SETI Institute

Home News About us SETI Carl Sagan Center Education and Public Outreach Publications Support us teamseti

What a Difference a Week Makes!

December 05, 2003

by Dr. Jill Tarter, Center for SETI Research Director

This entry was written on Monday Nov 24, and Jill has noted that observations have continued smoothly without any more surprises since then.

Monday morning as I headed down for a run -- well -- jog (stagger, gasp, sweat) around the dish, I ran into Murray Lewis who is on the science staff at the observatory.

How are things going? he asked.

Terrific, I said.  Last night we were actually not scheduled on the telescopes but we were able to run a bunch of statistical tests with our noise source to fine tune our detection thresholds within 1/10th of a sigma.

What this means (for those who may not understand the statistical terminology) is that weve tweaked up the system to within an inch of its life.  When you worry about a 1/10th of a sigma you dont have much to worry about!

I remembered the same sort of early morning discussion Id had with Murray just a week prior.  Wed had two nights on the telescope.  The first night, the telescope was not working properly and failed to move in azimuth.  That probably had something to do with the 27 inches of rain that fell over the previous 5 days it isnt clear what was actually broken.  The telescope crew took apart the hardware control system, and put it back together, and it started working before our second night of observing.  Fortunately, because we always spend our first night of observations doing drift scans to prime our RFI (interference) database, the telescope didnt need to move, so and we could and use it productively on the first night.

By the time I first chatted with Murray, wed had a night of active observing and what a strange night it had been.  Wed masked out the bands that wed learned from the first night were contaminated by RFI, and yet we still had strong RFI signals interfering with our observations.  We stopped observing stars, performed more RFI scans, and the results were the same as the night before!  So we masked out the newly-offending bands and started observing stars again. More RFI at other bands!  And whatever these new signals were, they had a suspiciously repeating form of modulation a sinc function, always about 200 kHz in width.  By the time I shared my puzzlement with Murray, we were beginning to suspect maybe the drive motors on the antenna were generating noise that we were detecting with this new S-band feed and receiver (we are an early-adopter of this very wideband system). 

At Arecibo, we are under enormous pressure to use our observing hours efficiently, and we are reluctant to stop observing stars to do testing.  Therefore, we continued observing while investigating the reappearing RFI.  We considered many explanations, and systematically ruled them out.  We scratched our heads.

We thought about a lot of things for three nights, until we finally bit the bullet and decided to sacrifice observing time and go into full-blown test mode.  After many hours of fussy work on Friday, we found the error buried deep within our software.  This problem had never surfaced in the past because the two separate racks of equipment that interface to the telescope and deal with the frequency tuning and conversion had been at two separate telescopes (Arecibo and Jodrell Bank), tuning to the same frequencies.  This is the first observing run where we have put the racks side by side and asked them to observe two contiguous blocks of frequency, and in this configuration, we experienced the spurious RFI because our frequency calculations were slightly in error.

We corrected the code, saved the data that we had collected during the previous nights, and started again.  If during the next few weeks, there is an interesting candidate coming from a star at a frequency range previously observed during our first three nights, well do some reverse engineering in order figure out what the actual frequency was in the early data, to get some history for that signal.  Theres a lot of system testing that can be done in the lab, but finding surprises is the norm when a new system first looks at the sky.

By my second pre-jogging conversation with Murray, I was able to tell him of our excellent observations Friday night (the sort that get boring before the end of the shift which is a good thing, and why I have lots of loud samba music to play through the night).  With luck, Murphy (not Murray) will be done with us, and things will stay routine and boring, without any more surprises.  At least until the boredom is broken by the detection of the signal we are all searching for.