Multiple VFD's on Single Iso Transformer w/ GFCI

[amptramp]

Hello all,

The application consists of four 1/2HP 115V 1[Φ] In / 230V 3[Φ] Out VFD's powering 1/2HP gearmotors that drive simple conveyors. The conveyors are dual narrow urethane V-belts riding on steel guides carrying egg flats. The VFD's will be loaded at approximately 75%. The machine has ground fault protection for personnel (Class A device) due to machine operation in a washdown environment. Also, the machine requires CE rating and therefore has an EMI/RFI filter on the incoming line. The drives are tripping the GFCI. An Isolation Transformer is being proposed to prevent the GFCI from tripping.

Common mode chokes on the motor leads are being proposed in lieu of shielded cable. A toroidal ferrite core has been applied (5 turns) to the PE (ground) conductor feeding the VFD enclosure. The GFCI is upstream of the VFD cabinet at the power feed to the machine. The drives manufacturer representative recommended increasing the carrier frequency in an attempt to minimize the leakage current.

1) What are the design considerations when applying multiple VFD's on one isolation transformer?

2) Should the transformer be K rated and/or shielded?

3) What are the design considerations when applying VFD's with a GFCI?

4) Is it practical to utilize VFD's and GFCI's, or in other words, is there any hope of preventing the GFCI from tripping?

5) Any other comments on the application?

Thanks everyone for your time.

Amp

 

[DickDV]

Feed the VFD's directly from the 120V single phase supply and the rest of the 120V single phase loads thru the GFCI.

A CE rated VFD has an internal noise filter which will trip the GFCI. You can't help it.

Interesting suggestions tho. Increase the carrier frequency to reduce leakage to ground! Exactly the opposite is true.

 

[amptramp]

DickDV:{/b]

My thoughts exactly on the carrier frequency. However, a Fluke 360 Leakage Current Clamp was used to measure the current in the PE (ground) conductor at the different carrier frequencies. At the high carrier freq, the current in the PE (ground) conductor decreased! I don't quite understand this result as it is counter-intuitive to my understanding of common mode leakage current.

 

[jraef]

As far as I know, you will never get a cheap little residential Class A GFCI to hold in with a VFD in the circuit. I have had this come up time and again with people wanting to put VFDs on fish pond and swimming pool pumps and it has never worked. Bender Systems makes a GFCI unit that they claim will work ahead of a VFD, but I have never used it.

http://www.bender.org/BENDER_Products/life_guard.html

Is this in the US though and/or are you concerned with the NEC? If so, I don't believe GFCIs are required for hard-wired systems like conveyor motors, only if they are plugged into a socket with a portable cord.


"If I had eight hours to chop down a tree, I'd spend six sharpening my axe." -- Abraham Lincoln
For the best use of Eng-Tips, please click here -> faq731-376

 

[Skogsgurra]

Jeff and others

This application is an 'as wet as wet can be' application. It utilizes a frequency inverter for the pumps(s) and can be connected to a grid with a standard RCD/GFCI.

Test protocol (pages 10-12):

http://www.cyrba.com/PDFdocument/Test Report CIP Electronic System.pdf

Same inverter built into the system as I have told about earlier. I have measured around 0.7 mA leakage current in an application with a 1.5 kW inverter.

Gunnar Englund

http://www.gke.org

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

 

[ScottyUK]

CE marking -> European destination? If so use a 230V VFD, make life easier for the end user.

If you must have an RCD (it's going to Europe - what's a GFCI? ) then consider one for each drive. I have a 15kW A-B 1336 drive to power my lathe running on my domestic RCD, and it doesn't trip the RCD. It does trip the MCB from time to time. Don't forget that most European RCDs are 30mA trip, not 10mA common in the US. 10mA trip types are available in the UK (and possibly elsewhere in Europe) but are not often seen. Might be worth checking with your customer what he is expecting from the RCD.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[jraef]

Class A GFCIs need to trip at 6ma if I remember correctly. 30ma is considered equipment protection, not personnel protection.

I for one am confused.
115V input and Class A GFCI? That would be North America.
CE listing? That would not.


"If I had eight hours to chop down a tree, I'd spend six sharpening my axe." -- Abraham Lincoln
For the best use of Eng-Tips, please click here -> faq731-376

 

[amptramp]

The machines serve both markets (US and Europe) and as such a one size fits all design is being implemented. It is possible the GFCI tripping issue is in the US only. The history of the machines is new for me. The machines are used in hatcheries.

I find it interesting that RCD's in Europe trip at levels high enough to be fatal (30mA). Are these devices for personnel protection?

In the US, the machines are typically sourced by 120VAC (240VAC in Europe).

The NEC requires all 125VAC circuits in wet/damp locations in Agricultural buildings to be GFCI protected (Article 547). This is a washdown application.

 

[amptramp]

Also, the machines are portable and are plug and cord connected.

 

[ScottyUK]

At 230V the cable capacitance can become troublesome. That effect is roughly halved at the lower utilisation voltage in the US. We have the option to use 10mA RCDs, but it is seldom taken because of nuisance tripping. Our final distribution circuits (in the UK) are typically not radials where a 10mA RCD might be acceptable, but are ring circuits serving a number of outlets.

On the 'fatal' comment, 30mA is at the bottom limit of respiratory paralysis and is below the 'let go' threshold. Kinda depends on the reference! I'm sure 30mA RCDs have saved a few lives compared to not having them.



----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[itsmoked]

All the references I see say 4mA is fatal. Hence the US 5mA. I think the English are using their 240V 10mA units as a way to control the population growth.

amptramp; As for your Fluke. I'm not familiar with that one but it strikes me as it could easily not be seeing the leakage that's at a high frequency. It could well be rolling off faster than the leakage is increasing which would give you your unexpected results.

VFDs all have different amounts of leakage. I'd check around and find one with the lowest. I also have one next to me that is plugged into a domestic GFI outlet and never trips it. So I know it can be done. It even sez in the manual "Don't bother trying to run this VFD thru a GFI".

Using one GFI per VFD for VFDs as small as you're using would probably work fine.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[Skogsgurra]

30 mA RCDs save lives. I really do not see why the 5 mA 'unpleasant level' is used.

This document is about European appliances and RCD trip levels:

http://www.epicair.co.uk/documents/zone_information.pdf

Gunnar Englund

http://www.gke.org

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

 

[dpc]

FWIW - my medical instrumentation book has a graph indicating test data showing that 0.5 mA is the minimum perception threshold. 6 mA is the minimum "let-go" current for women, with a mean of 10.5 mA. Minimum "let-go" current for men was 9.5 mA and the mean was 16 mA.

I think these "let-go" levels are the basis for the US 5 mA level for GFCI protection. But the 5 mA level can be a nuisance at times, that's for sure.





David Castor

http://www.cvoes.com

 

[Skogsgurra]

'Let Go' current is a very important parameter.

My thought is that the European 230 V level makes sure that at least 30 mA are produced when you touch a live wire and that the 100 milliseconds breaking time demand is what saves lives.

When you have a lower voltage, around 110 V, then the current produced is less and the 'must operate' limit is set lower to make sure the GFCI operates reliably.

Gunnar Englund

http://www.gke.org

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

 

[amptramp]

Thanks for all the comments so far. Don't quit on me yet though. How about the multiple VFD's on one isolation transformer and related questions?

Also, another aspect of the machine installation is the hatchery often has GFCI (RCD) protection for the receptacle powering the machine, so it is not feasible to provide GFCI (RCD) protection for each individual VFD. I must attenuate/minimize the leakage (common mode) currents for 4 drives. Is anybody familiar with the different types of GFCI (RCD) devices, i.e., Type AC, A, and B. I believe the device jraef mentioned must be a Type B as it is VFD "rated".

Also, I found this link to an interesting publication on electric shock and protection means. There is a graph on page F2 showing body current level versus duration time and associated categories of "damage". It seems to be written from an IEC point of view.

http://www.electrical-installation...._files/EIG-F-protection-electrical-shocks.pdf

 

[ozmosis]

It's not just the value of the residual current, it is the type of residual current to be monitored by a Type B RCD. Type B are, amongst other aspects, designed to operate on waveforms distorted by a rectifier (in the case of most VFD's: a 6 pulse rectifier).
This is important for RCD's looking at the opposite side of nuisance tripping; i.e. when they don't trip due to a waveform a typical type A RCD would not recognise.
If you have a fault where a distored waveform is evident, and the RCD does not pick this up then it is a bigger problem of not tripping than doing so.