Measuring floating voltage signals?

[Nathan Vaidyanathan]

I am trying to measure voltage drop across a resistor (0.5 ohm) using PCI 6221. The current thro the resistor will vary from 5mA to 450mA. I have configured the input channel as differential and the limits as +1V to -1V. I have 50K bias resistors connected to AI GND as recommended by NI.  When there is no voltage drop across the resistor, i  have no problem and i read close to 0V. When the voltage drop across the resistor is more than 50mv, i have no problem and i read the correct value. But When the voltage drops to 5-10mv, i always read a value close to 100mv. Could someone explain why there is a jump in the value i read? It looks like when the input is floating, the analog input is pulled to some value. How can i prevent this from happening?

[EEsRULEtheWORLD]

Hi Nathan,

If you suspect that your voltage is floating, you may want to implement bias resistors in your measurement.  Bias resistors will provide a DC path from the amplifier inputs to ground.  For more information on this check out this KnowledgeBase article.

In your post, you mentioned that you were connecting to AIGND.  It sounds like you are not actually connecting your signal differentially.  You may want to change your software configuration type from differential to RSE to see if this changes your reading.  Check out the following table for a visual description on the different configuration types and how to connect bias resistors for a given configuration:

If you are using a differential configuration and also using a floating voltage source (one that is not plugged into an outlet), pay attention to the way the bias resistors are implemented in the red circled selection above.

Let me know if this still does not solve the issue.

[Nathan Vaidyanathan]

Hi Sara,

I do have 50K bias resistors connected between AI+ to AI GND and AI- to AI GND as depicted in the figure. But i still have this problem. I did try both differential and RSE measurement. I got identical result. When the input voltage is around 3mv the value is read is 100mV. What is the minimum voltage the PCI 6221 can detect? Have anyone used it to detect signals in mV range?

[EEsRULEtheWORLD]

Hey Nathan,

I apologize for missing that information in your original post.

The PCI 6221 is capable of acquiring mV signals.  If you choose to use the +/-1 V range and utilize the full 16-bits of resolution on this card, you will be able to detect at least (1-(-1) V)/(2^16) = ~30 uV. 

Have you verified that voltage should be 3 mV with another oscilloscope?  If you have and found that the voltage read by the oscilloscope is indeed 3 mV, the next step is determining if this is a software issue.  How are you reading this voltage (custom program, example program, test panels, etc.)?  If you are using a program that you created, try running Test Panels to see if this same voltage is being read by your card.  If it is, then it's possible that the channel has become damaged.  If using another channel also gives the same results, you may need to get this card repaired.