Short circuit between analog input and output?

[Doctor_G]

Hello everybody,

I'm trying to set up a PCIe-6361 to supply a voltage and measure the voltage drop over two resistors. The device actually needs two voltage inputs with a common ground, but for now, I'm just setting it up with one to make sure it works. Attached is a simplified schemicatic as well as the breadboard for the DAQ connections. 

What I'm getting is that, even without a device, there is somehow a drop in voltage across the upper resistor. Why is this happening? Measuring with an external voltmeter, the drop only happens when I have attached the AI0 leads across the resistor, so it isn't anything funky with my breadboard. I have the LabVIEW program configured for a differential measurement, and the supply voltage is only in the range of 0.2-1V so I don't think it's above the common mode voltage limit.

Any thoughts? I also notice that the noise in the AI voltage measurements increases significantly when I connect the sourse... if there's any way to reduce the noise, that would be really helpful too.

Resistances are 39.5kOhm (top) and 609 Ohm (bottom.) 

[Doctor_G]

Quick update: 

The problem I'm have seems to be related to the sampling rate. 

To recap, I have ao0 connected to ai0+l and one side of a resistor, with ai0- connected to the other side. On a separate area of the board, I have ao ground connected to ai1- and one side of a resistor, with ai1+ connected to the other side. There are no connections between the two resistors. 

When I have the sampling rate set to around 10k Hz or less, I don't get any voltage drop across either resistor (as I would expect.) When I have the sampling rate set to 1MHz, I see a large voltage drop across the upper resistor (connected to ao0 and monitored by ai1) and can confirm with an external ammeter that around 5mA of current is flowing through ai0+ (pin 66). 

I tried backing down on the sampling rate, but the issue only seems to really disappear around 10kHz, but the noise level is still very high.

Any thoughts? I need to do frequency analysis with this setup and would really like a sampling rate higher than 10kHz!