PXI-5122 - multi-device multi acq. average issue - NaNs on channel 0

[Jan_Hettler]

Hi,

I've encountered the following problem. I'm working on the multidevice ((PXI-5412) + 3x PXI-5122) multi-channel application and I would like to add multi acq. average as a processing step with Add Waveform Processing.vi. 

So far, I've been able to write a  programme that does the specified thing, but I've encountered a strange problem. When I select only the 0 channels to be digitized, I receive an array full of NaN instead of the signals. No error or warning thrown. When I select the channel 0 and 1 to be digitized, I receive only the averaged waveforms for channel 1 on all devices. Channel 0 is again full of NaNs.

Funny thing though is, that when I put the Fetch Scalar Measurement.vi in front of the Fetch Array Measurement.vi, everything works just fine (even with None options selected). I get the averaged waveforms for both channels on all devices. The programme also works, if I select None as a processing step in Add Waveform Processing.vi and then just fetch the multi acq. average using Fetch Array Measurement.vi. Nonetheless, both of these are just workarounds not solutions. I've tested the the multi acq. average on each digitizer separately (see the second .vi) and it works just fine. So I think that the problem has probably something to do with multidevice setup.

Any ideas?

Kind regards,

Jan Hettler

[Zody]

Hi Jan,

Did you manage to solve this problem?

It is pretty difficult to replicate the system due to high number of oscilloscopes.

Can you confirm that you have the same behavior with one waveform generation and two oscilloscopes?

Best regards,

I.R.

[Nathan-P]

Jan_Hettler,

Which version of the NI-Scope driver are you using and on what OS?

[Nathan-P]

I was able to reproduce the behavior you were seeing with a USB-5133 digitizer.  I didn't need multiple devices, just one.  I was able to reproduce it by removing the Scalar Fetch from the niScope shipping exampled called "niScope EX Advanced Measurement Library.vi".  Right now this is a known issue and R&D has been notified of this bug (CAR#219351).  You can track it with each release of the NI-Scope driver, a list of fixed CAR's are included.

I noticed that in the example VI's that ship with the NI-Scope driver, the Fetch Array measurements is always done after a Fetch Scalar measurement, and this works like you mentioned.  This may have been why it was never caught, since preprocessing is designed to affect the scalar measurements, by providing filtering.  It may actually be better to not setup "Multi Acq. Average" as a preprocessing step since it seems to work more consistently when the option is included on the Fetch Array Measurements VI.

[Jan_Hettler]

Sorry for  a delyed answer. I was not in the lab today. To sum the things up:

I was also able to reproduce this error with only one digitizer used. I've even tested it in different chassis (PXI and PXIe) and it's always the same. Channel one produces just an array full of NaN's and the error exclusively affects 0 channel(s). 

Generally it's not that serious issue, but it's annoying. Especially if you want to apply more signal processing steps using the AddProcessingStep.vi prior to fetching the measured using FetchArrayMeasurement.vi. 

Kind regards,

Jan Hettler

[Jan_Hettler]

I am running 32-bit installation of LabVIEW 2011 with NI-Scope 3.8.5 on Win7 64-bit and 32-bit LabVIEW with NI-SCOPE 3.8.7 on Win XP 32-bit.

[Nathan-P]

Right, from the CAR I mentioned above, the following settings will not behave properly (array of NaN) in a preprocessing step:

1.) addChannels
2.) SubtractChannels
3.) MultiplyChannels
4.) Dividechannels
5.) MultipleAcquisitionAverage

Currently the workaround you found is the reccomended solution.

[Jan_Hettler]

OK, thanks a lot for cooperation.

Just one last remark. Somebody should consider renaming the Voltage Amplitude option in Scalar Measurements, because it doesn't actually return the classical voltage amplitude (Vpp/2) 🙂

It's a bit confusing then - http://forums.ni.com/t5/High-Speed-Digitizers/5112-Vamplitude-vs-Vp-p/td-p/504724

Kind regards,

Jan Hettler