Hi Jack,
Some of your wording confused me a bit. When you were mentioning that you were using a 'calibrator' on the inputs. Do you mean that the PCI-6143 is the calibrator or is some other device doing the calibration?
One thing you mentioned is that the level seems to be high for 10 samples and low for 5 samples. So you are saying that the time of the cycle is independent of the sample speed?
If you are concerned with a problem due to impedence matching you may want to take a look at this Tutorial:
Impedance and Impedance Matching.
One last thing to try is to take the measurements with different channels. Does the behavior remain the same? Is it the same for all of the channels that you are reading data?
Regards,