VFD Common-mode noise 1

[seamus000]

Suspect that VFD Common-mode noise is effecting instrumentation throughout plant site.

Have installed isolator on a single loop going to chart recorder and it clears-up. However, it is out of the question to stick isolators on every loop in the plant.

Have line reactors on the VFD outputs, have added chokes, no change in noise level. AB Drives.

Any suggestions on how to:
a) identify the actual source, and cause?

b) how to correct?

Thank you in advance.

 

[GusD]

Hello seamus)000

You state that VFDs CMVoltage, could be the cause of all your instrumentation problems.You mention that when you installed an isolator on a single loop,it seem to clear up.
Did the problem clear up all over the plant, or just on that loop.We use many VFDs,AB,ABB,Cegelec and others.Individual systems may gives us problems from time to time ,but I don't know of any problem or problems that would affect the entire Plant.I guess if you only have one power transformer,that could be true.
I just came across some information, that appears to go directly to your problem.The radiated EMI noise can be picked up by comunication,instrumentation or signal equipment.It appears that the reccomended cures are also specific to the Freqs. encountered.Above 200 Mhz,
Simple feed-through capacitors for the AC Supply,coil
suppressors and proper motor and control shielding to
cables,motors and enclosers.
14 KHz and up.A special filter for the AC Supply lines.
Conducted narrow band Harmonics at the lower freq.end,are
best dealt with one of the time.This should be evaluated in conjunction with the power system.

Somebody came up with all this info,so it must have worked one time or another.

On a personal note I would like to add the following .
In any plant, equipment malfuctions,breaksdown,burns up and whathave you.When this piece of equipment fails to deliver ,in a very short time, some leader,manager or other appears to remind us that our company is loosing serious dollars until I,or someone else fixes the problem.
Maintenance people on shift work,most of the time working alone,have to find the problem and get it to work as fast as they can.Even in the old days,We did have all of the equipment we have today,SCRs ,VFDs, EDDY clutches and all the others.Experience told us that 95% of the problems, are something much simpler like a broken wires,dirty contacts, etc,etc.
By fixating one's attention to solve a particular fault, that may or may not be the source of the problem,it stops us from looking at the broader picture of possibilities.
I still give myself this same advice from time to time.

Good Luck.Lets us know what you find.


GusD

 

[jbartos]

Suggestion: It may become very difficult to mitigate noise caused by very strong noise generators such as large VFD with very little noise mitigation within. There are some industry standards available dealing with permissible noise levels and noise mitigations. Also, some papers have been written on this topic. Visit

http://www.ab.com/drives/techpapers/ieee/ieee.html

for:
Gary L. Skibinski, Russel J. Kerkman, and Dave Schlegel, "EMI Emissions of Modern PWM AC Drives(text), graphics" IEEE Industry Applications Magazine, Vol. 5, No. 6, November/December 1999, pp. 47-81. (Requires Adobe Acrobat 4.0 or higher). WINNER OF: IEEE IAS magazine prize article for 2000. Russel and Gary will accept the award at the IEEE IAS conference in Rome.

etc.

 

[RadarRay]

While some of this re-iterates Gus' advise, some of this goes into more detail.

I have seen two types of noise problems on AC drives and servos. One is when you supply power to the motor, the other occurs as soon as you enable the drive. I would be willing to bet that you get the noise as soon as the drive is enabled, even if you are not drawing any current. Is this so??

To troubleshoot, you should really connect an Oscilloscope to your signal before you start so you have an idea of the magnitude of the noise. Then you can see what changes occur. Also look at the signal with no drives energized, with the drive enabled but drawing no current and then drawing current.

In general all these rules apply. First your system has to have a good, KNOWN, earth ground. I say "known" because I have seen customers use steel beams they thought were ground when in fact they were not adequet.

Second the power cables must be connected properly. The shield on the power cables ( you are using shielded or armoured cable, I assume) should be continuous all the way back to the power unit. We use a large amount of Eurotherm drives, and they provide terminals to connect the motors leads in the cabinet. Along with the 3 leads they always provide a terminal to connect the shield to. I have seen customers, tie the shield and the armour to the cabinet as the cable entered the cabinet and left about 3 feed of cable unprotected. As bad as having the cable unprotected is the fact that any currents in the shield now run as they like though the cabinet and not directly to the power unit as they should. You can argue that it is all grounded but in fact, the paths taken by the current are different then intended.

This is what you can do on the signal side. I assume the cable you are using for the signal are twisted shielded. You can now try to tie the shields at different locations, and will get different results. In order of preference, I tie the shield to ground in this order, when trying to resolve noise problems.
1) At the drive cabinet only.
2) At the sensor only.
3) At both ends
4) At neither end

I have seen all the above work, but usually best results are obtained using the first two approaches.

If none of these work, describe the system a little better, as there are a couple of other solutions which could be beneficial.

 

[Farkel]

I saw you had put line reactors on the motor end of the drives, but not on the line end. This would be most helpful.