Hello! My old SBIG ST402 does not seem to do well under very dark skies - the darkest pixels of an image frequently are darker than the darkest pixels of my master darks - so I’ve needed to do a lot of image manipulation in order to reduce the data. (Or let the moonlight raise the background dark pixel values.)
The QHY miniCAM8 seems like a good camera to use with my 8" Mead LX200 working at f6.3. It would be less expensive than purchasing a camera with a similar chip size and with an add on filter wheel and J-C filters. However, it would come with Sloan filters.
I know we’ve talked about Sloan vs. J-C filters in the past.
My question is how easy is it to request sloan comps for a field? I would prefer to continue to image the targets that I’ve been imaging, but Sloan comps do not come up for most of my targets.
With modern catalogs, would Sloan comps be commonly available for pretty much any field if I request them from the sequence team? If not, I would look to upgrade my system to a different CMOS camera with J-C filters (and probably a different OTA).
Also, are transform fields readily available? Landolt fields? M67? Best regards.
Mike
Sloan standards in Landolt fields are here:
https://ui.adsabs.harvard.edu/abs/2002AJ....123.2121S/abstract
Besides APASS g,r,i data (see the paper just posted in the JAAVSO), there’s also the sky-wide ATLAS ‘refcat2’ catalogue. This can be queried at the CDS VizieR catalogue utility as item ‘j/apj/867/105’. Probably best in that instance to use only stars fainter than about mag 11. Finally, in the south (up to +15 Dec) there is SkyMapper4, VizieR item ii/379. The zero-points of all the above are a bit jittery at the ~0.05 mag level between any two surveys. (APASS DR9 can be 0.1-0.3 mag off, however.)
\Brian
Thanks! So, if I requests Sloan comps for a field, it sounds like the sequence team would be able to do so without difficulty? Best regards.
Mike
Thanks! I think I’ve managed to figure out how to use my old SBIG ST402 with BVIC filters under very dark skies by converting all images to 32-bit float. I’ll keep the possibility of upgrading to a Sloan photometry CMOS camera in my back pocket at the present time in case I need to upgrade in the future. Best regards.
Mike
If this is happening often in non-defective pixels, it suggests you may have a latent issue somewhere in your process, as this really should not be the case. Even under “dark” skies, there is always going to be some sky background, which means a nonzero number of photons hitting the detector,[1] whereas there should be very close to 0 outside/external signal in your dark frame(s).
If you are using a single master dark and scaling it to match your exposure times, this could arise from a mismatch in temperature/gain/offset, inconsistent/incorrect bias subtraction, or potentially that your dark current isn’t constant in time for some other reason. An N-second on-sky exposure should have at least as much signal (barring defective pixels and noise effects) in each pixel as an N-second dark frame with the same settings, assuming N ≫ 0.
Unless you are doing very short exposures where shot noise might outweigh the contribution from the background ↩︎
Hi Mike,
This issue “the darkest pixels of an image frequently are darker than the darkest pixels of my master darks” is probably very simple.
My guess would be your “master” darks weren’t shot under the same conditions, location, temperature, cooler setpoint/ambient temp, or input voltage is.
Another easy mis-step is having a Pedestal value added to the image somewhere in the process. eg subtract two images, and add 100.
Bias level of the sensor may have changed.
If you are running off portable DC power, instead of AC, then you might be putting in a different voltage, and that is affecting cooler operation and inputs to the internal power section. Electrolytic capacitors age over time, and this can reduce effectiveness of power regulation.
So if you’re at a remote location instead of the normal one, you’ll want to redo your bias and darks.
You didn’t say what software you are using or what your process is.
Cheers,
Colin
Hello! Thank you all for your input. I shoot master darks at the temps and integration times that I would use, so I don’t scale or use biases. Typically, I shoot between 25 and 50 raw darks (or more, depending on the run) and average them for the master. When I shoot an image, I select whatever temperature is appropriate and use the same integration time and temperature master dark. For darks, the ST402 has an internal shutter mechanism. I also cap the scope. Otherwise, all darks and target images are taken at the same site and with the same equipment.
No pedestal. The SBIG ST402 cannot adjust the pedestal or offset.
I don’t know why the darkest pixels on checking the histogram images can be 200 to 300 ADU under the master dark at times. When I calibrate these images with MPO Canopus at the native 16-bit depth, it gives a wrap error. Converting all frames to 32-bit float seems to allow Canopus to handle the negative pixels. It is a small added step, but seems to work.
The only change that I can think of from past work is that I always shot from my house in town. So, there was always more background light. I can’t think of anything else or any other test to run. There have been some past forum comments about negative pixels under truly dark skies, so that may be the issue. That, and the 15 year-old CCD is more sensitive to such things.
Again, thank you for the feedback.
Mike
Yep, no problem. Just say you require Sloan filter values and they will be there!
Mike, sequence team
Thank you! I appreciate the guidance.
Mike
Mike,
Are you taking the darks frames during daylight? Is it possible that your system has a small light leak that is causing your darks to be higher? I would suggest that you always save processed images, flats, and darks in floating point rather than 16-bit integers. If you average several darks and then save the result as a 16-bin integer, you will loose some data. The issue is even more important for flat-field correction.
Allen
Thanks! Night time, telescope capped, with the cameras shutter in place for darks.
Mike
This can happen with any camera under normal circumstances. (A non bias subtracted master dark is an average of many darks so any particular pixel in the individual darks (and in the lights taken in a perfectly dark sky) will have values above and below this mean value. Dark subtraction can then create negative values in the dark subtracted light frames but these are to be expected and the data reduction software should handle these correctly
Just dont forget to tick the appropriate boxes for the filters you want in VSP!
Ok, I’m confused now… but that is easy to do these days with AIs running around acting like they know everything… 
Mike, How about you upload one (1) raw science frame, one (1) dark frame and (1) one flat field image that you process and gives you your problem at 16-bit resolution? Any other images you use to process to get the final answer image you do photometry on, like maybe one (1) bias if you use them or etc.
CCD image processing ins’t this hard… 
Jim (DeY)
p.s. Oh, upload the final science image you get for the same set of images.
p.p.s Where are the wrapped problem pixels located? On bright saturated stars or random pixels locations?
OK. Here are the 4 images at 16 bit - the raw frame, dark, flat, and processed frame. The histogram for the raw frame shows the darkest pixel below the darkest pixel on the master dark.
All processing and 32 bit float conversion were performed with MPO Canopus version 10.
Of note, the master 15 sec dark was obtained with an average of 50 raw darks taken at night with the OTA capped and the SBIG ST402 shuttered. There were about 75 raw flats taken at 1 sec integrations during morning twilight with histograms showing ADUs of about 20000 to 30000. The scope was pointed about Azimuth 270 degrees and altitude 88 degrees. The master dark used to process them was 1 sec with a stack of 50. The 75 processed flats were then stacked into a single master V flat. Thanks. Mike
V Master Flat 01April2026.FIT (767.8 KB)
SX HER_LIGHT_V_15.00_4.89_2026-06-17_0008_P.fits (767.8 KB)
SX HER_LIGHT_V_15.00_4.89_2026-06-17_0008.fits (767.8 KB)
Master Dark_15sec_5C_March2026.FIT (767.8 KB)
And here are the images after 32 bit float conversion. The histogram of the 32 bit float processed image does not show the wrap problem of the 16 bit processed image. However, the darkest pixel is at 0 ADU.
Mike
V Master Flat 01April2026_P_32BitFloat.FIT (1.5 MB)
SX HER_LIGHT_V_15.00_4.89_2026-06-17_0008_PRaw_32BitFloat.fits (1.5 MB)
SX HER_LIGHT_V_15.00_4.89_2026-06-17_0008_P_32BitFloatProcessed.fits (1.5 MB)
Master Dark_15sec_5C_March2026_P_32BitFloat.FIT (1.5 MB)
I processed only the 16-bit images. True something not being removed at the pixel level.
The raw science image looks ok by itself. However, if one looks closely you can see that this is the source of the linear warm/hot pixels in the final processed image. Appears in both 16 and 32 bit processing.
The dark looks ok by itself with no linear structure apparent.
The flat has a lot of large dust donuts but also strange linear banding features on one side. The flat has very low signal, somewhere around 6,600 dn. One should try for say around 30K - 37K dn values.
There is a temperature difference where the flat was taken at -5C while the science and dark were taken at +5C according to what is in the fits header.
So bottom line is your science image has the linear structure stuff in it.
Be nice to know if that pattern is stable, unchanging in your science images if you took a series of images one after another. I’ll leave that testing to you.
Jim (DEY)
p.s. That SBIG ST-402 probably should be replaced with something newer utimately IMHO!
Did you ever contact SBIG? SBIG sells 9 micron pixel cameras still but probably high $$$$$$! There are other CCD cameras with probably smaller pixels that might not be so pricy. I love my Atik cameras! ZWO, QHY, SBIG… other vendors out there.
My apologies. I sent the wrong flat! I have them at +5C, 0C, -5C, and -10C at all integration times that I might use.
The dust donuts are close to the sensor. With the SCT OTA closed tube, dust on the corrector plate does not show up, being too far from the sensor, according to the testing I did a few years back. I could get rid of the dust donuts with a cleaning from SBIG, though I doubt that it is worth the cost at this time. They do not carry any spare parts for the ST402.
The camera itself is (or was?) linear from about 2000 ADU to 60000 ADU, so the ADU level in the flat should be fine, though I probably should retest it.
The data looks consistent with light curve data at present, either by converting to 32-bit float, or simply subtracting 200ADU to 300ADU from the appropriate dark frame.
I’ve been looking at new cameras. The Minicam8 looks like it might work well with the 8" Meade LX200, but it has Sloan filters, hence the subject of the post. I would appreciate feedback on this camera, if anyone has used it. Thanks and best regards.
Mike
The pixel pitch is tiny at 3.9-microns for the miniCam8 but the array size is much larger so you should get similar total field of view. There are other brands with larger pixel sizes you should consider perhaps. The integrated filter wheel is interesting and nice. The sensor looks good, back-side illuminated, high QE! Probably can’t go wrong with it at those prices. Only you can decide if Sloan filters are what you want.
If you don’t want the Sloan filters you could get one of the LRGB+ sets. In my experience RGB can be transformed pretty well to the standard Johnson-Cousins system. The L is just a clear filter and you could get a few narrow bands for fun imaging!
Good luck.
Jim (DEY)
Thank you! I appreciate the guidance.
Mike