Data!

The past couple weeks have gone really well. I finished my detector and was able to start taking some preliminary data. My photmoeter is really just a high time resolution rate meter with many input channels (I have 8 active and can run up to 14 with some slight modifications). Apparently, a multi-input rate meter does not typically exist in the market. One of my colleges here at VERITAs needed such a device in order to determine some base characteristics of some new pixels that were installed in the camera of Telescope 3 (in preparation for a major overhaul of all the cameras that will take place in ~2012).

He was pleased to hear that my photometer, while not quite ready for taking astrophysical data, was more than ready to do the simple task of keep track of the rates on each of the pixels. The procedure was simple: Install my device and calculate the rates for a variety of different trigger thresholds. A trigger threshold is the minimum size of a signal that we require in order to register a 'hit'. If the threshold is set too low, we will trigger on electronic noise and light from the night sky, if it is too high, there will be too few incoming photons and it will take too long to acquire data. Hence, we must find the middle where we're not triggering on electronic noise or night sky background, but the threshold is also low enough to provide enough statistics. This measurement is called an "L1 Bias Curve".

After a few nights of L1 Bias Curve-ing, I caught a lucky break. There was about 45 minuts of moonless time available, which is below the cutoff for which VERITAS starts gamma-ray observations. Typically, one uses this time to take calibration data, which is often not as involved as taking gamma-ray data, but the 45 minute window was right before the Sun came up, from 5 - 5:45AM, and no one wants to stay up that late to get 45 minutes of calibration data.

Considering that I wasn't going to get *any* dark time at all, I got permission to take data during this time and took 45 minutes worth of observations on the Crab Nebula. THere is a well known ~33ms optical pulsar inside the Crab nebula, but it is very faint (Magnitude 16.5, this corresponds to something like 4000 photons / m^2 / s reaching the Earth in the wavelength range that VERITAS is sensitive to) and hence very difficult to observe with a traditional optical telescope. VERITAS has such a huge collection area, ~100m^2 that it makes it feasible to detect the pulsar for long enough runs.

I calculated that in order to get a reasonable signal-to-noise ratio (ie, 5 sigma for a detection) VERITAS needs about 2 seconds of data in order to detect the optical pulsar. The catch is that a VERITAS time detector doesn't have alot of live time (ie, it can take data for ~200ns at a time for at most 500 times a second - this correspinds to the detector being 'alive' for about 1/100th of a percent of the time). This means that in order to get two seconds of live time, one needs to take about 5.5 hours of continuous data. This is a significant amount of data, and VERITAS time is already oversubscribed by a factor of about 5 (ie, we have requests for 5 times more time than we actually have available). Conversely, my detector has 100% live time, I can continuously take data, so 2 seconds of live time is 2 seconds of data. This makes detecting things like the Crab Optical Pulsar much easier, but the devil is in the details.

Remember I was talking about thresholds? Well, the light coming from the Crab comes at various intensities, so the magnitude of the pulsar light could often dip below my threshold, which I have set so that I'm not triggering on noise. This essentially just amounts to needing to take longer data runs in order to get a detection. Assuming I'm only triggering on 1% of the Crab photons (the others being too low intensity), I still only need about 90 seconds of data. So, 45 minutes should be much more than enough to get a detection, unless I'm capturing some absurdly small fraction of the incoming photons.

I haven't had the chance to analyze any code just yet, I've been too busy with other things, such as working on the actual alalysis code and filtering out sources of noise (or at least trying to, it is very nontrivial to do so). I have determined that there's a 1kHz source of noise somewhere in the VERITAS system. Usually, VERITAS operates with a threshold higiher than that of the of the noise, so it's not a major issue, but I am operating with slightly lower thresholds that put me right on the border of being in the area where I can pick it up.


So that's work aside. I'm currently sitting in a garage having my car worked on in preparation for my depature from the Whipple Observatory on Saturday. Just some routing stuff, but I'm also having the timing belt changed. It's not due just yet, but I'd rather not have it break while I'm somewhwere with a population density of zero.

Stops on the way home include the Johnson Space Center (Houston), the Marshall Spaceflight Center (Huntsville, AL) and the Glenn Research Center in Cleveland. I'm really excited for these, but particularly I was invited to the Johnson Space Center by someone who works in ISS operations. I got word yesterday that it'll be ready by the time I'm in Houston.

Well, I have to finish (/start) packing up my room in preparation for my departure. There's US Thanksgiving dinner on the Mountain soon, too, I heard a rumor that there is 40 pounds of turkey lurking somewhere.