Past Observing Campaigns
Green Bank, West Virginia

Diary 1997

Observing Run 1997: March 10 - 31

11 March, 11:00 EST
A problem with the telescope declination drive and voltage levels going to the receiver delayed the start of observations for several hours. When those problems had been resolved, we checked out the system with an observation of Pioneer 10 (Happy 25th Anniversary!). Then it was on to stars 4035 and 4036, starting at 1750 MHz and working up in frequency.

Later today we will run test observations with the Woodbury antenna.

17 March, 20:30 EST
A very hectic week of observing ended this morning. We have a break for a few days while two other experiments use the telescope.

An unusual number of problems has made this a very difficult week. Starting with the telescope drive problem mentioned above, to a computer cable that wasn't connected, to a failing amplifier in the IF subsystem,... It was a week of headaches. Even the weather didn't cooperate. We had to cut short one of our tests with the Woodbury antenna due to tornado warnings in Georgia.

In spite of all that, we still managed to observe.

23 March, 17:00 EST
A few days off the telescope gave us a chance to track down some problems in the receiver. An amplifier was going and some pins in a cable had shorted, resulting in sporadic bad sensor readings and sudden changes in receiver power levels. Since we resumed observing on the 20th, with the exception of sporadic problems with one CD detector, everything has been working well.

As I write this, one of the stars being observed is Epsilon Eridani (#4010) one of the stars observed in Project Ozma.

28 March, 22:00 EST
Observations have been going well at Green Bank. We've had several eight-hour shifts without having to check any signals with an "off-source" observation (test to see if the signal goes away when we point away from the star). This means that our database has recorded the characteristics of most of the interference signals that are in this frequency band.

Today we succeeded in detecting Pioneer 10 at both Green Bank and Woodbury. This is a major step toward full two-site operation. We've been making progress toward this since last week, working through various software and hardware problems. Doing pseudointerferometry with the Follow-Up Devices requires a great deal of precision and everything must work perfectly. We'll do some more testing with signals from Pioneer 10 and Pioneer 6. Then we'll try running in two-site mode.

For a short time while writing this note I thought I might have news to upstage the success at Woodbury, but the system found the signal in the "off-source" position on the second try. Darn!

31 March, 08:00 EST
It was a dark and stormy night...

Observations had to stop two hours early due to ice forming on the dish. By 8:00 AM, when we were scheduled to stop, there was about an inch of snow on the ground and the wind was beginning to howl. Sounds like a good time to head back to California. :-)

Observing Run: May 5 - 19, 1997

The highlight of this observing run was the presence of two "rookies" on the observing team. John Billingham, Distinguished Senior Scientist, and Frank Drake, who conducted the first SETI, each spent a week at the telescope.

Frank Drake's Project Ozma search examined two stars, Tau Ceti and Epsilon Eridani, at 1420 MHz, the frequency emitted by neutral hydrogen atoms. Since our receiver is currently configured for the higher end of the observing band (1750-3000 MHz), Frank commemorated Project Ozma by observing the same stars at 2840 MHz, twice the hydrogen frequency.

We achieved two significant milestones during this run: two-site observations with an antenna in Georgia and operation of the system (at both sites) from the Institute office in California.

Observing Run: June 12 - 27, 1997

June 12, 21:00 UT
Installation of the Phoenix receiver and down converter was accomplished in record time, less than two and a half hours (including removal of the receiver from the previous experiment). We then checked the pointing accuracy of the telescope. Minor corrections were fed to the telescope control software.

Before proceeding with SETI observations, a few more tests were necessary. We observed Pioneers 6 and 10 from Green Bank and Woodbury. This gives a complete "end-to-end" test of the hardware and the software (including the calculations needed for pseudo-interferometer).

While running these tests, the signals from the two Pioneer spacecraft were briefly visible on the same waterfall plot. Warning: the gif file is 331 kB.

After the checkout was complete, both telescopes were pointed to star number 4030, a nearby (17 light years) M dwarf star.

Observing Run: August 11 - 25, 1997

August 12, 16:00 UT
Installation of the receiver and down converter went very smoothly yesterday, with the exception that too many people remembered to flip a switch. The result was that the receiver wouldn't respond to commands, yet everyone knew that the switch had been flipped. It took about an hour before we realized that the switch had been thrown one time too many.

Then we discovered that some of the Green Bank software for the interface to the telescope had been moved to a different computer. Our contact in Green Bank hadn't been told, so he had to spend two hours fixing things for us.

We finally started observing at dinner time and it was time to turn over control of the system to the Institute in Mountain View.

Two-site observations continued until about 4:30 AM EDT (why do these things always happen in the middle of the night?) when the Woodbury elevation motor controller failed. This will probably keep us in one-site mode for the rest of this observing period.

We've been observing stars 4070, 4006, 4007, 4017, and 4019.
- Peter Backus

August 13, 20:00 UT
Today was the weekly maintenance day for the telescope. While the Green Bank staff did some serious work on a bearing, we caught up on documentation and examined a temperature problem in our receiver. However, before maintenance began, I did capture a screen showing both the carrier and one of the sidebands from Pioneer 6. The strong narrow signal is the carrier. The weaker, slightly fuzzy signal is the sideband. They appear artificially close together in frequency due to the way I selected frequency ranges for the plot. The signals are actually separated in frequency by about 2 kHz, or about three times the width of the plot.

Our major focus of the day was our receiver. The low noise amplifiers in the receiver are kept at a temperature of 20 degrees Kelvin (20 degrees above absolute zero). Portions of the electronics and the waveguide are at 70 degrees Kelvin. All of these components are in a weather-proof box that is supposed to be kept at approximately room temperature. The thermoelectric heater/coolers for the box are supposed to keep the temperature constant but have been having trouble cooling the box on sunny days. Today we had an opportunity to check air flow within the box. We'll have to run a few more tests before coming up with a solution.

Observing is about to resume. More later...
- Peter Backus

August 15, 00:00 UT
Well, we're still observing with one telescope, but thanks to the efforts of our colleagues at Georgia Tech, Woodbury will be back "on the air" soon. It turned out that the drive controller was fine and the motor itself had failed. We expect to be operating Woodbury with the repaired motor within the next few hours.

In the meantime, observations at Green Bank have been proceeding under the "two-star protocol". In this mode we alternate observations between two target stars. The stars are selected pairs with angular separations of less than ten degrees. Any signals detected on both stars are classified as interference. In effect, each star serves as an "off-source position" (comparison) for the other.

Some example star pairs that have been observed are (4153,4154), (4089,4091), (3203,4081), and (4040,4139).
- Peter Backus

August 15, 03:30 UT
Two site observations resumed.

August 22, 03:30 UT
We have made good progress over the last week, finishing off our frequency scans on stars 3203, 4012, 4021, 4040, 4066, 4068, and 4139. We're currently working away on about another dozen or so stars.

Yesterday morning, the drive failed on the Woodbury antenna again. But the crew down in Georgia had a spare drive controller swapped in by mid-day -- meanwhile we did some observing in our one-site/two-star mode.

Last night we began observations of positions selected from the list of interesting signals detected by Project META, the Harvard SETI project (Horowitz & Sagan, Astrophysical Journal vol. 415, no. 218, Sept. 1993, pp. 218-235). It is most likely that META's interesting signals were RFI. However, it is also possible that one or more of the META observations could be real ET beacons that were amplified by the effects of the interstellar medium for a short period of time (of order 10 minutes). This effect is analogous to the twinkling of stars caused by the earth's atmosphere. For the interstellar case, it turns out that while on average the scintillations make the signals look weaker, occasionally they boost the signals by up to an order of magnitude (see "Scintillation-Induced Intermittency in SETI", by Cordes, Lazio, & Sagan). We can now use our system to begin testing the possibility that some of the anomalous events in Horowitz & Sagan's list could be real ET transmitters. This is possible because our system is at least 30 times more sensitive than the META system was and we should be able to detect the hypothetical transmitters even without the scintillation boost. Also, our powerful two-site observing system will allow us to exclude RFI - something META was not equipped to do. There were five good candidates from the META survey at 2840 MHz (twice the hydrogen line frequency). Since our telescope beam is smaller, we have to do a grid of pointings around each META position, which takes about 2.5 hours. We did the first grid last night (no signals detected), and hope to finish this week. We'll do the larger list of META candidates found at 1420 MHz when we begin observing over the frequency range 1200-1750 MHz later this year (this requires installing a different flare section in our microwave feed).

Observing Run: October 14 - November 3, 1997

Oct 14
The receiver was installed smoothly. The telescope operator checked pointing and we were ready to go at noon. Woodbury was ready to go as well, so we started right out observing in two site mode.
-John Dreher

Oct 15
We are running preliminary tests on AutoSchedule mode. This is new software that automates the selection of starts to observe.
-John Dreher

Oct 16
Tests this morning showed that we can indeed detect Pioneer 10 when it is treated as a star. However, we've uncovered what may be a problem with 'stars' whose coordinates are entered directly from the .cf file rather than from the database.

Observations in two site mode are proceeding smoothly.

Harvest moon and frost on the pumpkins...
-John Dreher

Oct 18
All's quiet and everything appears to be working perfectly. In fact it is so quiet that I keep checking to see if the FUDDs are really working, since I have not had a single candidate confirmed by the main FUDD all night! This seems to be correct, however, all the candidates have been right at detection threshold where we expect noise events alone to trigger detection.

Almost finished observations at S-band of Tau Ceti, which Frank Drake started observing when he was at Green Bank in May.
-Jill Tarter

Oct 24
Thanks to a particularly nifty piece of software called SharedX and our frame relay lines between Mountain View, Green Bank and Woodbury, we have been able to really have a virtual observing console in Mountain View that controls both the Green Bank, West Virgina NRAO 140 foot antenna and the Georgia Tech Research Corporation 100 ft antenna in Woodbury, GA. We have been close to this goal for several months now, but the installation of this software has made it possible for the telescope operator at Green Bank to observe for us from his control console at the 140 foot, and for Phoenix staff to follow along. It also lets us observe from Mountain View, with the Green Bank operator able to follow everything we do. This means that Phoenix personnel don't have to travel quite so much. I think this technology is here to stay!
-Jill Tarter

Oct 29
Observations are proceeding very smoothly. We are concentrating on finishing up the re-observations of Horowitz and Sagan events (described in the Astrophysical Journal 1993, vol 415, p. 218) reported as being seen at twice the HI line frequency. If the correct explanation of these transient events is that they are intrinsically weak extraterrestrial sources briefly amplified by scintillation, then our more sensitive observations should find them. We are also finishing up S-band observations of all the stars reported to have planetary companions, as well as completing as many of the target stars already observed over part of the S-band frequency range as possible. The main difficulty during the past few days has been the loss of our frame relay for periods of several minutes. This interferes with efficient telescope control.
-Jill Tarter

Oct 31
As we approach the end of this observing session, we begin to prepare for the next. This morning we turned on the L-band (1000-1750 MHz) amplifier in our receiver in preparation for tests on Monday morning. Observations continue at S-band with an emphasis on completing the 1750 - 3000 MHz for as many stars as possible.

I have updated the star-frequency plot summarizing observations of the initial list of stars. This does not include the Horowitz candidates or the recently added stars with known companions.
-Peter Backus

Nov 3
During the last few hours of this observing session we decided to test some of the software and hardware needed for the next session. In December, we will switch to L-band, 1200-1750 MHz. We haven't observed in this frequency range since the end of the Parkes deployment in July, 1995. As we've made changes to the hardware and software in the interim, it seemed prudent to run a few tests.

Preliminary results indicate that the software was fine but that a switch in the receiver was stuck. This switch is in the cryogenically cooled portion of the receiver, so replacement is a little tricky. Fortunately, we have a few weeks before the next session.

After the tests, we went back to observing at S-band. The last star observed was 5008. The total number of stars observed throughout S-band (1750-3000 MHz) is 104, with another 31 having been partially observed. We returned control of the telescope to the Green Bank and Woodbury operators at 7:50 AM EST.
-Peter Backus and Jill Tarter

Observing Run: December 1 - 22, 1997

Dec 3, 09:00 UT Mountain View, CA
The start of any observing run is a busy time, but when we change to a new frequency band there's even more to do. First, we have to check out the receiver itself, measure the focus position for the new feed antenna, calibrate the test signal over the frequency range, etc. Then we have to become familiar with a new radio frequency interference (RFI) environment. The last time we observed at L-Band (1200-1750 MHz) was 18 months ago and half a world away (Australia). The one thing we know for certain about RFI is that it's getting worse, especially at L-Band.

On Monday morning (Dec 1), the receiver installation and pointing checks went fairly smoothly. There were a few problems such as a sticky switch in the cryogenic section of the receiver, but by late afternoon we were searching for terrestrial intelligence (RFI). We do this by pointing the antenna straight up (at the zenith) and running a series of short (roughly two minutes per 20 MHz band) observations over the entire frequency range. Since we are not pointing at a particular star (the sky is drifting through the telescope "beam" or field of view) any signals detected are from terrestrial technology. After a few passes through the frequencies, the database was primed, and we were almost ready to go.

During any observing run, in order to operate at both Green Bank and Woodbury, we must perform a daily end-to-end system test by observing a spacecraft. This is particularly tricky at L-Band since there are no suitable spacecraft transmitting in that range. So, we switch in our S-Band (1750-3000 MHz) amplifier and observe through the L-Band feed horn. [Note: this is done with a set of switches in the cryogenic section of the receiver.] Although the feed horn is not well matched to the higher frequency of the spacecraft signal (2292 MHz), we should still see it. [Note: Woodbury uses a new wideband feed that covers L-Band, S-Band and beyond.] At 12:05 UT on December 2, Pioneer 6 rose above the horizon for both telescopes. The signal at Green Bank was clearly visible on the computer screen. The Follow-Up Detectors saw the signal as expected at both sites. [A minor glitch occurred when a frequency synthesizer at Woodbury had an incorrect HPIB address. The synthesizer was reprogrammed.]

The results of the RFI scans identified a number of 10 MHz bands (the bandwidth of our spectrum analyzers) that were so heavily contaminated with terrestrial signals that it really doesn't make sense to look for ETI signals in them. We programmed the scheduling software to avoid these frequencies and then put the system in Autoschedule Mode. The first star observed was number 4069. This was followed by some more system performance measurements.

Then it was on to Tau Ceti. Unfortunately, after about an hour of observing, the temperatures began to rise in the cryogenic section of the receiver. This is not good. We will continue to observe, at slightly reduced sensitivity, until regular telescope maintenance on Wednesday morning. It is very likely that we will be unable to observe for several days until the receiver is fixed.

In the meantime, we're finishing observations of star 4028. As I write this, the system has just checked a candidate signal that was seen at both Green Bank and Woodbury. The system automatically directed both telescopes to point to an "off" position and observe again. If the signal came from the star, it would not be seen in the off position. In this case, the signal was clearly visible on the screen. It is probably due to a satellite.
-Peter Backus

Dec 4, 10:30 UT Mountain View, CA
Well, we're still observing and the receiver is still warming. After talking with the Green Bank staff, we've decided to observe at slightly reduced sensitivity for another day. This will allow an Observatory receiver to be prepared for use by other astronomers while ours is repaired.

We're currently observing some of the stars with known planetary companions. As we step through the frequency range, it's sometimes interesting to watch the coarse resolution (643 Hz) spectrum display. This shows any strong signals in the band. Usually these strong signals are due to satellites. Signals from satellites often show a modulation in signal strength as they pass through the side lobes of the antenna. Other strong signals are due to "astrophysical interference" such as the emission from neutral hydrogen atoms at 1420 MHz.

Note: for background information about the plots seeMCSA Data Displays.
-Peter Backus

5 Dec, 23:30 UT Mountain View, CA
We're still observing! Thanks to persistent investigation by a member of the Green Bank staff, our suspicion was proven correct. The cryogenic refrigerator in our receiver was fine. The problem was much more low-tech. A swivel connector in a hose supplying helium to the receiver had worked loose and was crimping the hose. (Liquid helium cools the low noise amplifiers in the receiver.) After repairing the connector, and allowing the receiver to cool down again, we were back to observing. Total time lost was about five hours.

Currently we're observing Mu Andromeda, one of the stars with a known companion.
-Peter Backus

12 Dec, Mountain View, CA
Now that this observing run is underway, things are going smoothly indeed. We continue to see small fluctuations in the of the 70K stage of our cryogenic receiver, but these can be easily understood as due to slight contamination by oxygen that probably happened when we were experiencing some problems with the helium lines to the compressor in our refrigerator. There has been no degradation of our sensitivity, and after this run we expect that even this small fluctuation can be eliminated by purging the system to remove the contaminant.

It's winter in the West Virginia mountains, so it has begun to snow. Most nights it hasn't amounted to much, but early Tuesday morning the operator shut the system down for an hour when the snow started to build up in the dish. The routine procedure everywhere is to tip the dish towards the horizon and dump the snow out. Green Bank has a unique procedure to deal with those cases where some snow remains stuck to the antenna; the operator throws a baseball at the dish! It doesn't hurt the surface, which is made of steel and built like a battleship, but it takes care of the snow. Two tips: don't stand too close, or you'll become a snowperson, and look for the holes in the snow cover on the ground, as they will help you find your white baseball!

As with all projects, little software bugs continue to surface. On 10 December we discovered another small discrepancy with the way "time" is interpreted at Woodbury vs. Green Bank. A time that is an exact multiple of the hour (hh:mm:00) at one site is rendered as (hh:mm-1:60) at the other, and the software got confused and didn't recognize them as being the same thing! This is just one of a number of standard "zero" problems against which everyone bumps their heads when trying to communicate between different software and operating systems.

This bug came to light when we started to observe the "Horowitz and Sagan" candidates that had been published at L-band frequencies. These are ultra-narrow band detections that are substantially stronger than the expected noise levels (>31.7 times the mean noise power) which looked like the kind of signal the META project was hoping to find, except that minutes later they had disappeared, and were not seen again hours or days later. One explanation is that these really are weak ETI signals that just happened to be fortuitously amplified by interstellar scintillation during the initial detection, and faded away thereafter. Since our sensitivity limit is much better than META's (we use a larger antenna) we are trying to test this explanation by finding the faint, unamplified signal. We looked at all such signals reported at 2480 MHz with our S-band system without success, and so far the two positions we have observed at L-band have also turned up negative. In fact there was an intermittent signal found at a nearby frequency but it changed frequency in a peculiar way, going from -0.8 Hz/s to +0.3 Hz/s, and is unlikely to have registered a detection in the META system. We saw the signal when looking at two different directions on the sky, so we know it was RFI, entering our system through the antenna sidelobes. We are trying to see if we can attribute it to any particular satellite.
-Jill Tarter

18 Dec, 00:15 UT
Observations are proceeding with little excitement (darn). We've observed all but one of the Horowitz-Sagan candidates. The last should be observed later today. I've updated the star-frequency plot to show both L-Band and S-Band observations at Green Bank. This new plot also includes the observations of stars with known planet-sized companions. (Note: These are not yet listed in the Star List web page.)
- Peter Backus

more about Past Campaigns

SETI