 W, corresponding to an APD light level of about 26.21 nW.The peak-to-peak voltage at this frequency was
31、 noted. Then, the output voltage was measured for anincreasing series of frequencies. The normalized results of this test (along with two other tests performedat difiering power levels) are provided in Fig. 7. Note that as the signal magnitude decreased, the dicultyin measuring the signal amplitude
32、increased due to a substantial drop in the signal-to-noise ratio (SNR).All curves in Fig. 7 show a 3-dB roll-ofi point to be at about 1.75 MHz, based on an initial normalizedpower level of 1.5 dB.B. Amplitude ResponseFigure 3 shows the test setup used to determine the amplitude response of the APD.
33、For this test,a Pnsequence of length 291 (511) bits was input to the transmitter at a data rate of 1,024,000 bps.This resulted in a laser signal modulated with a Pnsequence exhibiting square-wave characteristics. Thisis representative of the ETS-VI signal format. The light level to the APD was varie
34、d with the variabletransmission beam splitter, and the light power at probe 2 was noted along with the peak-to-peak voltageoutput from the APD. To determine the actual light power at the APD, the probe 2 level was multipliedby 2:518103, as described in Section III.A. The results of this test are sho
35、wn in Fig. 8. From thisflgure, it is apparent that there is a linear relationship between input light power and output voltagelevel.1The feedback resistance, along with the current sourced by the APD, determines the amount of gain the operationalamplifler provides, according to the equation AV= iDRF
36、, where RFis the feedback resistance and iDis the diodecurrent.5181614121086420104NOMINAL SOURCE POWER = 2.6 nWNOMINAL SOURCE POWER = 26 nWNOMINAL SOURCE POWER = 42.5 nW105106107FREQUENCY, HzFig. 7. The APD frequency response (gain set to maximum).0.00.51.01.52.02.53.03.54.001020304050SLOPE = 92.54
37、mV/nWAPD MODULE OUTPUT, VFig. 8. The APD amplitude response.V. APD With Amplifier Module CharacterizationThe frequency response results of the APD module set at high gain showed band limiting beginningat about 1.25 MHz. In an efiort to extend the bandwidth of the receiver, the gain of the APD module
38、was reduced by a factor of flve, and an amplifler module was added to the telemetry chain after theAPD module. The resultant frequency response is depicted in Fig. 9. The 3-dB roll-ofi point is at about3.75 MHz. This represents an improvement in bandwidth of approximately 3.3 dB.61051061412108642010
39、4NOMINAL SOURCEPOWER = 9 nW107FREQUENCY, HzFig. 9. The APD with amplitude frequency response(ADP gain reduced).VI. BER CharacterizationAs a flnal test, BER characterizations were performed on the APD module and on the APDmoduleamplifler combination. Figure 10 summarizes the results of this test. Cle
40、arly, there is a sub-stantial performance improvement from utilizing the APDamplifler combination. The BER was quitesensitive to background noise. Any light not generated by the laser afiects the results. Simply movingaround the darkened laboratory could cause an increase in bit errors. Additionally
41、, if the laboratory lightswere on, the communications link could not be maintained. This was the case even though a shroud wasused at the APD.APDAPDAMPLIFIER1021031041051061071081090.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0LIGHT POWER, nWFig. 10. The ADP BER characterization.7VII. ConclusionsThe APD conflg
42、uration used at the receiver band limited the received signal to about 1.75 MHz. Thisphenomenon likely is caused by the high-gain setting on the APD circuitry and the low light levels. Fur-thermore, the band limiting is likely to cause intersymbol interference and so, if possible, the bandwidthof th
43、e receiver should be increased. To this end, the gain of the APD was reduced and an ampliflermodule was added to the telemetry chain. This resulted in the extension of the 3-dB bandwidth to about3.75 MHz. The input light power level-to-output voltage curve is quite linear and highly repeatable. TheBER response of the APD is quite good. However, a substantial improvement (3 dB) was realized withthe addition of an amplifler. The combined APDamplifler circuit was able to achieve a BER of 106atlight power levels of 0.45 nW.8
