Skip to main content

Exercise #4: WBY Calibration procedure

6 posts / 0 new
Last post
WBY
WBY's picture
Exercise #4: WBY Calibration procedure

The calibration procedure I will follow follow varies slightly depending on whether I am using my home observatory (well, really a telescope shed) or the Meyer Observatory of the Central Texas Astronomical Society. AT CTAS the Princeton instruments camera cools to -50C regardless of the ambient. At my home observatory my ST7 camera has much more limited cooling. I set the camera to whatever increment of 5 deg. C allows the cooler to work at about 85% of capacity after stabilizing at the temperature set point. The Princeton Instruments camera stays at exactly -50C. my camera keeps within +/- 0.2 deg C of the set temperature as long as i stay below 100% cooling capacity. Then, I focus and start the night's run. The CTAS camera is always pointed N up. The camera in my home observatory can be rotated and can rotate through a small fixed angle on its own at the joint between the filter wheel and the camera. After positioning on a target I make sure that the camera in my observatory is rotated CCW hard against one of the fixed stops. Then I calibrate tracking, and adjust the FOV so the target and comps are well positioned. I note the x,y positions of the target centroid and one other star centroid that is relatively near the edge of the frame to be able to note and correct if the drift is more than a few pixels, and to confirm that no camera rotation has occurred during the image sequence.  If I think that I may have to reposition the camera during a sequence, due to a meridian flip, for example, I will take flats in all applicable filters before starting the sequence of images. Then I will take another set of flats at the end of the sequence. At both observatories we use an adjustable light source so that flats though all filters can be taken to approximately half well depth in 5 seconds. At CTAS we take dome flats. At my observatory I use a light box.  I always take a complete set of flats in all applicable colors before repositioning the camera.  If the camera isn't repositioned then I take all the flats immediately after the night's runs. After flats, I set up an automatic sequence to take 30 bias frames and 30 darks for each light frame duration plus a set of 30 darks at 5 seconds for flat correction.  I visually check the first bias frame when it downloads to make sure it looks "normal" for the camera. I generally  close the dome (CTAS) or the roof (home) when the darks and flats are being automatically taken.

I do not scale darks. From the standpoint of standard deviation in dark current and deviation in unpatterned noise, i would be better off just using 30 dark frames of the longest exposure for the night. However, hot pixels can get into the non linear or full well capacity of pixels and therefore don't scale properly.  As long as temperature can be maintained dark acquisition can be done in daylight hours. Both observatories remain relatively dark during daylight hours and any part of the image train that can admit stray light can be covered easily. I do not take flats in daylight. Even with a light box that laps over the end of the telescope, there is too much opportunity for stray light to enter the telescope.

I use library darks and bias frames only when, for some reason, they could not be taken on the night of the run or within a few days of it. Pixel responses, and readout noise can change with time and I have observed changes in some cameras over a period of several months. I never use library flats.

 

Calibration

Before taking data, a directory is set up with the name format "Object Name_YYYYMMDD" using the UT date (Object directory). All light frames, Bias ,Darks, and flats are saved to an "Original" sub-directory of the Object directory for each object even if that means common calibration frames are duplicated several times. The Original directory is given read only access. The entire contents of the Original sub-directory are then copied into a "Working" sub-directory of the Object directory. Master calibration frames are saved in the Working folder with a names Master Bias.fits, Master Dark_<seconds>.fits, Master Flat_<Filter>.fits." Automatically generated log files of the calibration chain are also stored in the Working directory. At the the same time that the Working sub-directory is created, Calibrated", and "Photometry" sub-directories of the Object directory are also created. The Calibrated light images are saved in the Calibrated folder. The output of the photometry measurement tools and the associated log files are also stored in the photometry directory. Directory creation and file movement are manual operations at this time.

All calibration frames are at the same temperature.

Bias frames are averaged into a master bias

Individual dark frames are bias subtracted using the master bias frame, and dark frames of equal duration are median combined into master dark frames.

Flat frames are bias corrected using the the master bias and dark subtracted using the master dark of equal exposure duration. Then the individual flats for a given filter are median combined and normalized to make a master flat frame for each filter.

Light frames are bias corrected using the master bias and dark subtracted using the master dark of the same exposure duration. Then they are flat field corrected using the normalized master flat with the same filter designation.  This process is highly automated using Mira-Pro. I Create lists of the light, bias, dark and flat frames using a Mira-Pro utility, create a calibration tree or modify a stored template and do the entire calibration, including header modifications, if required, in one step for all images.

Although not highly automated except for the steps Mira-pro automates, I think this is a reasonably complete process and I plan to continue using it.

Aaron Price
Aaron Price's picture
comments

[quote=WBY]

Light frames are bias corrected using the master bias and dark subtracted using the master dark of the same exposure duration.

[/quote]

I wonder if you even need to do bias correction? If the darks are the same length as the light exposure, then bias should be included in the subtraction of the dark image.

[quote=WBY]

As long as temperature can be maintained dark acquisition can be done in daylight hours. Both observatories remain relatively dark during daylight hours and any part of the image train that can admit stray light can be covered easily. I do not take flats in daylight. Even with a light box that laps over the end of the telescope, there is too much opportunity for stray light to enter the telescope.

[/quote]

I am interested as to why you are more concerned about ambient light interfering with flats than you are with darks. I would think that if you think there is enough stray light to mess with a flat, that would be enough to mess with a dark too.

HQA
HQA's picture
flats and darks

[quote=WBY]

Light frames are bias corrected using the master bias and dark subtracted using the master dark of the same exposure duration.

[/quote]

[quote=Aaron Price]

I wonder if you even need to do bias correction? If the darks are the same length as the light exposure, then bias should be included in the subtraction of the dark image.

[/quote]

Mathematically, subtracting a master bias followed by subtracting a master dark is the same as subtracting a master (bias+dark), if the temperatures are matched.  We do this with APASS, since the images are 4kx4k and doing one subtraction instead of two is a time-saver.

[quote=WBY]

As long as temperature can be maintained dark acquisition can be done in daylight hours. Both observatories remain relatively dark during daylight hours and any part of the image train that can admit stray light can be covered easily. I do not take flats in daylight. Even with a light box that laps over the end of the telescope, there is too much opportunity for stray light to enter the telescope.

[/quote]

[quote=Aaron Price]

I am interested as to why you are more concerned about ambient light interfering with flats than you are with darks. I would think that if you think there is enough stray light to mess with a flat, that would be enough to mess with a dark too.

[/quote]

For flats, you have scattered light entering the front of the telescope as well as any light leaks around the camera.  For darks, you are only concerned about the light leaks around the camera.  So taking flats after sunset is a good idea.  I also prefer to take darks after sunset, since getting rid of light leaks is sometimes more difficult than you realize - the eye is not nearly as sensitive as the CCD.  In almost every location (except CTIO/Chile!), you get plenty of cloudy nights for which you can take dark frames.  Taking darks every night is not essential, since the camera is temperature controlled and the darks are repeatable - you might save some time by just accumulating a master library of darks vs. exposure time for a given temperature setting.

One other question - you take 5 second flats.  For SBIG cameras and their photometric/vane shutter, this is ok.  For other cameras, such as Apogee or FLI that use an iris shutter, this is pushing your luck - you will get "shutter vignetting" depending on how long it takes the shutter blades to open.  What kind of shutter does the Princeton Instruments camera have?  I'd suggest using 10-second exposures if possible.

Arne

WBY
WBY's picture
Bias Subtraction

You are right. If the darks are the same time and temp as the data frames I don't have to do bias subtraction. I do it anyway as a matter of course because I have a standard calibration pipeline set up and if my darks end up not being exactly the same duration as the image I am correcting I don't have to worry when the darks auto scale. Sometimes I get really stupid late at night when I am doing darks and get the time wrong. I make a 60 second dark, for example instead of a 50 second dark. My image reduction pipeline is set up to always use the closest master dark time and scale if necessary. Always bias correcting eliminates the potential for making a mistake. I suppose the bias subtraction adds a little noise, but not much.

WBY
WBY's picture
Flats, Princeton Insturments Camera

Arne,

Your comment about 5 seconds is good one. This camera has an iris shutter, but it seems to be fast compared to, say, the shutter on an FLI camera. We should compare 5 second vs 10 second flats and see if there is a difference in vignetting. We will check it.  

 

Thanks.

WBY
WBY's picture
Stray light for flats vs. darks

When taking darks the camera shutter is closed the scope is capped and the observatory is closed. When taking flats the shutter is open and the observatory is open since my light box won't clear the roof. Stray light can enter the optical path into the front of the telescope even though the light box extends about 5" down the sides of the telescope and is covered in black craft foam. Also, my focuser is not closed. It has two parallel plates with an open gap in between (Fingerlakes DF2). No problem in the dark unless someone shines a flash light at it but in daylight you can introduce large gradients from light entering the focuser. The telescope is a solid tube. even if I drape a thick black cloth like polar fleece over the light box and telescope, I can still get stray light in through the focuser which causes gradients in the flats.

 

With the camera shutter closed I don't have the problem and If I am taking darks during the day, the telscope is capped and the observatory is closed and quite dark inside.

Log in to post comments
AAVSO 49 Bay State Rd. Cambridge, MA 02138 aavso@aavso.org 617-354-0484