VFD Cable Screening Correction in MEW252?

[SillyOldDufferModerator @sillyoldduffer]

'On the Editor's Bench' in MEW252 (aka MEW251!), concludes with a paragraph headed "VFD issue".

In it Neil says "I have been contacted by a couple of electrical engineers about Laurie Leonard's VFD article in MEW 251. They both point out that the screen for the motor connection cable should be grounded at BOTH ends to be fully effective in stopping interference and leaving one end unconnected could make it ineffective."

Unfortunately within my experience, I think that Laurie's article is right, and the correction is wrong! As Laurie explained in his article, "It is usual practice to earth one end of the screen (one end only to prevent the introduction of a circulating current path via the earth system). (My bold)

Common mode interference is a well known problem in electronics and radio engineering. VFD's are a potent source of high frequency energy. High frequency currents behave very differently from AC and DC power. HF power can jump gaps and ignore 'earths' in entirely counter-intuitive ways. Earth loops are avoided because they are likely to be an effective radiator.

As far as I know, EMI screening advice consistently recommends a single point RF earth, just as Laurie said. But I know little about 'best practice' when it comes to wiring motors to a VFD: what do the team think?

Thanks,

Dave

Edited By SillyOldDuffer on 18/02/2017 15:32:23

[SillyOldDufferModerator @sillyoldduffer]

[Ian ParkinParticipant @ianparkin39383]

All motors in my workshop with vfd's are mainly connected with 4 core flex from the vfd

on a few its screened and i have only connected screen at the vfd end

I have 18 such motors in all on my machines from 100 watts to 3kw

[David JuppParticipant @davidjupp51506]

Control cable screen should be connected to ground at one end only, because it is the control cables that are vulnerable to interference due to circulating currents.

Motor cable screens should be connected at both ends to minimise RFI problems created by the VFD (conducted via mains supply or radiated).

Much of the time you won't notice the difference – occasionally you might, others in the vicinity or on the same supply might.

[Ex contributorParticipant @mgnbuk]

Ulimately, the correct approach surely is to comply with what the particular inverter manufacturer recomends in the specific device documentation & not try to generalise !

Nigel B

[Andy HoldawayParticipant @andyholdaway]

I'll throw my tuppenceworth in here – we install around fifty or so VFD's in industrial air conditioning units a week, ranging in size from 4kw to 70kw, and the advice from Siemens is that the 3 phase wiring screen from VFD to motor should be earthed only at the VFD due to circulating currents. A demonstration video was presented to show how circulating currents could 'cook' the bearings if the wiring was not done in this way.

I should also add that our A/C plant has been tested and verified by a third party to ensure it is compliant with current EMC standards (radiated and conducted) with no problems, neither have we had bearing problems caused by heating.

These units go in to retail and industrial sites, including data centres. I'm sure we would have heard by now if we were causing a problem with RFI. In the home workshop I think the worse that could happen would be a bit of white noise on your radio!

Andy

[Michael BriggsParticipant @michaelbriggs82422]

As David says, Rockwell (Allen Bradley) for example advise : Shields of motor and input cables must be bonded at both ends to provide a continuous path for common mode noise current. and Connect the shields of control cables only at one end. Regards, Michael.

[Michael BriggsParticipant @michaelbriggs82422]

Nigel B has the answer.

[Neil WyattModerator @neilwyatt]

OK, this is the explanation I have received (essentially the same from from two independent sources):

"The purpose of the screen for the connection between a VFD and the motor is to reduce the electromagnetic emissions due to the nature of the motor current. If the screen is connected at one end only it is completely ineffective in reducing the emission of magnetic interference. The explanation for this is well covered in many texts, not to mention VFD users manuals, so I won’t repeat it here. I would guess that your author is confused between the dilemma of circulating noise currents in screens that can be an issue to solve in screening of small signal instrumentation and the need to provide a closed circuit by earthing both ends of the screen for it to be effective as a screen against magnetic induced interference. Small VFDs normally have a metal terminal frame for clamping the motor cable screen to, referred to as an “EMC grounding flange” in the case of the Jaguar VFD. It appears your author has ignored this and not used it. At the motor end a proper EMC cable gland should be used. These are metal and include conductive fingers to completely encircle and earth the screen. Plastic cable glands such as illustrated in your photo 27 have no place in VFD to motor cabling if the screening is to be effective. "

In brief, most people are familar with screens to stop signals within being affected by noise, these work best if circulating currents are avoided.

In this case we are trying to keep the noise IN and any holes will let it out. Your microwave doesn't keep the radiation inside better if there is a hole in the casing…

Neil

[Neil WyattModerator @neilwyatt]

The IMO manual I have says the screen should be earthed at both ends to comply with the RFI directive.

Also, I'm happy to take the Seimens advice at face value, but this seems to be a case of balancing one possible negative against another, not a suggestion that screening is less effective is one end is unconnected?

Neil

[Andy HoldawayParticipant @andyholdaway]

At the risk of shooting myself in the foot, I have just looked up the recommendations given by Danfoss, as we may be using some of their VFD's soon, and their installation instructions say that the screen should be earthed at both ends!

Our A/C units form a pretty big Faraday cage in themselves, which will help as far as radiated emissions are concerned, but it's interesting that two major manufacturers quote conflicting information. Time for a bit more research, methinks!

Andy

[MuzzerParticipant @muzzer]

There's no right or wrong, just what works for each installation. And no recommendation will claim to ensure compliance with any directive or standard. I've found both approaches to be the requirement in different situations. Grounding both ends can cause nice resonances, just as grounding one end can. The motor will be grounded anyway but usually with a protective earth (fault currents etc), rather than an EMC grounding arrangement, so there is always the potential for ground loops, resonances etc etc, no matter what you do.

Insisting there is only one way and it's my way is not going to end well. I've done a lot of EMC compliance testing of mains connected equipment over the years, from 10s of Watts to 10s of kW and it's far from a science although there are certain lessons you learn along the (hard) way.

Most VFD units sold to the likes of us seem to have no EMC filter as standard anyway, so they are most likely appalling and I doubt many would come anywhere near meeting the formal requirements. You can often buy an optional filter for the VFD input and you could do worse than fitting one. It's also possible to fit a filter at the motor side but this isn't quite as straighforward.

Murray

[David JuppParticipant @davidjupp51506]

I have come across the damage to bearings by circulating currents issue that Andrew mentions – in a moderate size HV motor (a couple of MW or so if I recall correctly). That was caused by the 11kV (not VFD) supply because one bearing was not insulated from ground as it should have been (an accidental short to ground via an alignment shim).

I honestly don't know enough detail to say if VFD for larger drives are more probe to this issues than in small motors where both bearings are usually connected together via the frame.

[Author]

Posts