4070 open-circuit resistance measurement

[jjsomer]

I am using a PXI-4070 card to make a 4-wire resistance measurement to determine ground continuity. However, with no connections on the front of the card, the resistance measurement (using NI provided example - DMM single resistance measurement) is -90 ohms. I expected a return of NAN. What gives?

[T_Marsh]

Hello,

The problem with measuring resistance in 4-wire mode with nothing connected is that there is a disconnect between the current drive voltage and the sense voltage.  If you were actually measuring resistance and had everything connected correctly the leads of the +Iex and the +Sense lines would be connected as well as the –Iex and the –Sense lines.  Even if there were no path between the + and – those sense lines would provide feedback about the current output voltage.  Because of this the DMM could tell that the current drive was having to go to the max voltage limit to be able to drive the appropriate current, and would therefore give an overlimit voltage.  

When I did testing over here using the Soft Front Panel I found that when I was using ranges anywhere to 100 Ohm to 100 KOhm I would get overlimit for no connections as expected.  If I used the 1 MOhm range or above I would get very odd readings that would drift about if I had nothing connected to any connections.  However, if I had the wires connected as listed above I would get over limit as expected.  This is probably because on the 407x when testing resistance above 1 MOhm uses a Max Test Voltage of 10.5V.  Ranges 100K and below use 1.05V or lower as the Max Test Voltage.  The voltage on the feedback may have drifted above 1.05V but not above the 10.5V.  This would make sense if the AI on the feedback had a pull up resistor to 5V.  Once the feedback drifted above this Max Test Voltage then the DMM would return over limit.  

The best solution for you would be to make sure you actually have a circuit wired up to measure resistance in 4-wire mode.  As long as there is a path between + and + and also – and – then you should be OK.  If not, then consider measuring in 2-wire mode first to determine continuity and then switch to 4-wire mode if need be.

Regards,
Travis M

[J_Bray]

Are you using Offset Compensated Ohms for this 4-wire measurement?

Message Edited by J Bray on 10-03-2006 05:41 PM

[jjsomer]

Connecting the 4070 to my circuit and creating an open circuit load, yielded the same results.  I did use the Soft Front Panel and saw the same results that you were seeing - basically, an over limit situation up to the 1MOhm limit.  However, using the DMM example for single resistance measurement, which emulates exactly how my application is performing the 4-wire resistance measurement, an over limit situation is never detected.  Instead, the DMMRead function resulted in a resistance in the range of -50 to -120 ohms.  Maybe there is a difference in how the Soft Front Panel application and the DMM example application make those measurements.

As far as the offset compensated ohms, I have tested my software and the NI example both ways.  Using offset compensation, the DMMRead function returned a value very close to zero, i.e. 0.00000006 ohms.  My software has been using offset compensation in the past.

It does seem like when I make a 2-wire measurement, an over limit result always occurs.  Therefore, I am making a 2-wire measurement first to determine an over limit situation.  Then following up with the 4-wire measurement.  This is not an ideal software solution, so if you have any other ideas, I would like to know.

[T_Marsh]

Jim,

I was not using the Offset Compensated Ohms when I did the measurement.  If I try it with OCO enabled then I get readings right around 0.  This makes sense because OCO takes one measurement with the current flowing, one without and then subtracts the two.  Since in an open circuit you will have no current flowing you should get about the same reading each time, resulting in 0 Ohms net.

jjsomer,

I tried out the standard shipping example for both LabVIEW and a C based example in CVI.  In both of the examples I was getting the same results as I got from the Soft Front Panel.  The only time I could get odd readings, some of which were negative and of varying values, was when the leads were all independent as if nothing was plugged in.  

The Soft Front Panel and the example program should be making the same basic calls into the driver and should return basically the same results.  The Soft Front Panel only calls a few additional configuration steps because there are more options than in the basic shipping example.  

Is it possible that there is a short in the circuit between the either the + excitation and the + sense or the – excitation and the – sense?  One way to test would be to jumper directly from red to red and from black to black on the DMM itself.  By doing this you have basically changed a 4-wire measurement into a 2 wire measurement and you should get exactly the same results, which in this case should be over limit.

The 2-wire measurement is also a good pre-cursor test as you mentioned.