From Russian ASC: Alexander Smirnov, Boris Kanevsky
From RISDE: Yury Korneev
From NRAO Green Bank: Dan Pedtke, Toney Minter, Ann Wester, summer student Steve Hicks

SUMMARY:

The Russian radio astronomy satellite, "RadioAstron", is in the construction phase in Russia. The engineering prototype of the RF modules used in the satellite was brought to the US at the invitation of the JPL and NRAO for compatibility tests. The equipment was in Green Bank from May 20 thru May 28, then traveled to JPL in Pasadena where further tests will be performed thru June 10.

Results of the Green Bank tests show that the X-band timing system performs well with our earth station. Overnight 16 hour phase tests show phase residuals similar to those logged from the VSOP satellite. However, the wide band data link at 15 GHz appears to have modulation problems. No acceptable data error rates were achievable in field tests with the Green Bank earth station, despite good results achieved over the identical path with the NRAO satellite simulator. Time prohibited further tests to identify the problem.

Pictures of the Russian visit, including setup views, some spectrum analyzer and scope displays from the tests, and off-site activities, were posted at www.gb.nrao.edu/~dpedtke/radioastron/. See also a report by Toney Minter discussing the round trip phase test results.

CHRONOLOGY:

Thursday 5/20
Picked up visitors and equipment at Dulles Airport and traveled to Green Bank. Minor delay at customs for the equipment due to the invoice being in Russian and needing to be translated. Handled by JPL.

Friday 5/21 Late start due to late arrival last night. Equipment was unpacked and set up in the lab. There was a 2 hour wait for warm-up, to allow the crystal oscillator to reach frequency, and the TWT to stabilize. The system was found to have a command problem, it wouldn't turn the transmitter off, but otherwise it locked to the incoming X-band uplink generated by their test set, and transmitted on X and K band.

Alexander, Boris, and Yury

Saturday 5/22
This morning the command problem disappeared. However, the system wouldn't lock at X-band. We moved to the other transponder and the system worked. We then did spectral plots of the output using their test set as the reference. These 7 plots were of the K-band spectrum with and without modulation.
Note that they use an approximately 14 MHz crystal oscillator doubled 9 times to 7215 MHz as an uplink signal. K-band power output was measured at +40.5 dBm (11 Watts).

Monday 5/24
The Russians attempted to repair their command test set in the morning. Apparently the command problem was with one of the logic relays inside (there is no semiconductor logic used in the test set). A connector was damaged on one of the transponder modules, and this was re-attached. However, that didn't fix the lock problem as hoped. After lunch, we moved the equipment setup to the field test site upstairs. On air tests were not possible until tomorrow due to the on-site observing schedule.

Tuesday 5/25
After system warm up, we performed the first on-air tests with the earth station (see block diagram). We achieved 2-way lock easily, surprisingly without the uplink antenna connected, in fact terminated (our earth station was aimed at the window where the equipment was located). It was curious that this lock occurred at a power level transmitted by the earth station very close to that required to lock with the antenna connected. It appeared that there must have been a leak in the X-band receive system. X-band power output was measured at 400 mW.

A 10 minute round-trip lock test of phase stability showed good results. But the data path was not good. Received spectrum at the earth station was not smooth or stable. This was probably due to being in the near field of the earth station, and a tree with leaves blowing was partially blocking the path. We decide to reflect the signal off of the site water tower. Stable signals were then possible but the error rate was still found to be high.
An overnight phase stability test was started around 5:30 PM.

Wednesday 5/26
Overnight phase stability test ended at 9 AM. Results looked very good. Plots were made of phase, phase fit, and phase fit error vs. time. This information will be detailed in a report by Toney Minter.
Data error rates to the station were still poor. Valid sync is seen most of the time, but frame parity error rate was perhaps one part in 10 to one part in 100. We decide to try the NRAO satellite simulator at the Jansky lab test site, over the same path with the same antenna. A K-band amplifier was required to get the signal strong enough for the link. The setup was not completed today.

Thursday 5/27
Tests with the NRAO satellite simulator show good signals with no errors at 18, 36 and 64 MHz data rates, and poor error performance at 72 MHz (unknown cause). Switching back to the VIRK module, the same error rates as before existed. Thinking that the power supply spurs might be causing a problem in the demodulator, we ran the transponder from an external 14.71 MHz reference signal. This was a suggestion from Yury. This indeed cleaned up the spurious products in the 15 GHz unmodulated carrier, indicating that these 40 KHz products were not coming from the TWT switching power supply, but from the reference. This clean source was then used while transmitting data to the earth station, but the data errors remained. At the end of the day we once again tested the satellite simulator, and results were still good.

It must be concluded that the problem is inside the VIRK-M module, probably with the modulator. A check of the demodulated analog signals coming from the Russian test set demodulator show much jitter (see scope photos). They are feeding this demodulator with 1 mW of signal, and it has no clock recovery circuitry - the clock is provided directly around the RF link.

The ECL modulator inputs were measured for DC bias with inputs disconnected. The bias point was at half-way point for ECL, -1.4 volts. Apparently, the external signals are expected to be AC coupled. Their schematic of the modulator driver shows that the termination to 50 ohms is AC coupled, with the DC bias applied thru greater than 1K ohm. Even so, with our data generator connected (with it's internal pull-down resistors), the data levels were acceptable ECL levels (-1.1 and -1.8 volts), and they are buffered by internal ECL logic, so input levels would not affect the QPSK modulator.

Also note that during these tests, an anomalous mode was detected in the K-band transmitter, which the Russians were aware of. The K-band output appeared to have a 500 Hz square wave amplitude modulated onto it (see spectral plot). Power down and back on cleared this problem. Later it was noted that the fan set used to cool the TWT heatsink was not turned on, and this was blamed for the anomalous signal.

Friday 5/28
Transponder X-band lock range testing was done until 10 AM. These tests showed that the lock range was limited mainly by the search sweep range of the PLL, as the frequency limits were not power level sensitive. Input capture range was from 7215.2670 to 7215.2320 MHz (+12 to -23 KHz), and hold range was 7215.26724 to 7215.232240 MHz (+26.5 to -23 KHz). The hold range widened by about 1200Hz when the input power was increased from -70 dBm to -50. Note that hold range measurements were very sensitive to rate of change. We had to move in 1Hz steps, about 1 step per second, to hold lock at the extremes.

After the lock range testing, the equipment was torn down and packed. The equipment left Green Bank at noon.

CONCLUSIONS:

1) All tests requested by Larry D'addario were completed, with the possible exception of no X-band transmitter spectra were recorded.

2) Round Trip Phase stability for X-band timing system is good.

3) Data modulation at 15 GHz appears to have problems. Not enough time was available to narrow this problem down.

4) 15 GHz output from TWT has many spurs which appear to be from the reference signal.

5) Statements made by the visitors indicate that a replacement of the TWT with a 40 watt version (whether high orbit is used or not) is a certainty, along with a better X-band LNA (100K sensitivity rather than the current 700K). This would imply that these tests need to be repeated when final hardware is available.