Off with his head, I say! (the undergrad's, that is) :-)
From time to time I've been saying that any A/D board would benefit from having a tightly-coupled D/A run from the same clock, so that when things like this come up one could just hook up a 'scope to the D/A's output and quickly resolve, in real time, most questions about whether or not the A/D was OK. Dana Whitlow Arecibo Observatory On 2/15/2013 7:01 PM, [email protected] wrote:
I have been having alarming trouble with our 83000 ADC board that corresponds to the adc83000x2 block in the CASPER library. The ADC board seems to have stopped working. The ADC board may have been damaged and I need it to work before I defend my thesis. Before ordering another ADC board, I want to make sure that my configuration is not the real problem. The first time the problem was seen was about 18 months ago, but that time I discovered the ADC board was not seated properly on the ROACH board. Three tries of re-seating it solved the problem. About a year ago, undergraduate student began using my ADC board, and sometimes they remove the DC block from the inputs to the ADC. The 83000 ADC can be damaged by DC offsets. Attached a test configuration in MDL format and as a PNG picture, and a text file showing a "hexdump -C -v" of the contents of the snaps. I input a crude sine wave to the ADC. Do you agree that the configuration should show raw ADC values corresponding to a sine wave instead of what is in the text file? If that is a valid test configuration then my board may have been fried and that will cost.
