LO terminator Dick Plambeck, 1/9/2004 Version 2, 4/17/04 Version 3, 10/27/04

Transcription

1 LO terminator Dick Plambeck, /9/00 Version, /7/0 Version, 0/7/0 Function: Provides 00-0 MHz phaselock reference signal (LO ref) at each antenna. Incorporates fiber directional coupler to send echo signal back to the control building. Brief circuit description: The LO reference signal is transmitted from the control building to each antenna on a singlemode fiber link. A fraction of the incoming laser power is split off by an optical directional coupler inside the LO terminator box and is sent back to the control building on a second singlemode fiber for the roundtrip phase measurement. The straight-through signal from the coupler is sent to a photodiode (Fermionics FD0F-FC/APC-PCB). U (MAX00) supplies DC bias to the photodiode, and also provides a convenient current monitor. The RF output of the photodiode is amplified in low noise amplifier U (RF,. db noise figure), and two following amplifiers (U7, ERA-; U9, ERA-). The total gain is approximately 7 db. The output level is controllable via a digital attenuator (U, Hittite HMC7, 0- db in db steps) located just before the ERA- amplifier. Log amps (AD) measure the RF power before the attenuator and after the final output amplifier. Optical couplers. 0 db couplers are installed in LO term serial numbers -0, db couplers in serial numbers -0. Probably we will switch all boxes to db couplers in the future. Design considerations. Typically a fiberoptic link has a loss of 0 db. We expect to pump the Fiber-Span lasers in the control building with + dbm RF input power, and want the LO terminator boxes in the antennas to provide at least + to + dbm to drive the harmonic generators for the X-band phaselocks. Hence, the LO terminator box must have a gain of at least 0 db. We provide db to allow for optical losses of up to 7 db in the fiber system (the expected loss for a link with connectors, several splices, and km of fiber is to db). Note that RF loss is the square of the optical loss because the modulated laser intensity encodes the amplitude of the RF waveform. Some sort of AGC, variable attenuator, or limiting amplifier is needed to keep the output power in the target range, whatever the optical loss. A key requirement is that small changes in optical power (which might be caused by fiber flexure as the telescope slews between source and calibrator) should not cause substantial phase changes at the LO frequency. Tests of a prototype LO terminator module showed that this requirement was not met with an AGC loop and a voltage-variable attenuator. With this system the phase change at. GHz was of order. degrees/db. Flexing the fiber causes up to +/- % changes in the optical power, corresponding to % (0.09 db) changes in RF power, which then leads to 0. degree changes in the. GHz reference phase. When multiplied up to 0 GHz, this is a 0 degree phase jump, which is unacceptably large.

2 Instead, we chose to use a digitally controlled attenuator with discrete steps. With digital control, the attenuator settings can be left untouched during each observing track (and probably during the entire time an antenna is at a particular pad). Note that if the attenuation is set to zero and the optical loss is small, the output amplifier saturates at approximatly + to +9 dbm. The phase through a saturated amplifier changes slightly with RF level, so there is still some phase changes as the fibers flex. This is discussed more fully below. Ideally, the attenuator in the LO term box is used to limit the LO term output level to no more than dbm. Input: 0 nm optical signal on singlemode fiber, FC-APC connector. The expected optical intensity (allowing for db loss in fiber) is + to + dbm; the circuit can provide an LO reference output power of + dbm if the optical input power is greater than - dbm. "test in" SMA female provides a coax connection to the input of the first amplifier through a 0 db internal attenuator, for test purposes. Because of the 0 db internal attenuator, there is no need to terminate this port when it is unused. The test input can be used to operate a receiver in the lab if no fiber input is available. Output: Echo fiber signal, FC-APC connector, for round trip phase measurement. RF out, SMA female connector. The max possible level of + dbm is set by the saturation of the output amplifier. Indicator LED: green indicates adequate optical input power (photodiode current monitor > 0.V, corresponding to laser input power > 0. mw); with no internal attenuation, the LO reference output power should be > dbm. yellow indicates marginal optical input power (0.0 V < photodiode current monitor < 0. V, corresponding to 0. mw < optical pwr < 0. mw); with internal attenuator set to 0, the LO reference output power should be in the range + to +7 dbm. off indicates laser power is unacceptably low; LO reference output will be below + dbm.

3 -pin D-subminiature connector, male: db atten; gnd to activate db atten; gnd to activate db atten; gnd to activate RF in; RF level before internal attenuator = (0*RFinVolts ) dbm - V power in, 0 ma 7 DS0 -wire serial number out + V power in, 0 ma 9 db atten; ground to activate 0 db atten; ground to activate LM temperature sense output, 0 mv + 0 mv/c (note ) RF out; RF level at output SMA = (0*RFoutVolts ) dbm (note ) photodiode current, 0. V/mA; corresponds to 0. V/mW optical pwr NC. The LM is mounted on the circuit board close to the bias resistors for the output amplifer stage, so it tends to run C hotter than the chassis.. RFout can be double-valued above.00 V (e.g., RFout =.0 V might indicate 7 dbm or. dbm); it s best to keep RFout < V. Setting the internal attenuator. Generally it will be necessary to set the internal attenuator to some value > 0 to keep the final amplifier inside the LO terminator box from saturating. If the amplifier is heavily saturated, the LO reference phase becomes sensitive to the RF level, as shown in Figure. A safe choice is to set the internal attenuator to achive RFout = 0.9 V, which corresponds to an LO reference output level of dbm (+/- approx db). At this output level, a % increase in RF power (the max we think is likely from fiber flexure) changes the LO reference phase by 0.0 degrees, or the LO phase by degrees at 0 GHz. From Figure, one sees that saturation of the second stage amplifier can also be a concern if RFin exceeds 0.9 V. This should never happen for modules using a db optical coupler in lab tests with minimal optical loss, RFin ~ 0.9V for these units, when driving the units with a Fiber-Span laser modulated by + dbm of RF power. For modules with a 0 db optical coupler, RFin ~.00 V in lab tests; if optical losses at the site are ~ db as expected, then second stage saturation should not be a problem for these units either. To reduce second stage saturation, one must either attenuate the optical power into the LO terminator, or else reduce the RF modulation on the laser transmitter. Changing the attenuator setting changes the phase! Set the attenuator only at the beginning of an observing track.

4 Figure. LO reference phase for LO term # as a function of RF power, as measured with an 7 network analyzer. The RF input was applied to the coax test port of the LO terminator. The solid curve plots the LO reference phase as a function of RFout; for all these data, RFin < 0.7 V, hence saturation of the second stage amplifier is thought to be minimal. The dashed curve plots the LO reference phase as a function of RFin; for these data the internal attenuator was set to db, so saturation of the final stage amplifier is thought to be minimal. Thermal tests. The prototype LO terminator box was tested in a thermal chamber as shown in Figure. The results are shown in Figure. At. GHz, the difference in the phase drifts of the echo and LO term outputs is approximately 0.7 degrees/c. At 0 GHz this corresponds to a phase drift of degrees/c. To the extent that the temperature behavior is reproducible, it can be corrected to first order by monitoring the temperature of the LO terminator box. The thermal phase drift is due in part to the fiberoptic coupler. Tests of the phase vs. temperature for the circuit board only were made by connecting to the test input of the LO terminator. Results are shown in Figure. The phase changes by degrees/c at 0 MHz.

5 Fig. Thermal test setup for LO terminator. Fig. Thermal test results for prototype LO terminator. The bottom panel shows the chamber temperature, while the top panel shows the echo and LO term phases; unfortunately these have opposite signs!

6 Figure. Phase vs. temperature for the LO term with th 0 MHz LO reference signal input through the SMA test port, bypassing the fiberoptic coupler and photodiode. Phase noise. The phase noise of the fiber link and LO terminator was checked using the setup shown in Figure. A transmitter and receiver are independently phaselocked to a common synthesizer frequency. The reference tone for the synthesizer phaselock is sent via a Fiber-Span laser, a. km spool of optical fiber, and the prototype LO terminator box, while the reference frequency for the phaselock is sent via a short length of standard coax. The signal to noise of the beat note is measured on a spectrum analyzer using the 'adjacent channel power' function of a spectrum analyzer. As shown in Figure, the noise power in MHz wide bands just above and below the central tone is measured to be 9 db below the carrier power in the central 00 khz channel. This implies a signal to noise ratio of db, which corresponds to an rms phase noise ϕrms = sqrt ( Pnoise/ Psig)., which is identical to the value measured when the reference tone is sent to the transmitter phaselock chain via a short length of coax.

7 Figure. Test setup for measuring phase noise.

8 Fig.. Signal to noise measurement of the transmitter test tone using the adjacent channel power feature on the spectrum analyzer.

9 J SMA (TEST PORT) R.k 000PF C R R OUT REF R R 00 R NC BIAS CLAMP D FD0F U LT0 MAX00 U L NH + C 000PF + - -V R 0k R k U VCC RF IN RF R 0 R7+.9k + VCC RF OUT R NC 7 NC L Fc R 0 UB C7 LM9 UA LM9 + U VPOS 7 C R C 000PF L NH C V_DN V_UP COM AD R0 0k 00PF +V Q N700 RFIN ENBL VSET FLTR R9 0k 000PF U7 ERA- C R 0 R 90 D Q N700 S R9 90 D Q N700 S R 7 R N.C. C 0.09 L NH C 00PF R X OUT TER G Y LED TRI JDC-0- U HMC7QS IN CP R C7 00PF RF 9 V V V V V 0 7 RF R R0 C R7 00PF.7K C C 0.09 R9 +VIN R0 R R 0.09 C U9 ERA- R7 R R L NH C 00PF R9 N.C. R R 0 R 000PF C U AD FLTR COM X TER R VSET ENBL RFIN OUT JDC-0- PHOTO DIODE CURRENT 0.V/MA V_UP CP IN V_DN 7 VPOS R C C 00PF + - R k +V 0 R TL0 UB -V 7 U LMDM V+ J SMA R 00 U DS0 OUT R 00 DATA RF OUT +V 0.09 C C C C RF LEVEL OUT V=0db (0mv=db) -V C VIN L Fe +V C VI VO Fe L7 C 0.09 C uf C.7uf DB DB DB DB DB SCI J U0 7L0UA R 0 +V TL0 UA + - C -V 0.09UF R R7 00 RF level IN Fe : Ferrite 00ma 0 Ohm digitkey# 0-0--ND R k 0 Title Size Number Orcad B LO TERM - Date: -May-00 Sheet of File: C:\Design Explorer 99 SE\project\loterm.ddb Drawn By: Revision of

10