I’ve been trying to nail down the transformation coefficients for my Seestar S50 from TR, TG, TB to Johnson-Cousins(J-C) R,G,B. I don’t know if this is universal for these instruments or just my unit. I’m generally trying to capture eclipses on the Eclipsing Binary section legacy list and I think that untransformed TG is good enough for that. However, I want determine the transformation to J-C VBR for other stars.
Every clear moonless night I try to image a standard field or cluster to as it crosses the meridian for my location. When I generate coefficients with Transformation Generator, the results from imaging M67 seem to be systematically different from imaging Landolt fields. Has anyone noticed this? I’ve tried to carefully select stars in M67 with a good color range, are not overlapping with the S50’s resolution and have a good signal to noise ratio (SNR).
The images are taken slightly defocused because of the one shot color chip, and are calibrated with flats. They are also calibrated with darks taken during the imaging session so the temperatures are similar if not the same. I debayer the colors and stack to improve the SNR.
I’m next going to try to determine the quality of my flats to see if there are any systematics. I guess what I’m asking is: has anyone run across this issue with the S50?
I’m using NINA to capture data and given a clear night and good horizons, an avalanche of data can be produced from these units.
It is pretty unlikely that you need flats given the size of the Seestar’s sensor. Also, the FITS files it produces are already dark-subtracted, so you might actually be introducing some errors if you’re manually subtracting dark frames again.
Just how far off are your results from the Landolt fields vs. M67? Are your results for a given field consistent from night-to-night?
Thanks for replying so quickly. I’m using NINA to acquire my images so I do need to subtract darks. Applying flats are what I’ve always done for photometry as part of good practice.
I’m going to review all of my images of standard fields and reprocess them again and see if I get the same results.
Using a Bayer matrix sensor the transform coefficients will generally have more scatter from night to night. Many reasons why…
We really need to see the values of the coefficients and the least-squares generated coefficient standard errors to evaluate the significance of any “bias” you mention.
I keep a roll-up spreadsheet that contains all the transform coefficents, coefficient standard errors for the large number of sensor-OTA combinations I have used over the years. Due to scatter from night-to-night I use the long-term average transform coefficients. For one DSLR-OTA I have 60 coefficients taken over one year using a mix of cluster and Landolt fields. For one CCD-OTA I have 44 coefficients taken over four years.
I must admit that I have not analyzed the transform coefficents for bias evidence between Landolt and other standard fields. Any actual bias would be very small and not likely detectable by amateurs. All the scatter I’ve seen is attributable to small numbers of stars in the L-S fits and Bayer matrix sensors from slight focus differences which can cause more scatter in the transform coefficients.