ASI294MC - horizontal banding in bias

jean4zwo

Hello,

I just bougt the ASI294MC and the first thing I did was to take some bias frames. I downloaded the frames via USB3 port on to my windows 10 PC with SGPro. Gain was 120, temp -5°C. When opening the frames in Pixinsight and applying the STF (screen Transfer Function-stretch) Some brought horizontal banding becomes visible on all of the bias frames but not always on the same place in the image. So the sensor or cooling can't be blamed. To take a closer look I made 30 Bias frames and integrated them.

After integration I cropped from the middle a 2048 x 2048 square. From one of the bias frames I did the same crop. Then I subtracted from the cropped bias frame the cropped integrated bias frame and did an FFT (fast fourier transform) using the FFT module from PI. The result shows me an image with a vertical bright line in the middle. If the pixel distribution of the bias frame would have been Gaussian I would have only a bright point in the middle of the image. The vertical line seems to tell that there is a high frequency component in the downloaded bias frames. Maybe a badly shielded clock frequency???

The method used is described by QSI in their article: Interpreting Fourier transforms: link: http://www.qsimaging.com/ccd_noise_interpret_ffts.html

Can you tell me what your interpretation is of this phenomenon?

Attached the FFT and one of the bias frames.

Kind regards

Jean

Sam@ZWO

sorry to say there is such noise with 294

almost all CMOS cannot have very clean bias frame as CCD

but this is not very serious problem

because these noise is very small and can be averaged by stacking

or covered by sky glow

james1ca

Yes, CMOS sensors have pattern noise which will show up in each frame (bias, dark field, and lights). In fact, it might be better to stop using bias files and just calibrate with darks. The only good reason to use bias files on a CMOS sensor is if your processing software tries to perform dark field optimizations or scaling to correct for small differences in the camera's temperature or for small drifts in the camera's response. However, you shouldn't use dark field optimizations or scaling if your camera has a significant amount of amp glow and the ASI294 definitely has amp glow. Thus, you will probably get better calibrations if you stop using bias files and just calibrate with a well-matched set of darks.

In any case, even with calibration you will still have some residual pattern noise, since some percentage of that noise is random in nature. For that reason, some recommend dithering during capture (particularly if you want to use dark libraries) while others can get by with retaking the darks during each imaging session (basically trying to get the closest match between the lights and darks).