In the last post I explained that things were looking up. I’d found a new eclipsing cataclysmic variable – hopefully the first of many. Unfortunately, the discovery wasn’t quite as brilliant as I had first thought. It turns out that during the observations I had made, the CV must have been in a mild outburst state, significantly brighter than its normal quiescent state. The data from just a few night’s later looked like this:
The system was obviously much fainter than in the previous observations, and the eclipse is very difficult to spot. It’s still there, thankfully, but with a much fainter target it will be more difficult to determine an orbital period. It will still be useful of course, just a bit more challenging to assess.
Yesterday I found another eclipser. This one has a much shorter period, so it should be easy to accurately measure it for future study, even though it’s also quite faint.
Aside from the eclipse hunting, I’ve also been busy with the ULTRASPEC commissioning work. One of the most important tests when commissioning a new instrument is to check that it performs as well as expected. The main thing here is to compare the observed throughput of the instrument with a theoretical throughput based on a model of system. As you can imagine, the model of the system is where the hard work comes in.
I spent several slow and tedious days working on a spreadsheet which modelled the throughput of every component in the system, as a function of wavelength. This included the atmosphere, the telescope mirrors, 11 lenses of different thicknesses and compositions, filters, adhesives, and the efficiency of the actual detector. I had to be very careful running anything else on my computer at the same time, as it really struggled to manage such a hefty amount of data. The results however, do look rather pretty.
Code enthusiasts will lambaste me for abandoning programming in favour of the ugly spreadsheet. I am sorry, I really am, but ultimately I think it was easier this way. It was important to be able to visualise each component individually, and I think it would definitely have taken me a lot longer in Python. Plus, my supervisor encouraged me to use a spreadsheet, as they are much more transferable to other users. The second part of this webpage shows my recent work.
The final numbers show a noticeable under-performance when compared with observed data. It’s not disastrous, but it is annoying. We could attribute it to any number of other factors apart from the instrument optics themselves (e.g. under predicted atmospheric extinction, over predicted mirror or CCD quality), but ultimately it’s likely that the lenses are performing slightly worse than expected.
In other news, I’m heading to Poland on Tuesday morning for a workshop/conference, in Warsaw. I’m a bit apprehensive, mostly because I won’t really know anyone there, and the journey is a pain in the arse (overnight in East Midlands Airport before a 7am flight). I’ll be giving a short talk which you can see here, but the main reason for me going will be to find out how our robotic telescope can help get the transients phase of the new Gaia satellite project up and running. Here’s hoping it’s a productive week!