Acquiring AC voltage signal using NI 9206 9205 and cRIO

[phdieum]

Hello.  I am having difficulty accurately acquiring an AC voltage signal using a NI 9206 and cRIO.  I'm attempting to acquire voltage signals from two types of Magnelab Current Transformers: split core and rope.  In Labview, I first populate an array of size 2,500 with the sensor signal (DIFF mode), then compute the RMS of the array.  For the split core CTs, I am able to accurately acquire the correct voltage reading (verified by measuring the corresponding amp on the line using a Fluke 434 Power Analyzer.  For the rope CT, however, using the same array/RMS method, I am unable to accurately acquire the correct voltage reading.  Measuring the amp on the line using the Fluke 434 PA, the correct voltage from the rope CT should be 0.05v.  Using the 9206 (DIFF mode), the RMS from the array gives a voltage reading of 0.071.  Now the interesting part is when I measure the line voltage using two different True RMS DMMs, I get two different readings.  One multi-meter, a Klein CL2000, reads the correct voltage at 0.05v.  Another multi-meter, a Fluke 189, reads incorrectly at 0.071, the same as I get using LabView and the NI 9206.  I suppose the question is how is the Klein interpreting the signal differently than the Fluke 189 and the NI 9206 via LabView.  One difference between the split-core and rope CTs is that the rope CTs require an external power supply 12-30v AC or DC.  I am providing them with 12v DC.  I have tried multiple voltages in both AC and DC and still get the same incorrect reading.  I'm quite certain it is not a power supply issue, although perhaps the integrator in the rope is creating a unique signal.  Any ideas?  I appreciate any input.

Thanks,

J.Grant

[AnalogKid2DigitalMan]

Is the AC waveform a sine wave? What frequency range is it? What are the model numbers of the Magnelab transformers?

Never heard of a 'rope' transducer, how do they work?

-AK2DM

[Jake_G]

Hey phdieum,

It's odd that one of the DMM's and LabVIEW should agree, while the other DMM does not.

The RMS of a sine waveform should be a/sqrt(2), and the value that it should be is 0.5 while the incorrect number you are getting is 0.71.

I wonder if the Fluke 189 and LabVIEW are giving you a and not the RMS (i.e. 0.71/sqrt(2) = 0.5).

Does this work for other values?

[phdieum]

Hey AK2DM, 

The AC waveform for the split-core Magnelab SCT-1250 is basically sinusoidal.  The AC waveform for the rope Magnelab RCT-1800 is not as sinusoidal.  Take a look at the attached pic.  The top graph is the rope signal, the bottom is the split-core signal.  In this pic the two signals are on different circuits, but the top graph is the rope signal and should read 0.05v and not 0.078v.  Strange that the Klein DMM reads the signal correctly at 0.05v while the Fluke DMM and LabView with 9206 reads the 0.078v.

[phdieum]

Hey Jake, Very s

[phdieum]

Hey Jake, very strange indeed.  In LabView I populate an array of size 2,500 then take the RMS of this array.  I'm assuming it's dividing by the sqrt (2).  When I measure the signal off the spit-core CT, the Klein DMM, Fluke DMM, and LabView with 9206 all read the correct signal.   It's only on the rope CTs, that I'm getting different readings.  Take a look at the attached photo of the signal.  The top graph is the rope and the bottom the split-core.  These are on different circuits, but you can see the split-core signal is sinusoidal while the top graph the rope signal is not so sinusoidal.  Still strange that the Klein gets the correct reading on that not so sinusoidal signal. 

[AnalogKid2DigitalMan]

Ahh- now I see what a rope sensor is!

Is it possible to increase your sampling rate? Those plots look pretty rough.

Are you using an integrator on your rope?

One thing I can think of is maybe you have high frequency harmonics on your line that the Fluke and 9206 are picking up. The Fluke has a bandwidth up to 100KHz, while the Klein is limited to 400Hz.

What type of equipment/machines are on the bus you are monitoring current on?

-AK2DM

[phdieum]

Hey AK2DM,

I increased the sampling rate to 1us with no change.  I should have mentioned the plots were only a few samples so the wave could be seen.  The ropes do have built in integrators that are powered with 12v DC. 

I think I found a solution.  By applying avButterworth low-pass filter prior to the getting the RMS, I am able to get the correct voltage signal.  The high-cut off setting for the filter has no effect on the signal, but the low-cut does.  I can essentially tune the low-cut off so that my voltage signal is correct.  Does this make any sense?  Also on different phases of power supply, would the noise be different because it seems each rope CT needs i's own custom tuned low-cut off setting.  I guess this is to be expected on unique phases of power?  Thanks for your help.

-phdieum    

[phdieum · Accepted solution]

AK2DM:

Update - solution found

While the Butterworth low-pass filter sort of worked, it was NOT the solution.  I spent some time analyzing the rope CT signals by sampling at a very high rate.  The AC signal coming off the rope CTs is triangular NOT sinusoidal.  The RMS function in LabView assumes a sinusoidal waveform and essentially calculates the RMS as the average of voltage peaks (for each cycle) / sq.rt (2).  Since the AC voltage single emanating from a regular split-core CT is basically sinusoidal, the LabView RMS of the signal gives an accurate reading of the voltage signal.  Unable to find a RMS function in LabView for a triangular AC voltage signal, I ended up developing my own custom triangular RMS function in LabView.  My function analyzes the signal by finding the peak of every cycle within the signal, takes the average of these peaks, then divides this average peak by sq.rt (3).  see this link for formulas:

http://www.optics.arizona.edu/nofziger/OPTI%20380B/Lab%201/AVG%20and%20RMS%20Values%20of%20Periodic%...

Using the triangular method, I am able to accurately acquire the signal off the rope CTs. 

Thanks for your help!

[AnalogKid2DigitalMan]

phdieum:

Good work! I would not had expected a triangular waveform, but never have used a rope and an integrator.

Glad you discovered the source and came up with a solution.

-AK2DM