Trouble calibrating flats

[johnw3d]

Hi all.  I'm having a bit of trouble calibrating my flats.  From the same flats run, half will scale the dark master successfully, the rest warn of no correlation with the dark master.  I've been studying carefully all the posts and Vincent's tutorial on master generation, I think I've got them right, at least all the lights calibrate OK - integrate the biases to get a bias master, then calibrate the darks with that bias master, then integrate the darks, so you get a scalable, bias-subtracted dark master.   

I'm wondering if the large difference in exposure times between the dark master subs (for 2x2 eg, 5 mins @ -25ºC) and the flats (for 2x2, 1 to 3 secs @ -25ºc) is making this troublesome.   The camera is a QHY9M with a KAF-8300 chip, running with a pedestal of 500, offset of 112 and gain of 10% to give me about 80% of full-well on max exposure.   The master bias averages about 740 ADU, the bias-subtracted dark master averages about 70 ADU.   Am I running with an inappropriate pedestal, should that be included in the calibration setup (it's not in any FITS key in the files)?

The masters comprise 25 subs, might that be too few to get acceptable SNR?  I see Vincent suggests a minimum of 10 and Mike around 100.

Any suggestions?

Here are some good flats:

Writing output file: /Users/john/Pictures/AstroMBP/caps/2015-07-25-IC1396/preprocess/IC1396_1sec_2x2_G_frame15_bs.fit
Dark scaling factors:
k0 = 0.006
Gaussian noise estimates:
s0 = 3.573e-03, n0 = 0.953 (MRS)
52 FITS keywords embedded
Writing FITS image: 32-bit floating point, 1 channel(s), 1674x1248 pixels: done

Writing output file: /Users/john/Pictures/AstroMBP/caps/2015-07-25-IC1396/preprocess/IC1396_1sec_2x2_G_frame16_bs.fit
Dark scaling factors:
k0 = 0.007
Gaussian noise estimates:
s0 = 3.568e-03, n0 = 0.953 (MRS)
52 FITS keywords embedded
Writing FITS image: 32-bit floating point, 1 channel(s), 1674x1248 pixels: done


Here are some uncorrelated flats from the same run:

Writing output file: /Users/john/Pictures/AstroMBP/caps/2015-07-25-IC1396/preprocess/IC1396_1sec_2x2_G_frame12_bs.fit
Dark scaling factors:
k0 = 0.000
** Warning: No correlation between the master dark and target frames (channel 0).
Gaussian noise estimates:
s0 = 3.567e-03, n0 = 0.954 (MRS)
52 FITS keywords embedded
Writing FITS image: 32-bit floating point, 1 channel(s), 1674x1248 pixels: done

Writing output file: /Users/john/Pictures/AstroMBP/caps/2015-07-25-IC1396/preprocess/IC1396_1sec_2x2_G_frame17_bs.fit
Dark scaling factors:
k0 = 0.003
** Warning: No correlation between the master dark and target frames (channel 0).
Gaussian noise estimates:
s0 = 3.571e-03, n0 = 0.953 (MRS)
52 FITS keywords embedded
Writing FITS image: 32-bit floating point, 1 channel(s), 1674x1248 pixels: done

Thanks very much,
John.

[pfile]

Hi John,

i don' t think this is totally unexpected because of what you pointed out - the difference in exposure time is huge.

i've always calibrated my 8300M flats with a master bias alone, and it's worked OK for me.

rob

[jkmorse]

+1 on only using a master bias with flats.

[johnw3d]

Thanks chaps.  It appears that the Preprocessing batch script in PI is attempting to do a dark subtraction on the flats as part of it's preprocessing workflow, and there doesn't seem to be a way to disable that, so perhaps I'll just do the flat calibration & integration by hand.

Thanks again,
John.

[johnw3d]

Confirmed again that the PI preprocessing script is indeed attempting a dark-calibration of the flats. 

Is it a safe assumption that the scaling of the master dark for the flats would be so substantial that there would be essentially no effect if some of the flat subs were not dark-subtracted, given the averaging in integration that follows?   I do find that script a time-save and finger-trouble reducer.

Thanks again,
John.

[pfile]

yeah, i assume that it's probably a no-op. i guess it would not be that hard to hack the script to not calibrate with a dark if the exposure length is less than some threshold.

rob

[Geoff]

Thanks chaps.  It appears that the Preprocessing batch script in PI is attempting to do a dark subtraction on the flats as part of it's preprocessing workflow, and there doesn't seem to be a way to disable that, so perhaps I'll just do the flat calibration & integration by hand.

Thanks again,
John.

You can still use the BPP script in two steps:
1. Run with flats and biases only. That will generate master flats and master bias (You can safely ignore warnings about no darks, lights)
2. Run with darks, biases (or master bias), master flats and lights.
Geoff