Variable Frequency Drive tripping out the earth leakage in my house.

[FrankE]

Just remember that - residual currents are summed

I swapped out every MCB for an RCBO to prevent nuisance tripping, each room on it's own ring main and put the fridge and washing machine on their own circuits to reduce the consequences of ring mains being down.

My Iso-Tech ILCM is unfortunately in the 'repair someday' box and it was bloody handy when it was a working unit. When it worked I had an IEC lead with a section of sheath stripped from mid cord to measure earth leakage from units with IEC inlets.

Could you borrow a CT to plug into a multimeter, a PA tester or an earth leakage clamp meter?

It's a good idea to have separate circuit for your bench (with isolating transformer and emergency stop) and workshop equipment.

[AllTheGearNoIdea]

There is a RFI link that can be cut taking the filter capacitors out of circuit.   

Dear sweet WTF. Lets cut out the circuitry that lets the unit comply with IEC regulations for RFI and EMI

Or perhaps you could look into RCDs that are especially manufactured for just this problem. Look up type F & B RCDs. F generally for single phase VFDs and the B and B+ models for three phase industrial machines.

Dont see the point in looking at RCDs that are not allowed for domestic installations.   Im not loosing much sleep over the EMI for a machine that  going to be run for a only a few minutes at a time every few months or so.  I would rather have a safe installation.

Chris

[AllTheGearNoIdea]

Just remember that - residual currents are summed

I swapped out every MCB for an RCBO to prevent nuisance tripping, each room on it's own ring main and put the fridge and washing machine on their own circuits to reduce the consequences of ring mains being down.

My Iso-Tech ILCM is unfortunately in the 'repair someday' box and it was bloody handy when it was a working unit. When it worked I had an IEC lead with a section of sheath stripped from mid cord to measure earth leakage from units with IEC inlets.

Could you borrow a CT to plug into a multimeter, a PA tester or an earth leakage clamp meter?

It's a good idea to have separate circuit for your bench (with isolating transformer and emergency stop) and workshop equipment.

Hi frank

Thanks for the commnet, I do already have an isolated bench in my electrical shop but this is for the garage where i have a  new single phase milling machine using a VFD.  I have ordered a sensitive clamp meter and I am waiting for it to arrive. I am going to investigate the standing leakage current for my installation as it could already be high and the machine start up is just pushing it beyond the limit  of 30mA. Im also going to borrow a installation tester to check out what its actually tripping at.

Many Thanks Regards Chris

[Zero999]

[imghttp://][/img]

I was thinking about something similar to this. The reason im considering bonding the neutral to earth on the secondary of the isolation transformer is so that a real earth fault on the machine will clear the circuit breaker.

Does anyone have any thoughts on this.

Regards Chris

The only issue with that is, it voids any protection against shock, given by the isolation transformer and you then have a circuit which is not protected by an RCD or isolation transformer.

Instead, you could install a leakage current monitoring system to warn you if there's an earth fault.

[AllTheGearNoIdea]

Hi the circuit does show an RCD on the secondary side of the isolation transformer. That’s why the neutral  is strapped to the earth conductor.  But I cannot say I am happy with any of these arrangements.

I thought I had hit gold with my previous post about disconnecting the RFI filter on the drive but on close inspection I cannot find the link.  It could have already been cut at the factory or I may have the wrong model of VFD. Unfortunately the model number is obscured by a transformer that cannot be removed without removing the whole panel back plate. Bloody annoying.

I did however spend an hour investigating tonight. I disconnect all the various circuits that are pulled in by the main contactor including the VFD.  I then put them back one by one and power cycled the machine.

Well much to my surprise I found that it’s not just the VFD Tripping the RCB. The main spindle motor has a auxiliary mains powered fan for forced air cooling at low speeds.  When I connected this fan the RCB tripped out about 40 percent of the time at switch on and about 40 percent at switch off. 20 percent of the time it was OK.  Well this is a first for me, the RCB Tripping when the load is disconnected ?

With the fan motor disconnected I did a 500 volt insulation resistance test on the motor and it passed at nearly 1 Gig Ohm.  I also powered the motor with the neutral disconnected and the live connected and  this should have tripped the RCD if there was a real earth fault but it did not trip the RCD.  I’m starting to think it is an issue with a spike but this is a relatively small motor so would not have expected this caused a trip.

I also plug transformer based welders and welding inverters into this supply these don’t trip the RCB so I’m really at a loss.

It’s annoying I cannot power the machine up off the RCB protected supply without Tripping  so I am unable to make measurement. Running via a transformer hides any potential fault so again cannot make a ground reference measurement.

I’m starting to think I should run the machine off my petrol generator. I’m assuming that neutral and earth are the same point on a generator  but I can at least get some real measurement without the RCB tripping out constantly.  Or strap neutral of the secondary to the machine earth pin for testing. It’s no longer  acting as an isolation transformer but this should be fine fo testing.

[calexanian]

I have had this problem myself. I tracked it down to RFI down the mains line. I re bundled the motor leads in metallic conduit and warped the line cord a few times in a ferrite and made a common mode choke. That fixed it. VFD's are bassically little low frequency radio amplifiers and the high frequency/relatively high current has a way of getting into things and causing problems.

[Edison]

Equivalent circuit for measuring

[Electro Detective]

I'm not sure if an AC leakage clamp meter will show much, especially if the issue is high frequency based or chopped or distorted waveform/s via cabling bleed kicking all over the shop

There may even be a freaky DC component in there too 

Even the better big dollar AC clamp meters might extend from 2khz and stretch to 10khz,
most roll off at 500hz to 2k, they are really geared for 50/60/400hz apps afaik

A cable break current measurement with a 100khz AC bandwidth multimeter might be a better way to go,
even a Fluke 87V may not cut it, it rolls off at 14 to 20khz  and your issues may be at 40 to 70 khz 


The transformer is obviously killing off the generated leakage nasties because it's a LF device for starters, coupled with the isolation/magnetic/air decoupling factor 

Not sure how a generator would cope with such a VFD load btw, especially if the gen has onboard regulator and cutoff electronics, so 'try before you buy' 


If you don't find an issue with the VFD itself or a safe workaround, I'd be going the way I do it for -temperamental- devices

I plug them into a properly rated isolation transformer with FULL ground/earthing on the input and output and the metal box chassis and transformer body,

and check that the central earth/ground bonding point isn't a halfassed job done with flimsy washers/screws/bolts,
squashed on to the paint or powder coating by some Friday afternoon manufacturing process from across the the ocean...  

and transformer OUTPUT neutral and earth ground tied together


Most MIS-LABELED so called 'isolation transformers' come wired this way anyway, aka 'medical isolation transformer' labels etc
so all you have to do is open the box and verify for yourself visually
and or do it the easy way with a good multimeter on Ohms continuity check (not buzzer) if the box is rivetted/pita to get into.
Naturally you do all this with the tranny in the ON position and disconnected from the death mains!   

Basically if the earth/ground reads ok everywhere expected, and the transformer neutral and earth/ground read as connected 'zero ohms',
you should be good to go and 100% legit,
no DIY hackery or isolation guesswork and prayer required     

With a transformer wired this way you also have the option (recommended) to plug the VFD into a portable RCD/GFCI (and maybe a MCB too) and into the transformers output,
and if it trips due to a mishap, your earth leakage device upstream at the switchboard won't know or care
You can also use the TEST button on the portable RCD/GFCI to ensure the function works

otherwise... it's oscilloscope time, with all the pita 'proper procedures' to suss out safely wth is going on where...
without taking out the scope    or zapping the already frustrated VFD owner...   

[Zero999]

Hi the circuit does show an RCD on the secondary side of the isolation transformer. That’s why the neutral  is strapped to the earth conductor.  But I cannot say I am happy with any of these arrangements.

Sorry, I missed that, but adding the RCD after the transfromer would just move the nuisance tripping elsewhere. I suppose at least it's not the whole house being taken out, so it's an advantage.

And the correct term for the equipment to detect insulation failure, in a circuit powered off an isolation transformer is an insulation monitoring device.

I thought I had hit gold with my previous post about disconnecting the RFI filter on the drive but on close inspection I cannot find the link.  It could have already been cut at the factory or I may have the wrong model of VFD. Unfortunately the model number is obscured by a transformer that cannot be removed without removing the whole panel back plate. Bloody annoying.

I did however spend an hour investigating tonight. I disconnect all the various circuits that are pulled in by the main contactor including the VFD.  I then put them back one by one and power cycled the machine.

Well much to my surprise I found that it’s not just the VFD Tripping the RCB. The main spindle motor has a auxiliary mains powered fan for forced air cooling at low speeds.  When I connected this fan the RCB tripped out about 40 percent of the time at switch on and about 40 percent at switch off. 20 percent of the time it was OK.  Well this is a first for me, the RCB Tripping when the load is disconnected ?

With the fan motor disconnected I did a 500 volt insulation resistance test on the motor and it passed at nearly 1 Gig Ohm.  I also powered the motor with the neutral disconnected and the live connected and  this should have tripped the RCD if there was a real earth fault but it did not trip the RCD.  I’m starting to think it is an issue with a spike but this is a relatively small motor so would not have expected this caused a trip.

I also plug transformer based welders and welding inverters into this supply these don’t trip the RCB so I’m really at a loss.

It’s annoying I cannot power the machine up off the RCB protected supply without Tripping  so I am unable to make measurement. Running via a transformer hides any potential fault so again cannot make a ground reference measurement.

I’m starting to think I should run the machine off my petrol generator. I’m assuming that neutral and earth are the same point on a generator  but I can at least get some real measurement without the RCB tripping out constantly.  Or strap neutral of the secondary to the machine earth pin for testing. It’s no longer  acting as an isolation transformer but this should be fine fo testing.

I know I'm probably clutching at straws here. What sort of motor does the cooling fan use? Is it brushed? Does it have a filter capacitor which is leaking AC current to earth and tripping the RCD? Or perhaps it's EMI which is interfering with the RCD, causing it to trip?

[Elasia]

I might be lambasted for this but I've also had to deal with this issue working on projects out of the house with similar motors.

Put the motor on a dedicated circuit with proper grounding and circuit protection just like in a commercial / industrial setting and toss the resi rcd.  Problem solved.  When you go to move.. remove the extra feature so the locals don't bust your balls.  I got an entire lab in my basement of gear that would just eat a rcd/gfcis lunch for breakfast.

Yes they provide great value at protecting a nation of darwins at large but in some specific use cases they are just garbage.  Newer models are getting better though at dealing with current inrush and voltage spikes since you can get ones with microprocessors in them now that can figure out if its real or just another motor.

[sibeen]

There is a RFI link that can be cut taking the filter capacitors out of circuit.   

Dear sweet WTF. Lets cut out the circuitry that lets the unit comply with IEC regulations for RFI and EMI

Or perhaps you could look into RCDs that are especially manufactured for just this problem. Look up type F & B RCDs. F generally for single phase VFDs and the B and B+ models for three phase industrial machines.

Dont see the point in looking at RCDs that are not allowed for domestic installations.   Im not loosing much sleep over the EMI for a machine that  going to be run for a only a few minutes at a time every few months or so.  I would rather have a safe installation.

Chris

I will be the first to admit that I don't work with domestic situations, nor do I live in your jurisdiction, but I'm wondering why a type F RCDO isn't suitable for a domestic installation. I've had a very brief flick through AS/NZS 61009.1:2015 "Residual current operated circuitbreakers with integral overcurrent protection for household and similar uses (RCBOs)" and cannot find anything that precludes the use of a type F. AS/NZS 61009.1 is reproduced from IEC 61009-1 with some national modifications added.

[fubar.gr]

VFDs can also cause the opposite problem: preventing RCD from tripping.

This is due to the DC component they inject into the power lines. The current transformer in the RCD saturates and won't trip should a genuine ground leakage appear.

The solution to this is to use B type RCDs instead of the most common AC type. They are expensive though

[AllTheGearNoIdea]

Just a couple of comments.

The machine is in an outbuilding about 20m from the house. My house is a TN-C-S I have no earth conductor coming into my house. The earth wire going to my power sockets is actually is actually connected to the neutral where the power supply comes into my home. Now this is a crap system has lots of issues and is done to save the utility company a few bucks is copper wire back to the utility transformer.

Now my understanding of the regulations is that plug socket have to be RCB protected  in the U.K.

It’s doubly important that if you have a outhouse away from the main property that this supply  is RCB protected.

The supply to the outhouse is a spur from the plug sockets in the kitchen.

So it’s probably not practical or allowed within the regulation to have plug sockets in my outhouse that are not RCB protected. This rules out rewiring or putting in a dedicated supply as it would still have to be derived from the rcb.

A few post mention changing the rcb. I’m not and expert regarding these however for a domestic application the maximum trip current is 30mA the Tripping time is also defined. So although other RCB may be available they may not be useable for a domestic installation in the UK.

If your an electrician in the U.K.  perhaps you could advise

Many thanks

[capt bullshot]

That's a bad situation for you, alas.

From the technical side, it is clear that this setup won't work if you obeye the laws and regulations ...
Another clear thing is, neither law nor regulations can override laws of physics, but laws of physics won't even care about any law or regulation.
So my advice would be to follow law of physics to get your setup working, or to have a real safe (in the sense of laws and regulations) outhouse, since there won't be any dangerous moving parts there ...

BTW, the power distribution system you describe is quite the same here in Germany, there's no separate earth wire within the Low Voltage Distribution network.
And the yearly increased requirements from law and regulations bring up quite a lot of  here too.

So either you are forced to obeye these regulations (e.g. because you're working in a company and no one is willing to take responsibility) and do some  stuff to get your stuff up and running, or you just choose to take responsibilty for yourself and ignore parts of the regulations ...

[sibeen]

The machine is in an outbuilding about 20m from the house. My house is a TN-C-S I have no earth conductor coming into my house. The earth wire going to my power sockets is actually is actually connected to the neutral where the power supply comes into my home. Now this is a crap system has lots of issues and is done to save the utility company a few bucks is copper wire back to the utility transformer.

Just a quick note on that. It's not a crap system. It's a multiple Earth Neutral (MEN) design which is used all around the world. At the switchboard the neutral and the earth are combined. In the UK they use the plumbing piping or such to form the bond to earth. In Australia we are much more civilised and require that a dedicated earth stake be installed for each domestic premises

A few post mention changing the rcb. I’m not and expert regarding these however for a domestic application the maximum trip current is 30mA the Tripping time is also defined. So although other RCB may be available they may not be useable for a domestic installation in the UK.

This is why I mentioned the type F RCD in my original post. These are made for a domestic application with the type B and B+ for industrial/commercial. They are designed to tolerate the high frequency noise components that things like VFDs put out. They also comply with the 30 mA tripping and can be selected with different tripping current to suit the type of cable that you are feeding.

[Kjelt]

The machine is in an outbuilding about 20m from the house. My house is a TN-C-S I have no earth conductor coming into my house.

I am not uk qualified.
If allowed I would replace the powercables to your building. Starting from where the mains enter the main house you create a special power group with only a large safety fuse (20A or so) use very thick cable (6mm² or so) to cut the losses.
In the outbuilding this cable gets its own cabinet and create a new seperate ground (should be struck and tested). From that basis you can continue properly distributing power with seperate groups and rcb to each machine.

[sibeen]

kjelt, I was going to suggest that but I'm not sure what the rules are in the UK for that. In Australia you can do as you suggest, run a supply to a new board and form a new MEN point at the new distribution board. The RCBO would then be installed in the new switchboard to run the GPOs within the building.

[Kjelt]

In Holland they loosened the rules such that the owner is responsible after the main switch.
That means you got to stick to the rules or if fire breaks out the insurance company might not pay out but it gives lot of room to do what you need.

[AllTheGearNoIdea]

Hi Sibean

Thanks for the heads up on the F type RCBO I will ask an electrician for some advice on this.

Regarding the TN-C-S. I disagree on that it is a crap system with no avantages but saving a few bucks in copper. Its dangerous and has many flaws. I girl was killed in the UK only 1 week ago after touching an outside tap.  It looks like the property had a neuttal fault.  Im also a radio ham TN-C-S is a nightmare as you cannot easily ground outside equipment as you have potential fault cuurrent paths down Coax cables etc.

I think I need to investigate a bit further. I know a currently clamp wont be fast enough to catch the high speed transients etc but it will show a normal earth leakage fault and can discount this as the problem.  If it look to be purely just a transient issue there are measure that can be taken.

Chris 

[Elasia]

Hi Sibean

Thanks for the heads up on the F type RCBO I will ask an electrician for some advice on this.

Regarding the TN-C-S. I disagree on that it is a crap system with no avantages but saving a few bucks in copper. Its dangerous and has many flaws. I girl was killed in the UK only 1 week ago after touching an outside tap.  It looks like the property had a neuttal fault.  Im also a radio ham TN-C-S is a nightmare as you cannot easily ground outside equipment as you have potential fault cuurrent paths down Coax cables etc.

I think I need to investigate a bit further. I know a currently clamp wont be fast enough to catch the high speed transients etc but it will show a normal earth leakage fault and can discount this as the problem.  If it look to be purely just a transient issue there are measure that can be taken.

Chris

I use a I think its either 6 or 4 gauge ground spike at my main panel drove in deep into the ground... not having a spike ... ehhhhh

[AllTheGearNoIdea]

OK this is what I am suggesting for my isolation transformer wiring.  I know its not giving isolation to ground but is this safe to use ?

Note i would probably link the primary and scondary earths

[Elasia]

Why not just get an industrial rcd made for inverters with the higher mA and call it a day if your going that route?  Your just tossing money if you ask me.. you would just need a dedicated circuit to your main panel and avoid all your normal house wiring then wire your rcd in and the motor after it

Also I'd highly recommend a ground spike at your main panel with a proper ground bar... then that bar gets tied to your neutral in your panel usually with a high gauge bond wire.

[AllTheGearNoIdea]

Well I think if the above circuit works safely its my cheapest option.  Just bought a 3Kva isolation transformer for £98 and the 100mA RCB for secondry for   £18.    So In theory I get to run my machine safely without modifying the house wiring.   For very little money.

Chris

[AllTheGearNoIdea]

Wired up the transformer today all working 

[Kjelt]

Do test the rcb with a resistor leaking 35mA to gnd see if it trips.