Skip to main content

Exercise #4 - Calibration Procedure (CMJA)

1 post / 0 new
CMJA's picture
Exercise #4 - Calibration Procedure (CMJA)

I have a Meade 16" ACF on a Mathis fork mount in a Home Dome.  The CCD is a QSI 516ws with a non-antiblooming detector and internal BVRI filter wheel. This camera can cool to -35 below ambient which is easy to get during cold during winter nights but I can manage to get down to -10 to -15 in summer.  I usually wait for about 30 minutes for the CCD to reach thermal equilibrium, and cool as low as possible while ensuring than the CCD cooler is operating at a power level of not more than 80% (per the OEM recommendation). The SCT design and sealed system of the imaging train (I have a focal reducer in front of the CCD optical window) means little or no dust can create a problem for taking flats, as long as I don't rotate the camera or change anything else in the optical path. The mechanical shutter of the QSI 516 means that I can get darks in the dark.

After learning about a calibration pipeline, I decided to acquire ACP software to run my dome, telescope and CCD camera. The nice thing about ACP is that imaging runs can be scripted.  ACP has "directives" to take auto twilight flats at dusk and/ or at dawn, as well as bias and dark frames.  The user has complete control over the parameters under which these frames can be taken. For flat frames settings include:

  • The mean pixel value I want for the flats (I set this at the default 32768);
  • The acceptable variation in mean background ADU (I set this at the default 4000);
  • Minimum exposure (I set this at a recommended 0.5 seconds); and
  • Maximum exposure ( I set this at 10.0 seconds)

ACP uses user-generated "plans" to tell ACP scripts how to handle things. For ACP's autoflat script, a flat plan is required which basically outlines the number of flats for each filter (BVRI) and in which filter order the flats will be taken. Binning values can also be set-up in the plan for each filter.

Because the sky is more transparent to longer wavelengths i.e. red, there is an order that the flats for each filter are taken. If I'm doing multi-band (BVRI) imaging, my filter order at dusk will be BVRI, at dawn it will be IRVB.  ACP determines the correct altitude/azimuth to point the scope to optimize twilight flats. As flats are being taken, ACP issues "micro-slews" to ensure that the same stars do not show up in different flat frames (yes, even in twilight a few stars can be seen in the flats). I inspect and discard any flat frames which have any visible stars. I occasionally get 1 or 2 frames showing up, but I take lots of flats which are median combined to get a master flat.

I'll take between 15 and 30 flat frames depending on the season (twilight lasts a long time during June at my latitude).

After ACP completes flats (and while awaiting toward the end of astronomical dark), ACP will "chain" to another plan to take flat darks and darks. I'll inspect and discard any flat darks and darks that have major cosmic ray hits.

I don't take bias frames, preferring instead to take darks with the CCD cooled at the same temperature and that match the science frame exposures I anticipate to use for the night. I rather avoid calibration involving bias frames since a little noise can be introduced. I recently collaborated on a project where the target was down around 19 mag. and SNR was low.  I didn't want to introduce any more noise that I had to. Most of my images for photometry targets are in the 30 to 180 second range, but I've had to take 10, 600 second exposures and stack to get enough SNR for photometry.

Flats, flat darks, and darks are saved to the same folder where MaxIm DL looks to generate master frames. MaxIm DL is great at setting up calibration groups to create masters.

Once science frames proceed during the night (run by another ACP plan) each science frame is automatically calibrated immediately after download. Prior to doing autocalibration, ACP saves a copy of the image with "-RAW" appended to the science frame's name. This enables me to still have the original uncalibrated science frame if, for some reason, calibration had undesirable effects.

The next day I can perform data analysis using either MaxIm DL or AIP4WIN, either of which allows me to create a photometry report in AAVSO format.  After hearing about VPhot at the 2012 SAS symposium, I will look into leveraging it's capabilities.

I recently implemented a darks library at 2 degree C steps and plan to rebuild this library semi-annually or quarterly. But I  need some empirical data to determine how frequently I should do this.

The other thing I'm wondering about is whether I should save all frames as 16-bit integer, 32-bit integer,  or IEEE float format.

Feedback welcome.


AAVSO 49 Bay State Rd. Cambridge, MA 02138 617-354-0484