Week 10: nearing conclusions
This week I felt like I didn’t really accomplish anything. I’m sick and my whole body hurts, and there’s the whole global pandemic while I’m trying to figure out summer internships. I tried to get photometry to work for the binary targets in our Taurus subfield, but I couldn’t get our algorithm to treat the targets separately. Good thing is, this seems to be a me problem because my group partner was able to do it just fine.
The first thing I tried to do was change the input parameters on the front end around. I explored a wide range of aperture radii, sigma thresholds, and FWHM values netting me results no different than the last. The problem is that I can’t get the algorithm to recognize the blended PSFs as separate stars and so when the algorithm goes to match my target list with its own list of recovered stars, it just matches the two separate targets in my list to the brighter of the two components of the binary system.
So I tried changing the distance that the algorithm will look for a match to something smaller than the distance between the centers of the two PSFs. To do this I added a keyword to the “diffPhot_IndividualStars” function called “max_sep” which specifies this search separation. Again no luck: lowering this search separation meant the algorithm simply couldn’t match my target list to any recovered star. It appears at a search separation of about 3.5 arcseconds, the algorithm will match the dimmer binary targets to their brighter companions’ recovery. This leads me to think maybe I need to keep the search separation below this value and change the star extraction parameters further. This also went nowhere.
I also tried making a suite of generalized light curves for the TESS field but that kinda crapped out too. Here are some results of that.
Update 8:10pm, 4/14/2020, I did get my code running on at least the most easily recoverable TESS targets. Below is a gallery of their comparison lightcurves.
In terms of finding relationships or analysing our data in any meaningful way, Lena put together this really cool plot of best fit period (for both filters) and mass. I think there’s such a scatter in Ha-off because the SNR for our that filter is much lower, and perhaps the Ha trend, strangely is a better representation of the period-mass relationship, even given that we expect accretion to be more stochastic and in general Ha period fits for these objects is probably unreliable.
Getting poster outline feedback was fun because it was really just Noah roasting me for not having very much on our outline, which is fair. My grasp on reality right now feels like flying a burning biplane blindfolded. We’ll probably manage to scrape something together for our poster draft.
Until next week.