Noise Problem

[Tom Droege <tdroege2@earthlink.net>]

In getting TOM2 and TOM3 running I had a noise problem.  Some of the TOM3 
images were messed up.  The usual cause of this is picking up a noise spike 
that causes an extra byte to be written into memory.  This cause all the 
data after that to be offset by a byte.  This is obviously useless, and can 
be easily observed in the image.

The problem turned out to be TOM2 turning the chiller on and off while TOM3 
was writing out images.  The cure was to move the telescopes onto a 
separate AC circuit from the chiller.  If any of you with Mark IVs have 
this kind of problem be sure to consult with me as I have a lot of 
experience with noise.

In general I try to build systems so that noisy operations never take place 
at the same time as precision operations.  This in general means to keep 
computers away from the analog circuitry because a computer that you 
thought you had nicely synchronized with the data operation may wake up and 
start collecting garbage just as you are making the accurate 
measurement.  This is why I designed the scanner with chips instead of 
doing it in the computer.  I know when it is making noise and can avoid it.

Tom Droege