Amp glow is heat (IR) based. All it means is that the electronics of the sensor, typically the amplifier section that reads the sensor, is putting off some heat, and even if it is just a little, and even though the sensor is cooled, it will cause some areas of the sensor to be warmer than other areas and those warmer areas will have more dark current than the other areas and appear brighter. It is like taking a thermal image of an object and the temp of the object not being uniform all over.
It is unavoidable. CCDs have it as well, but to a lesser extent, at least with CCDs designed for long exposure imaging. I think CCDs are able to contain it better because all of the readout occurs on the edge/corner and that is where you find the glow. I look at it as being just part of the dark current. In fact, you measure it by taking a dark frame and subtracting a bias frame, leaving just the dark current. It is only a problem when it is either excessive or inconsistent. If it is excessive, then it carries with it additional noise and if it is inconsistent, then it can't be completely calibrated out with a dark frame.
Personally, all of this glow talk sounds much worse than it is. The sensor isn't perfect, and none are, but it is the size of an 8300 and it is keeping up with an 8300. I appreciate the immense amount of work that Jon has done in measuring the sensor. It will help everyone who uses the camera. But the more pressing issue is getting the drivers to work consistently, and progress seems to be being made on that. Or the users are exhausted and just not complaining as much.
Edited by syscore, 18 June 2016 - 07:44 AM.