STF-8300 firmware upgrade makes DC offset variation worse

My STF-8300 has always had about 8 out 1000 bias frames with anomalous DC offsets (after a previous upgrade cured the 8 out of 1000 frames that were totally blank). Most of these are low although, in the past, some of these were quite high (>2000 instead of ~1039). After the recent upgrade to stf-8300usb.bf2 to version 2.60 the rate of these anomalies has increased to about 150 out 1000 with only 1 or 2 high values. By low I mean ~950 instead of ~1039 ADU. By high I mean ~1120 instead of ~1039 ADU. The anomalous values also apply to one and two minute darks and lights but are a bit harder to see. Normally the average clusters around ~1039 within 3 ADU so these stand out like a sore thumb.

My software and firmware are up to date and the images were taken with CCDOps (0.09 second exposures). The image average was performed by my own software but I have checked it using Nebulosity4 and ImageJ. I have also looked at overscan as a way around this nuisance with considerable success. The anomalous values apply to the overscan pixels as well as the normal image pixels .

This could be just my camera or it could be a design or firmware problem so it would be nice if someone else would take ~1000 bias frames and inspect the average of each image to see if they see the same anomalies.

 

I forgot to mention that I did a setUSBoffset to no effect and the auto bias feature was off.

 

Should I remove the overscan to take images? I have 1000 with overscan of 200.

 

I also forgot to mention that the standard deviation of each image is nearly identical to the others, i.e. the DC offset is the problem (not the gain).

 

Bias_075 has a low average of 914.10 and an rms of 28.94

 

Bias_100 has a normal average of 1041.62 and an rms of 28.95. Both files were taken with and overscan of 200 but the averages and rms values only included the normal pixels.
There seem to be three distinct areas of overscan pixels whose averages are 873.27, 917.24 and 948.83 for Bias_075 and 1000.83, 1045.52 and 1076.27 for Bias_100.

 

After upgrading stf-8300ga.bf2 from version 2.24 to version 2.25, I took 100 Bias frames with no anomalies and then 1000 more, also with no problems. I don't know what you did but it certainly works! Thankyou! I suggest that you include this bias offset check in your suite of tests that you run after each software and firmware change. I will run some further tests to see if the overscan calibration scheme I produced has any advantage over no overscan in the dispersion of the image averages.

 

The rms dispersion of the image average without overscan calibration is 1.164 ADU and with it is 0.376 ADU. The overscan calibration uses the darkest of the three areas of overscan pixels with overscan set to the maximum (200). Now that the initial problem is solved, using the overscan calibration probably has little effect on the image quality. Getting the DC offset correct (so that the sky background is measured correctly) will have some effect on Flat calibration.