VFD noise on other sensor output such as : Thermocouples, Accelerometer, Torque meter

[jftak01]

Hi,

I'm using a PCIe-6351 Daq board (and SCB-68) for my test rig. I'm collecting different kinds of sensor outputs through this DAQ board such as temperatures, acceleration, and torque meter output.

beside that, I have connected a VFD (Baldor H2 inventor) using a USB cable to the PC case. I use this VFD to control motor speed. (http://www.baldor.com/products/accontrols/vs_inverter.asp)

Here is my problem:

when I turn on the VFD I get a lot of noise on every signal. This noise is like a peak which repeats every 5-6 seconds.

So it's not 50-60 HZ noise of outlets.

You can see related uploaded video clips on :

http://www.megaupload.com/?d=4NHDIVV7
http://www.megaupload.com/?d=2461Z7KN

I have uploaded some pictures too.

Is there any way that I can filter this noise in Labview?  ( I have tried Filtration VI in labview, but couldn't find a correct setting) 

How can I get rid of this noise?

I really appriciate your help on this matter.

Thanks.

-Jeff

[Kyle_A.]

Hi Jeff,

Check the grounding for your devices/VFD.  Is there any potential ground loops?  Can you shield your signals? 

Do you know what kind of filter you are looking for?  There are filters avaible in Express form and low level form in the Signal Processing palette.  The only other thing about filtering is to create a filter to either set limits for how high a voltage measurement can go or do averaging on a subset of data so that the signals will not see the spiking. 

I feel the best solution will be to find and eliminate the source of the noise, whether it is grounding issues or EMI. 

[AnalogKid2DigitalMan]

VFD drives/motors are inherently "noisy" in terms of EMI and RFI due to the fast risetimes of the drive waveforms.

AC mains and motor phase wiring should ideally be contained in their own metallic conduits that are earth grounded. If that is not an option, then the next best alternative is shielded cables that have the shield connected to earth ground.

VFD drive should also be contained in it's own metallic enclosure that is grounded.

Filters between the motor and drive may also help.

All sensor wires should also use shielding if possible.

Been there, done that....

-AK2DM

[jftak01]

Thanks Kyle for your reply...

and sorry for late response..

I don't know how to check the grounding of devices??

Honestly I don't know what kind of filter works for me?? (have you seen the video clip which i uploaded?)

but i tried almost all of filters in express VI ,, but none of them worked for me..  😞

I really appreciate your help on this matter

-Jeff

[jftak01]

thanks AK2DM for your reply... and sorry for late response...

Actually, Most of the sensor cables or shielded... but I don't know about the USB cable which is connects VFD to PC.!!!

How can I check the grounding of other connections..

thanks a lot.

-jeff

[Kyle_A.]

Jeff,

You would need to implement a low pass filter that is about half the rate at which you are sampling.  This can either be done with hardware on your front end, pre-processing, or with the low pass filter VI's in LabVIEW, which is post processing.  One thing to make sure is that these noise spikes are not railing your measurement, and they still fall within max/min voltage range on the card.  That is to ensure no damage will be done to your hardware.  Either way, it sounds like something is going on with the VFD that is causing this issue, though I cannot determine if this is expected behavior or not.

[JordanJMcBain]

I can very much identify with this issue.  I have a test bench with two motors and corresponding Baldor H2 VFD's on it. I have a variety of sensors such as accelerometers, hall effect sensors, and a torque meter.  The bench was put together "professionally" but still get this kind of noise.

The *only* sensor that I have this trouble with is the torquemeter.  I get an impulsive EMI blast every 12 seconds on this signal with just one drive turned on.  With the other drive turned on, the noise is excessive.  I have spent weeks troubleshooting assumed ground-loop faults with the torque meter company's tech rep - to no avail.

I have confirmed that this EMI is from the drive.  I hooked up a wire to the signal lead of an oscilloscope's VFD (acting as an antenna) and the noise I get on this oscilloscope is very similar to what I get on my torquemeter. (The noise is present only when the drive is powered up)

The noise from the VFD will not be cured by using a low pass filter.  Impulsive type events have very broad band noise.

My new approach is to try and use adaptive noise cancellation to get rid of the the VFDs' EMI on the torquemeter.  I gather the noise signal on my DAQ card (PCI-7851-R) by plugging just one wire into the positive port of an AI port and tying a 1 mega Ohm resistor from that port's ground to the AI port; this provides a means of collecting the EMI from the drive with an "antenna." The wire dangles by the side of the VFD with the other end taped up.   See the following for a description of adaptive noise cancellation with labview (also possible on an NI FPGA device):

http://zone.ni.com/reference/en-XX/help/372357A-01/lvaftconcepts/aft_noise_cancel/

I believe this to be the only sound signal processing means of addressing this issue.  I am still struggling to get it to work myself.  I'll advice as I progress.

Cheers