Pickup Noise (Horizontal Banding) On The SCAM Detector

There is a known problem with the SCAM detector picking up 60 Hz electrical noise. Because of the readout frequency of the SCAM is almost an exact multiple of 60 Hz, this problem shows up as bands that appear horizontal or nearly horizontal on the detector. These bands are NOT sharp, row to row. A vertical plot across them looks like a short-period sine wave, not like a square wave. This appearance strongly suggests that the noise is analog in nature, in that it affects the signal from the detector prior to digitization.

The noise pickup is unfortunately quite time-variable. It often is completely invisible, especially in single frames. It most often shows up in subtracted frames, and then most commonly when the subtracted frames have nearly the same flux level. Typical values for the noise, when it is seen, are on the order of 10-20 DN peak-to-valley.

Here is a sample image showing the banding:

This image was taken as a guider "cam" frame, when SCAM guiding was in use. It was taken shortly before dawn, with thin variable cirrus present. The large offset between the quadrants is probably due to the bright and variable sky illumination. In SCAM guiding, a "sky" frame is stored and subtracted, and the "sky" frame in this case may have been saturated or nearly so. Because the quadrants have slightly different gains, they saturate at slightly different times, and so one quadrant of the sky frame may have been fully saturated while another was not. (Caution: this is all after-the-fact rationalization, but this quadrant offset is commonly seen in subtracted images with high but different flux levels.)