Due to a confusion on my part, my report on gain variations in ea20 was partially incorrect. The correct statement is that those ~hour-long 10% gain variations are only seen at C-band. All looks good at X band and Ku band.