Replacement for NI PXI-4071 DMM

[JB]

Are there any recommandations for a PXI DMM for low resistance measurements (typically 90mΩ) ?

I'm currently using a NI PXI-4071 but its minimal range of 100Ω is quite high.

Moreover, I'm looking for a faster DMM (PXI-4071 = 1145ms for a 71/2 digit 4-wire resistance measurement).

Many thanks in advance

[Steve_B]

Hi JB, The PXI-4071 is the best DMM that NI has in our portfolio for making a 4 wire resistance measurement. The accuracy of the low resistance measurements are going to be limited by the onboard current source of the device. If you were to pair the 4071 DMM with an accurate, higher current supply you could implement your own 4 wire solution that would have better accuracy and a faster measurement because you wouldn't necessarily need 7.5 digits from the DMM to measure 90mOhms. A higher current source makes the amplitude of the voltage higher for the DMM. You would want to be careful not to introduce self heating of your DUT and can control that by selecting an appropriate level for your current source. 

The combination of uncertainties for the voltage and current would follow this example:

http://en.wikipedia.org/wiki/Propagation_of_uncertainty#Example_application:_Resistance_measurement

I would also recommend implementing your own OCO by taking the difference between 2 voltage measurements where the polarity of your current source is reversed in the 2nd measurement.

Hopefully this is helpful.

[JB]

Hi Steve,

> The PXI-4071 is the best DMM that NI has in our portfolio for making a 4 wire resistance measurement.

This is exactly why I choose it some months ago... and because I didn't found another one in PXI format with a lower range.

The DMM is used in an automated production line and is currently shared by two DUTs (through a matrix). The 7.5 digits resolution is indispensable to get accurate values. With this configuration, the duration of the measurement of the 3 DUT's resistances is unfortunately too long to respect the cycle time. (1145ms/measurement).

The PXI chassis was choosen with this probability in mind and the adding of a second PXI-4071 would be a quite easy solution. But this also restarts the search for a faster third party PXI DMM with a lower range for R measurements.

I guess you mean this setup. Do you have any experience with it ? Are 4 wire measurements possible ?

With 6.5 digits, the DMM needs 290ms for one measurement. Therefore this could be an interesting alternative to above mentioned solutions.

Best regards

[Steve_B]

Hi JB I have used the PXI-4130 with the PXI-4071 to make some very low resistance measurements using the current source at 200mA and the 4071 measuring at 7.5 digits on the 100mV range. I believe I was able to achieve an accuracy of about +/-50uOhms based on the calibration dates of my devices. I was not very concerned about speed for this test. I would encourage you to look at your error budget and run some calculations to see if a configuration would be sufficient for you and if the aperture times, etc are somthing you could work with. 

More recently I was taking some measurements of contact resistance for some relays to watch them age and I used a PXIe-4141 for 4 channels of simultaneous V & I measurements. My accuracy requirements were pretty loose so I was running very fast to reach close to a million cycles without taking too much time. Here is a plot of some of my data with Ohms vs Cycles as the axes. 

I was using a 100uA aperture time, a 1ms source delay, and +/-50mA current limit to bring the voltage over 1mV. I did not run an error calculation for this test because I was mainly concerned with the trends as opposed to the absolute resistance, but slowing down the measurements would help!

[JB]

Hi Steve,

Because I'm using contact probes to connect the DUT, this is a big concern : are 4 wire measurements possible with this setup ? I fear no. So this won't be a solution because of the varying resistance of the contacts.

Many thanks for your help

[Steve_B]

Hi JB, the setup with the 4130 and the 4071 is a 4 wire measurement inherently. The setup I was referring to with the PXIe-4141 was using a 4 wire measurement as well, it is referred to as remote sense with the NI DC Power driver. 

The 4141 voltage measurement accuracy is not comparable to the 4071 DMM so it may not work for you depending on your erro budget, but you would get 4 channels of parallel test that would not require a switch. If the test current can be high enough, again there is concern for heating up the DUT, the voltage can be large enough for a reliable measurement. The entire test I was doing with this setup was only measuring the contact resistance of some electromechanical relays that we were evaluating, they are not on our board. 

Regarding the contact probes, how are you making the connections now with the 4071? Any 4 wire setup should be able to make use of the connection you have now, regardless of using 1 or 2 instruments to make the measurement. Ideally you would want to have 2 connection at each side of the DUT so the contact resistance of your connectors is not measured.

[JB]

>The setup with the 4130 and the 4071 is a 4 wire measurement inherently

Yes, obviously ! Sorry for the stupid question.
 

The DUT is a 3-phase motor and R of the windings is only one type of the measurements.

The hardware consist of a PXI-1033 chassis with a PXI-4071 DMM and a third party matrix. In addition, there are two DAQ cards, one for each DUT. For several reasons, these cards are not in the PXI chassis but in the computer.

  >Regarding the contact probes, how are you making the connections now with the 4071? Any 4 wire setup should be able to make use of the connection you have now, regardless of using 1 or 2 instruments to make the measurement. Ideally you would want to have 2 connection at each side of the DUT so the contact resistance of your connectors is not measured.

Contact probes for 4-wire measurements (two contacts on one probe) are used to connect the DUT.

As mentionned before, the results with the DMM are satisfaying and the main point is the reduction of the test time. This can easily be achieved by adding a second DMM.

The solution with a power source has too many uncertainties (fast enough to avoid second DMM, accuracy...?) to implement it in this already running application. However it seems very intersting and I will examine it more in details if a future project requires low resistance measurements.

Many thanks for your help