It's been quite some time since there has been activity on this thread, as Brandon and I took the discussion off the board.
I identified that my program was not properly waiting 500 ms before taking a measurement using the SMU. When I "fixed" that, I no longer needed to extend the aperture time or use the higher order noise filtering property. I did run into a problem this week for which I found a fix, but would like to see if there are good explanations:
I configured the channels of my PXI-4141 SMU to measure currents between the HI S and LO S terminals (one channel at at time, with the force terminals not connected to anything and in remote sense mode), and had a test fixture in place in which I needed to measure 100 uA (10 V applied across a 100 kOhm resistor. Except I was seeing only about 80 uA. I set up my test system to step the voltage between 0V and 10V and observed that the current measurements would roll off after 60-70 uA. I also observed that this behavior would step up to about 80 uA when the wires were hooked up to my DUT in one direction, with an entirely different behavior in the other. It would jump from -60 uA to approximately "-120 uA" when the applied voltage changed from 6V to 7V. I also observed that the SMU Voltage was reading a non-zero value, and corresponding to the IR drop across my DUT resistor not represented by the current reading (about 2V in the case of 10V applied across the 100k resistor and a "reading" of 80 uA)
My mind went back to the fact that the LO terminals of the four channels are all connected and observations I made that the force and sense portions of the SMU are not entirely decoupled, as I once saw the SMU applying current through a DUT through the HI S terminals. I shorted the LO and LO S terminals together, and voila, by current readings looked like what I expected, with essentially zero SMU Voltage measurements.
As I stated above, I am interested if others can provide an explanation for what is going on- I suspect that the partial coupling of the force/sense circuitry of the SMU was not allowing the instrument to maintain its 0V setpoint when the current was above ~60 uA. I chose to connect the LO side terminals because of the knowledge that the LO terminals of all four channels are linked already. I am also interested to learn if there are potential downsides of what I have done, either in potentially damaging the instrument or compromising the accuracy of any measurements that it may provide.