VFDs and EMC 8

[Marke]

It is very important that VFDs are installed correctly in order to minimize interference to other equipment.
There are a number of basic rules that should be applied to the installation procedure to minimize problems and these are usually covered in the instruction manual for the VFD.

I have recently come across a number of VFD installations that are causing severe interference with flow meters and other equipment to the extent that some equipment has been damaged by the induced noise levels. These have not been installed in accordance with the manufacturers instructions.

I am looking for a simple test and/or equipment that can be used by an installer to see if the VFD installation is satisfactory. This does not need to be a full compliance test, but just a reasonable indicator that the installation is not "BAD", and that it is safe to install sensitive equipment near by.

Has anyone got any experience or suggestions?

Best regards,

Mark Empson

http://www.lmphotonics.com

 

[itsmoked]

Well my poorly installed 1HP VFD nearly obliterates the screens of all the TVs in the house whenever running. It also sends all radios into a buzzing fit.

It may sound ridiculous but I would actually try a battery powered cheap transistor radio. I bet the sign of all poorly installed VFDs is RFI that an AM radio would sense easily.

$5.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[dpc]

Before reading Keith's post, my first thought also was use of a portable AM radio. These are very handy for checking out EMI.

Nothing quantitative, but a good general indicator of stray RF. The directional characteristics of the ferrite core antennas normally used can help a little in getting a general idea on the source of the noise as well.

Ham radio operators often use them to track down noisy insulators, etc on power lines.

You could baseline the noise using an installation where you are not having any issues. I suspect even in a "good" installation, you will hear some hash in an AM radio.

 

[Skogsgurra]

Mark and Keith,

You are treading on sore toes here! And you probably know it, too!

"Static" in an AM radio is a sure sign. It is being used a lot by guys checking for bad installations in apartment houses. It is also easily understood by anyone. Switch the VFD:s off and static disappears. Easy. No discussion.

This problem is what makes me spend lots of hours on the road. This and bearing EDM.

I have been using a very simple circuit to detect and quantify conducted emissions for many years. It consists of a couple of blocking high-voltage, high-frequency capacitors to block off mains frequency, followed by an HF transformer that is good up to 50 MHz. There is also a 50 ohms resistor parallel to the secondary of the transformer.

With this simple circuit, I get a flat frequency response from about 20 kHz to 10 MHz. Response then drops 3 dB at just above 20 MHz and 10 dB at 30 MHz. 30 MHz is where the conducted emission standards stop.

I use this device together with an oscilloscope with FFT so that I can check for emission spectra in the range 150 kHz - 30 MHz. Very, very few installations meet the EMC directive. And there is very little to do about it. This is an official secret.

You can ask any supplier of VFD:s to show the curves for conducted emission. They will not show it to you. The reason is simply that most (I hesitate to say all, but I think that would be correct) PWM based VFD:s emit a lot more than they are allowed to. I regularly measure 20 or 30 dB above the "home, office and light industry" limits. Sometimes more than that.

Filter or no filter, the complete installation always radiates a lot more than it is allowed to. The reason is that there is no good ground plane to connect the filter drain wire to. So it is connected to the PE, which conducts the HF noise all over the building. It is a sad thing. And denied by all VFD suppliers.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[Valvecrazy]

Might could also use a Canary, like in the coal mines of the past. Animals seem to be very sensitive to this stray voltage from VFD's. I have replaced VFD controls with ATL starters on dairies, and within three days the milk production had increased by 50%. I have replaced VFD's with ATL starters on chicken facilities and the egg production increased dramatically. Same thing on cattle and hog feed lot facilities, and the weight gain increased almost immediately. A canary might just break out into an opera or a bad rap song if in the proximity of a VFD.

 

[Marke]

Hi Keith and dpc,
yes an AM radio certainly tells you that the drive is running, but I am looking for something a little more definitive so that the installers can use it to get an indication that they have a reasonable installation, good installation of a bad installation.
The problem is that it is easy to install correctly and minimize noise, and it is also easy to install incorrectly and make buckets of noise.

Last night I went on site to a very noisy installation, so noisy that the induced voltages actually destroyed some other equipment at great expense.

In New Zealand, we have regulations, based on European, covering the equipment etc, but it is the equipment and the installation that is important and that seems to be ignored by many.

Gunnar, your solution sounds like a good one, but it requires scope and/or spectrum analyzer and of course the ability to interpret the results after correctly driving the equipment.
I am looking for a turnkey solution, but if I can not find one, I may be forced into building one. I was considering a broad band transformer followed by a video amplifier with lots of AGC range and measure the AGC voltage to give an output that can be interpreted and give a pass/fail etc.
I do not wish to be inspecting every installation!

Thank you and best regards,

Mark Empson

http://www.lmphotonics.com

 

[itsmoked]

Gunnar's idea sounds, (typically), good.

But perhaps something like this:

http://www.worldofmanufacturers.com/open.php?

http://www.credencetech.com

Which oddly enough is made in my town?!

I think , Mark you need to characterize a bad site like the one you have now so you can say any number over xx is going to cause you problems.

I could talk to these guys if you want. I could even get one for you and ship it if need be.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[Skogsgurra]

Oops, Keith!

I forgot to mention that I have been using their (Credence) kit of near-field probes (ScanEM-EC and ScanEM-HC) for about ten years. Tell them that they have got a satisfied customer in Sweden when you go to see those guys!

They are little hand-held probes and bought them locally here in Sweden. So, I think you can find them almost anywhere.

They are not good at quantifying levels. But very good at sniffing out an installation and find from where the high levels come. Also to verify if a counter-measure works or not.

Always carry them in my instrument-case.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[itsmoked]

Hey!! HaHa! Small world Gunnar. I have that set too!! I couldn't remember what they're called and I'm not at my office. So I searched for EMI probe and "Santa Cruz" as I remember buying them about 80ft from where my office is now.

http://www.credencetech.com/products/product.php?productId=CTK019, CTK015

These right?

They work great. But since they are sort of ambiguously adjustable I too doubt their usability for Mark.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[Skogsgurra]

Yes! Those two.

The scope connection is very useful. The output reveals things that are hard to measure with ordinary scope probes.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[Marke]

Today, I have wound a broadband transformer and terminated input and output, and fed it via a 2n2 capacitor from the phase and neutral. On the output, I have connected a schottky diode and a millivoltmeter. If I connect a drive and motor to the supply, I get a small reading when they are running, and no reading when they are not. I think that there is the potential to make a plug in meter that measures across phase to neutral and gives a reading representative to the amount of noise there is on the input to the drive. This would be an indication of the conducted emissions which gets pretty bad when the drive is incorrectly installed. I will do a tidy up and then visit a good site and a bad site to get a measure of the range of output levels.
This will not in any way show compliance with the standards, but it should indicate "safe to connect other equipment", which is really what we want. Compliance is another issue and for others to sort out. We have the standards here, (same as IEC) but no one following up, so anything goes until something blows.

Best regards,

Mark Empson

http://www.lmphotonics.com

 

[Skogsgurra]

A small coincidence (it isn't really). Just got an order from the dominating energy company over here to go and find out why their large heat pump doesn't work any more. MW meter based on temperature and flow indicates broken wire in a 4-20 mA temperature loop.

- Any new inverters installed lately?" I asked.
- Yes, a couple of weeks ago, the guy in the phone said. - That's when the problems started. Do you think there's a connection?
- Yes, I said. Definitely.

I could give you tens of examples. Here are some:

Motor lab for large diesel: Vibration monitor quits working when 800 kW VFD brake regenerates.

Machine for wire drawing: Counters indicate crazy footage when new VFD driven blower starts.

Automatic Meter Reading: Utility loses readings when ventilation revvs up.


Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[Valvecrazy]

Skogsgurra, have the examples you mentioned been resolved? When grounding and filtering doesn't work, do you know any other way to reduce the noise to at least an acceptable level? Or am I asking for one of your trade secrets?

I also get called to systems where VFD's are causing interference with radio controls and other things. So far the only way I have been able to resolve the situation is to remove the offending VFD or VFD's.

 

[Marke]

Hello Valvecrazy

There are a number of simple steps that can be taken to reduce the EMC and these are usually described in the installation manuals. They are not trade secrets although there may be some extra sneekies around as well.

The first step is to ensure that the drive has been designed with EMC in mind. If it is a bad drive, you may be able to improve things, but it will still be bad. I have worked with some drives where there was no way that we could get them to meet the EMC standards under laboratory conditions, - they were just not able to meet the spec. Drives designed for use in Europe must be designed to meet the CE requirements and this will be defined in the manual along with any special filters required in order to meet spec.
The first step, is to use a proper screened cable between the drive and the motor. There are different levels of screening and they have very different results. The best cables have a very tighlty woven copper screen (sometimes double screened) and distributed earth conductors (between the phase conductors).
Second, terminate the screen correctly at the motor AND at the drive. The correct termination is to use a screened gland of clamp to secure and bond the screen to the drive case and the motor case. Pig tail connections are very inferior and produce a much degraded performance.
Third, if there are earth conductors in the cable, take them to the earth connection on the drive and motor.
Fourth, any control cables should be screened and the screen should be earthed at one end only.
Fifth, if the manual describe or specifies input and/or output filters as necessary for CE complience, then these should be used. It is important to use the types specified for best performance.

If this is done correctly, it will significantly reduce the EMC problems and in most cases, will overcome issues.
Further improvement can be achieved by the addition of some ferrite rings on the output conductors such that the thre phase conductors pass through the ferrite rings together, and the addition of an EMC on the input will help to reduce the conducted emmissions.

Definite DO NOTs.
Do Not connect an earth from the motor back to the main earth. The only earth connection should be to the output of the drive.
Do Not bond one end only of the screen of the cable betwen the drive and the motor. Both ends must be bonded.

Installation with submersible pumps.
Run unscreend cable from the pump up to a junction box on the well head.
The well head junction box should be a metal box with the box directly mounted on and bonded to the well head.
Run a screend cable betwen the drive and the wel head box with the screen(s) bonded at the drive and the well head box. An additional large flexible cable between the well head box and the drive output earth can make an additional improvement also.

If you do the above, you will usually overcome the EMC issues. It is dificult to get the installers to do the above, most argue that you must only bond one end of the screened cable if you use a screened cable.
The best performance is limited by the installation and by the drive design.

Best regards,

Mark Empson

http://www.lmphotonics.com

 

[Skogsgurra]

Valvecrazy,

If they have been resolved? there are different answers to that.

The one I am going to check next Monday will probably turn out to be a case of sensitive transducer or transducer installation that can be filtered so it doesn't react to the EMI. Such fixes are a lot easier than changing the installation. This is a megawatt motor and everything is big and heavy. Not easy to do anything about the installation.

Motor lab: Same thing. An 800 kW machine. Not easy to do anything about the installation. It has been done according to manufacturer's manual and manufacturer has had a go on it. Several, actually. I have tried to "ruggedize" the transducer installation, but no success so far. Machinery used for long-term testing. Not easy to get hours to work on it.

Wire drawing: Inverters had been installed according to manufacturer's spec. Had to filter counter inputs. Nothing more than 1 nF ceramic capacitors across counter inputs was needed.

AMR: Not much to do. My personal opinion is that AMR will have increasing problems as more and more inverters are being installed. One problem is that filters that are used to filter out conducted EMI also tend to attenuate the carrier. Seen that quite a few times.

Mark has given some very valuable advice above. But even if you follow these rules, there are still cases where it doesn't help. I guess those are the cases I get to see.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[Valvecrazy]

Thanks Mark
The ferrite rings and the double grounded cable are things I have not tried.

 

[Skogsgurra]

ABB Technical Guide #5 is worth reading.

About the "ferrite rings". They are usually either amorph metal toroids or iron powder toroids. Ferrites seldom work, too low permeability and to low saturation point to be useful. Mark knows what rings to use.

Such rings, also called Common Mode Filters are not only good at removing EMI, they also help keeping EDM in bearings low.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[Marke]

I recently visited a site where the electronic flow meter was not recording.
I measured 50V peak between the electrical earth connected to the flow meter and the main earth.
The inverter output was wired in screned singles with the screen bonded at the inverter end, but tapped off (open circuit) at the well head end.
A bonding strap was added to the screens at the well head end and that was bonded to the well head, and the problem went away.
Another installation, the cables (screened) went down the well with no bonding to the well head. The problem was that a nearby house could not use electric clocks (went fast) cordless phones were blocked, TV and video were interfered with.
A bonding strap from the screens to the well head fixed the problem (after 3 months of attempts by others)

These were easy fixes when done to instructions.
The problem is that there are so many "experts" out there giving the wrong advice.
Hence my search for a simple indicator that can be used to show improvments etc.

I believe that after the design of the drive, the most important thing is the low impedance path between the motor and the drive, and the screened connection is he most effective.

Best regards,

Mark Empson

http://www.lmphotonics.com

 

[Valvecrazy]

I was reading about a "swinging choke" that is supposed to help a lot with harmonics at part load conditions. Has anybody had any experience with these?

 

[dpc]

Definite DO NOTs.
Do Not connect an earth from the motor back to the main earth. The only earth connection should be to the output of the drive.

Marke,

Good info.

Just to clarify - the motor frame is grounded via the ground wire carried with the motor phase conductors coming from the drive? The motor frame has to be connected to earth in some fashion, at least in the US. The NEC does allow alterations in grounding to stop "objectionable" currents.