Best point to connect ground with single-ended measurements?

[davey31415]

I am supplying power to several potentiometers using a single DC power supply and am reading the varying pot voltages using a USB 6008 analog input.

Would it be better to connect the DAQ ground to the ground on the pot terminals rather than the ground terminal of the power supply?

If the power supply and daq device are both powered from the same AC outlet, is there a point in connecting the grounds at all?  Wouldn't this just add a ground loop into the system?

Thanks,

Dave

[Aaron_W.]

Dave,

We have a document that deals specifically with the question that you have. It is the Field Wiring and Noise Consideration for Analog Signals. If you look through the document it will give you various ways to connect different sources for different terminal configurations. The only issue with your USB-6008 is that it is not capable of doing a Single-Ended-Nonreferenced (NRSE) terminal configuration. Your best option for taking measurements is to use a differential measurement but you would be limited to measuring 4 total devices. If you need more channels than that your best option is to first ensure that all of your potentiometers are referenced to the same ground, and if they are you could try to place a 100K Ohm resistor (or similar value) between the ground of your potentiometers and the AI Ground of the USB-6008. This would be similar to using a AI Sense line, but I cannot guarantee how accurate these measurements will be with this type of configuration. 

Even with both devices connected to the same AC output you will get small differences in grounds for each device and that can lead to a ground loop if you connect the ground of your Potentiometers to the AI Ground directly. Hope this helps, let me know if you have any other related questions.

[davey31415]

Thanks Aaron.  

I placed a 100k ohm resistor between the ground for the pots and the AI ground on the USB-6008 as you suggested (I tried connecting to one ground and jumping all the grounds together).

When one potentiometer voltage changed the measured voltage on potentiometers which were not chaning also changed (I verified the voltage on these wasn't changing with a DMM). 

This did not happen with differential measurements.

I tried removing the 100k ohm resistor out of curiosity.  Without the 100k ohm resistor I am able to measure the potentiometer voltages correctly.  The measurements are not as precise as the differential measurements but they are accurate.

What problems may I run into when not using the 100k ohm resistor?

Regards,

Dave 

[Aaron_W.]

Dave,

With a grounded voltage source for your supply that you are measuring the only risk of creating a group loop is that you will have a noisy measurement, which is exactly what you are seeing. If this lack of precision in your measurement is fine for your application then I really don't see any harm in continuing taking the measurement that way. It would be best to use a Differential Signal, but I understand that it limits the number of available channels. Hope this provides some more insight on the situation.