Inverter Drive Installation--EMC Compliant 2

[Ahr35181]

I work at a large pharmaceutical plant in the UK & we have many VSD drives on pumps, agitators & fans etc, mainly Danfoss & ABB. Both manufacturers recommend segregation of supply, motor & control cables & the use of screened motor cables to comply with EMC regulations. Our installations don’t comply with these requirements, we use XLPE SWA cable & never segregate. In two of our switch rooms from which large drives are supplied I’m unable to measure EFLI & Ip at fuse boards whilst these drives are running, my robin loop tester indicates 240V, then drops to say 40V, goes back up to 240V etc etc!!! When I press the test button to try & measure EFLI I get differing high readings (16 ohms for example!!!). With the drives off I get good readings in the region of 0.01 ohms (we have an 11KV distribution network around the site, thus the low reading). We’ve recently had some kind of EMC survey carried out on the site, the report was positive, i.e. no problems. The large drives mentioned are obviously causing disturbance in the supply.

I’d be interested to hear people’s views on this, particularly anyone genned up on EMC. If it helps, all our inverter drives are fitted with filters, although I gather the filter is pointless if the instalation is not correct.

Thanks,

Alan

 

[jOmega]

Alan,

If you don't abide by the manufacturer's wiring and installation recommendations, then I'm afraid you deserve all the grief you get. Those recommendations are made based upon having had to clean-up after people who also thought they could ignore the recommended practices.

Clean up your installation wiring, and then come back and tell us if you are having the same problems...

Sorry I can't be more sympathetic...... but you see.... I've spent too many hours in the field cleaning up such messes that could have been prevented.

The cure is always more expensive after the fact than before the fact. If you have to re-do the wiring, to clean up the problems, you'll end up having to pay more than twice what it would have cost to do it correctly at the time of installation.

jOmega

 

[Skogsgurra]

I agree totally with jO, but if you read the "do it right-manuals" from manufacturers like Siemens (they have "Twenty simple rules for EMC-correct installation&quot Danfoss, ABB and Telemecanique, you will find that you start contradict yourself after having applied the first three or four rules. So the finer parts of the rules often simply cannot be abode to.

I have measured several installations where everything was formally correct (some of the installations were actually carried out by the inverter manufacturer) and still had around -30 dBu avg in the 2 - 6 MHz region. I was using the voltage probe according to EN 61800-3/A11 and the result is some 30 dB above the B limit.

The reason for this is - as I see it - that inverters are tested for compliance in test rooms where there is a good ground plane and ideal conditions in every other respect. The actual installation site does not have the same good ground plane and everything that the filters filter away (the HF "reject&quot is dumped to the PE, which distributes the HF all over your installation.

I have seen this in many installations and your experience is shared by many. I think that jO should be a little nicer to you. You did not do the bad installation. You are the victim.

I have cleaned up these things in several plants to a point where you still have EMI way above the EMC limits, but where your measuring instrument start to work again. I have found that the blocking method is a lot more efficient than the filtering method. By applying ferrite cores (yes, large ones) around motor cables and supply cables you can often block the propagation of the PWM fastest edges and reduce the EMI to levels where your equipment works again. And then some extra care and perhaps normal mode filtering at the instrument inputs often helps.

Good luck!

Gunnar Englund

 

[jbartos]

Suggestion: Ideally, notch filters for 60Hz (or 50Hz) will solve EMI problems. However, the efficiency would go down since a lot of energy contained in harmonics will be discarded. The high frequency energy is relatively small so the low pass filter would be sufficient as long as the harmonic content is acceptable.
VSD may have:
1. A filter suppressing dv/dt
2. A filter suppressing carrier harmonics
3. A filter suppressing harmonics due to DC to AC power electronic switching architecture (on the VSD output)
4. A filter suppressing harmonics due to AC to DC power electronic switching architecture (on the VSD input)

 

[ScottyUK]

I've seen a very similar problem with a Robin loop impedance tester while testing in on our engineering office block. In our case the instrument indicated that the PFC was over-range at >20kA, at the wall socket, which is simply not possible.

The problem seems to be harmonics, because this building is mainly fluorescent lighting and PC's. We repeated the test with the PC's switched off one night, and the Robin instrument behaved normally. The same instrument behaves normally when testing boards with primarily linear loads.

I expect your drives are probably putting harmonics into the supply in much the same way as our PC's were, and this is the source of your problem. I don't think the incorrect installation of the drive is neccessarily causing the odd effects you are seeing - I suspect they would be there even if the cabling was shielded. Your installation is still bad practice and you have given yourself an additional unknown factor to worry about.

 

[jOmega]

Alan,

Appology offered if you took my comments personally.

They were not intended as a personal attack, but rather, a commentary on the intelligence (or lack thereof) that dictates installations with total disregard of mfgr's recommendations, and then complain that they are having consequent problems. The "you" in my remarks was not a personal pronoun; but rather a collective pronoun directed at the "company".

"You can lead a horse to water....but you can't make him drink"​

Some of the things they can try (it's only money) are EMC and RFI filters on the mains.... They can also try Ferrite Beads on the mains... and on the motor leads... and on the signal / control leads...


jO

 

[ozmosis]

The guidlines issued by manufacturers have to be issued, by law. The guidlines have to be adhered to by the installer, again by law. Remember, the EMC Directive is there to protect OTHER users on the same point of common coupling(PCC): on the same public network. What you do in your own backyard is effectively not covered by the Directive(the law) but should be covered by the product standards available. The EMC Directive is 'complaints driven', meaning; if another user on the same PCC complains to the EMC 'police' (in the UK it is the Radio Communications Agency [RCA])then they will start to investigate where the problem comes from. If the installer has not adhered to the installation guidlines issued by the manufacturer, then the buck stops with them and they are in trouble. If the installation is good and there are still problems, the RCA investigates the tests carried out by the manufacturer. If these tests are deemed incorrect, or worse fraudulent, then the manf is in serious trouble. It's a question of responsibility all down the chain.
Note also, there are two aspects to the VSD specific standard (EN61800-3): Emissions and Immunity. Too many people get hung up on the emissions from the VSD's (quite rightly in a lot of cases) without looking at the immunity of the products the VSD is effecting. The VSD could be operating within the emission levels required for that installation but the products being effected may not have the necessary immunity class. It's about getting parity between emissions and immunity. Having said this, if the installation is bad, do not waste your time on filters (just yet). An EMC filter in a bad installation can often make the problem worse. My experience is that it's the simple things that cause problems. EMI is often a black art, high frequency interference has very different characteristics to a 50Hz supply and ensuring these unwanted 'electrons' have a good low impedance path to earth is what makes for a good installation. Bonding of the motor cable screens, bonding of the cabinet, equipotential bonding of the drive and motor and any other device connected to it is critical. Note the word bonding. A piece of round, small gauge earth wire is not often suitable for getting rid of the interference. It might meet the local safety standards (50Hz faults) but not 150kHz to 30MHz noise. Use braided cable or flat metal straps. More expensive, yes but it works. Planning is the key. Plan the installation with EMC in mind and the chances of getting problems once the system is up and running will be greatly minimised!

 

[jbartos]

Suggestion/comment: Often, it is seen that the product complies with industry standards and regulatory agencies. Then, the external installation/wiring should have a minimum interference impact. Also, control and power wiring should be addressed in the product literature or via the manufacturer's tech support. Larger projects, e.g. power generating stations have separate trays for different voltage levels, controls, and instrumentation cables. The only exception appears to be fiber optic cables that may be run with various other cables.

 

[rodar]

Hi
One thing that helps is making sure the contractors
pull a individual ground wire along with the phase
wires going out to the motor. I have worked with groups
that wanted to daisey chain the grounds at the motors
but this is not acceptable for EMC. The primary trouble
is the common mode voltage in the phase wires that
couple to the frame of the motor. These common mode
voltages originate in the drive itself so the path of
least inductance for their return is a ground wire with close physical proximity to the phase conductors for the whole length of the run of wire between the inverters and
motor. Bring each ground back onto the VSD panel and ground
it at the proper VSD drive.

Hope this helps
rodar