deviation in measurement value from old LCR meter

[lieven]

I have developed an automated measurement system on PXI platform using a DMM (PXI4072) and a switch (PXI 2530). I have developed the program in labview based on the tutorial and some example codes I have found. Nowadays we are testing this system, but we find that for some measurements the inductance readings deviate about 20% compared to our older system which uses a Wayne Kerr LCR meter. Can anybody explain where this difference comes from and/or how it can be compensated?

Thanks a lot!

Best regards,

Lieven

[lieven]

Might this have something to do with the following:

_Wayne Kerr LCR-measurement device has the possibilty to measure Q-factor, Z, ... at a specified frequency(eg. 1kHz) where NI DMM can only measure L, C, R.

_The coils I want to measure are far from ideal as they all have one or two air gaps, also ranges vary from 300µH to 750mH.

Regards,

Lieven

[lieven]

Hi,

As it seems that I get no respons on the forum or on the phone, I have looked further for a possible solution myself.

I found a discussion on the forum from someone who seemed also to doubt about the way of measuring inductances by the DMM:

http://forums.ni.com/ni/board/message?board.id=80&thread.id=1057

Only the frequency range seems to be the reason here not to go for the multitone measurement, however it is not that the measurements deviate only at small inductance values so there must be another reason... WHICH?

Regards,

Lieven

[BrowningG]

Hello Lieven,

From your previous posts, it looks like when you perform an LCR measurement on your PXI-4072, it deviates from measurements made on your Wayne Kerr device.  What do you mean by deviates?  Is the inductance measurement on your PXI-4072 constant?  Or does it fluctuate from one measurement to another? 

Most LCR meters use a single frequency tone (that is user specified) to determine the capacitance or inductance value of the device under test.  The PXI-4072 uses a multi-tone square wave to determine the capacitance or inductance value.  Also, many of the materials used to manufacture capacitors and inductors have different characteristics that depend on frequency.  In order to ensure that the results returned from one LCR meter to another are comparable, configure the non-NI LCR meter to use the effective test signal specified in the 4072 specifications.  The range used in the measurement changes the test frequency and current value, which can result in changes in the measurement result. 

The specifications for the 4072 specify several constraints on the measurement in order to get the specified accuracy.  Please make sure that these constraints are followed to ensure more accurate results.  Hope this helps. 

[lieven]

By deviation I mean that the L-measurements with NI DMM deviate up to 20% compared to the Wayne kerr measurements.

Altough this deviation of 20% applies only on a few models of coils, but a deviation of about 6-13% is very common when comparing measurement results of both systems.

Also sometimes there is no difference in measurement results of both systems.

The spread on measurement results is approximately the same for both systems.

Something that I do not understand is that two different models of coils that both have the same inductance value on the Wayne Kerr system and also on another LCR-meter(some other measurement system from a third company that we used long time ago), but on the NI DMM the measurements these measurement result do not equal the old values and even more, they are not measured as the same value anymore.

Some examples of the differences I have measured:

                      Wayne Kerr           Older system                   NI DMM

coil type 1            20,1mH                   19,9mH                     19,5mH

coil type 2            19,88mH                 19,9mH                      25,6mH 

coil type 3             341,4mH                 338,5mH                   433,6mH

So what I would really like to know is what causes these differences in measurement results. Certainly, what causes the difference in measured values on the new system, where on the old systems the values are the same.

Regards,

Lieven