I had hooked up the a channel to monitor the tank temperature. However, now I am seeing random noise and also zero readouts on every other sample.  

I had something similar over the last week when I was trying to set up autolocker and PID slow controls for the laser. The fast monitor signal reported by epics was very noise. There I thought it was just poor performance of the FSS locking loops and that I would need to get in and check alignments, polarisations etc.  Now I'm thinking there is something wrong with the chain from the ADC to what caget spits out.  

A calibrated voltage input of +2 V was put into the C3:PSL-VACTEMP_MON1 channel (acromag channel regiser 7 (out of 0-7).  A quick grab of the ouput is attached below, it doesn't look great.  Might have to look tomorrow to see if this is something to do with modbus dropping measurments or if there is some other issue.