Thursday, February 4, 2010

Looking at the Data

It's been several weeks since I posted. The major reason for the gap is that my father passed away in early January, and it's been tough. I'm very glad that this didn't happen while I was in Antarctica.

In the past few weeks, we (the royal we - mostly, it has been Steve Barwick, Jordan Hanson, Lisa Gerhardt and Ryan Nichol) have been looking at the data, which is flowing smoothly through the internet link. Mostly, things look very good, and the station continues to work well. This data falls into two classes: housekeeping data, relating to station performance, and triggered data, collected when the antennae see something).

Lisa Gerhardt posted some of the early (through early January) housekeeping data on the web Please remember that not all of this data has been calibrated yet. The temperatures in particular seem low. That said, a few trends are clear. One is that the wind kicked up shortly after we left (starting around Dec. 26th), and again in early January. So, there may be hope for the wind generator. A second is that the diurnal (day-night) variations are present, but small.

The antenna data is harder to describe here, but we are also making progress. Some background might be helpful here. We collect data (and call it an event) whenever the signals on two of the four antennae are above an adjustable threshold. These signals could come from background noise, natural sources, or man-made sources. When a trigger occurs, we record the data for each antenna for 60 nanoseconds (billionths of a second). This may seem like a short interval, but a real neutrino event should produce a pulse that is less than a few nanoseconds long.

One thing that jumped out early was that the trigger rate was partly periodic. We saw triggers every 60 seconds, as expected, These are the 'heartbeat' pulses that we create to check the detector. But, we also see other, unexpected periodic signals. Sometimes, triggers are separated by almost exactly 6 seconds - pretty clearly, a man-made source. The rate of these pairs varies over 24 hours, and we suspect that it comes from the switching power supplies that power the internet hardware. If so, this will disappear when the internet equipment is removed. We also see other events not related to these periodic signals. These might be 'thermal noise' the irreducible background associated with random noise due to molecular motion, etc. This thermal noise provides the 'natural' limit to detector performance. There are ways to reduce it (better antenna, lower noise preamplifier, etc.), but this thermal noise limit is our immediate goal.

No comments:

Post a Comment