Over range and NaN on 4071 DMM cards

[Wortel]

Hi,

I am trying to capture an AC voltage using the PXI-4071 DMM cards and have encountered a problem with the 'range' setting.

The voltage I am trying to capture is formed from a short circuit test, so it has a large spike when the switch is closed and then a small sinusoidal voltage that gradually decays to a negligable value.

The problem I am having is that once the switch has been closed and the spike has occured, the DMM card just stops recording any data, up to the spike the data is recorded as required. If I look at the array generated by the program it is just full of 'NaN' entries after the switch was closed, this made me think that it was a problem with the range setting. However I have adjusted this to a higher setting and it still gives the same results, even though the spike is much lower than the range.

The maximum value of the spike (measured on a different system) was 32 volts and I had the range value set to 50V, so this should have easily been high enough.

Also I would have expected just the values over the range limit to be NaN and then once the signal came back below the limit for the DMM to carry on recording.

Any suggestions would be greatly appreciated.

Thanks in advance,

Darren.

[kirsch]

Hello Darren,
That does sound odd. Go ahead and try this with the largest range setting, just to see if you still get the NaN or not. If you don't, let me know what the maximum voltage reading is.
What version of NI-DMM are you using?

Thanks!
Marcos

[Wortel]

Marcos,

Thank you for your reply.

I tried the tests at the maximum range setting (700) and I still experiance the same problem, the card just stops collecting the data after the spike. For very small values (around 5 - 10V) of the spike it is ok, but as soon as it gets higher the only data recorded is before the switch is closed, which causes the spike. The resolution of the data recorded is however very poor, we are only looking at a mV signal (after the spike) so using such a high range is not ideal.

We have now made a hardware change so there is a diode bridge inbetween the switch and the 4071 DMM which will limit the voltage to 5V. the actual spike is not needed for the final calculations and is removed anyway, so it is not essential.

Thanks again for you help.

Regards,

Darren.

[ninlmartijn]

Hello Darren,

This is Martijn from NI Netherlands. We can't find a explanation of the strange problem you encountered. I know that the problem is solved for this moment. If you want to look further into this problem, I have a few suggestions to look further to determine the problem.

- Is it possible for you to try this with another NI DMM?

- Can you try this in the softfront pannel

If there isn't a problem in the softfront pannel, will it be possible to zip your code and post it here?

Best regards,

Martijn Smeulers

NI Netherlands

[wildcat_600]

Hello Darren,

Another possible explanation may be that the spike exceeded the maximum Volts-Hertz product of the board.  Normally when this happens an error is thrown that is described in this Knoledgebase article.  Depending on your program the error that is usually thrown may have been ignored.  If you try Martijn's suggestion and try using the SFP you may see this error.  But I think limiting the voltage spike is a good idea. 

Have a good day,

Brian P. - DMM Product Support Engineer

[Wortel]

Martijn, Brian,

Thankyou both for your continued support in solving this matter.

This test is carried out on a 3 phase system and the spike could occure on any of the 3 phases, so in effect it has been tested on 3 different DMM boards already, each time giving the same NaN values.

At a sampling rate of 10k samples per second the spike only occurs for 2 samples, in fact on the old system sampling at 8333 samples per second it doesn't even record the spike on every test. With this in mind I'm not sure the SFP would be much use in testing this.

I think that the error described in the knowledgebase could be the one I am experiencing, I will calculate the V-Hz value based on the theoratical values and see.

We are going to try the test again on a new vacuum switch we have purchased. The vacuum should ensure that no such spike occurs so hopefully that will solve the problem.

Thanks again for your assistance,

Darren.