"When we configure the FPAA chip to be a straight wire from input to output, we notice that we get a reduced amplitude signal as the output. i.e. an input sine wave of 3V is output as ~1V.
We've set the board up as single-ended, but we tried changing some of the configuration switches and could get this to change. Why does the amplitude drop like this? Should it not pass the signal through unchanged? "
Depending on which version of the PAM-5000 you have, there are slightly different gains you should expect through the PAM unit. The version information is the first letter in the serial number, which is or should be on the RS232 connector. Versions are in alphabetic order, a Version L board is an older design than a Version M board. (Version Ra is the latest version, see the PAM Support page for a detailed explanation.)
First, using a single ended input automatically reduces the gain by half. That should give you 0.5V at your output instead of 1V. This is not limited to the PAM units, anytime you go from single ended to differential, unless great care is taken, you loose half the signal.
The second loss of signal is known to be the output buffer which converts the differential signal to single ended. This had a gain of approximately 0.9. (Something like 0.8695) This was corrected to give an overall gain of 1 but I have forgotten when (which version) that change was made on. It is at least correct on the R version, perhaps as early as the P version.
There also could be loading effects, depending on the signal generator you are using.
Other than telling you how and what to remove and resolder, there is an easier way to do this in the Anadigm:
Because of component tolerances, this gain might be slightly different for each of the outputs. Hopefully that will take care of your problem.