Peculiar electrical problem

[john fletcher 1Participant @johnfletcher1]

A non computing friend called around yesterday, and during the course of conversation he told me about his mill and lathe, I haven't seen either. He said he had fitted both with inverters, and normally would run one machine or the other in the interest of safety, amongst other things. More recently he had attempted to run both at the same time and the RCD tripped, The friend didn't know much about the electric supply to his workshop, other than it is a ring main system and is fitted with an RCD. This seems a strange situation, any advice would be appreciated. John

[john fletcher 1Participant @johnfletcher1]

[Peter Cook 6Participant @petercook6]

Quick internet search found this

Residual current protection in variable-speed drives – CIRCUTOR

it's advertising, but suggests that the EMI filters in invertors feed small currents to earth in suppressing the high frequencies.

Sounds like individually the invertors don't have enough leakage to trip the RCD, but when on together the sum total is big enough to so do.

[SillyOldDufferModerator @sillyoldduffer]

The RCD "Residual Current Device" disconnects when it detects an imbalance between line and neutral currents. The imbalance means current is flowing to earth. To an electrical engineer used to managing power, earth current indicates an electrical fault. RCDs are quite sensitive, set on the assumption that the fault current is flowing through a human being, and because 30mA is dangerous to health, they disconnect quickly – a few tens of milliseconds.

When an RCD triggers, it's important to ensure the earth fault isn't genuine, perhaps caused by the failure ancient rubber insulation, damp, frayed cables, broken connections, or other wiring faults.

However, the problem can be innocent. Enter villain stage left. It's an electronics engineer! He has a legitimate need to bleed small currents to earth. All electronic devices in the home will be fitted with suppressor capacitors, arranged to short interfering high-frequency currents to earth. Unfortunately, suppressors also allow a tiny amount of 50Hz current to flow as well. As a modern home contains dozens of electronic devices, it's possible for them to collectively leak tens of milliamperes and take the RCD close to it's trigger point. Powerful electronics, like microwaves and VFDs leak more earth current than tiddlers like TV sets, and are therefore more likely to pop the RCD.

I suspect the friend's house has an existing high level of legitimate earth leakage, and switching on both VFDs takes his RCD over the top. Proper EMC filters leak more current than basic suppressors.

I thought some of the newer cheapo mains testers measured earth leakage, but I couldn't find one on Amazon. Perhaps I dreamt it! Looks like a more expensive clamp meter is needed. Positioned on the device's earth wire, it will show if the problem is local. Positioned at the consumer unit it will show total leakage. The most leaky devices can be identified by unplugging them.

A 'competent person' could do the tests by wiring in great-grandad's AVO. Not recommended – it's the sort of dodgy job that zaps professionals. Clamp meters are much safer!

Dave

 

 

Edited By SillyOldDuffer on 24/09/2023 12:42:39

[old martParticipant @oldmart]

I was concerned about earth leakage when fitting the VFD to the Tom Senior and noticed that the EMI filter can be disabled to reduce earth leakage. Fortunately this feature was not needed.

[noel shelleyParticipant @noelshelley55608]

As Dave (SOD) has said a competent person could put an ampmeter in the earth line but NOT a task for an ordinary Joe. One of my boxes of tricks will test the drop current and how many mil seconds but cheapo it was not. Noel.

[Robert Atkinson 2Participant @robertatkinson2]

This fault is as already been said almost certainly caused by leakage currents that result from the design and normal operation of the VFD and filers in particular. This higlights the fact that almost all VFDs are industrial devices and not intended to be connected to the domestic mains supply.
Disconnecting the filter is not the correct solution.
Ideally the workshop would have it's own dedicated feed with an isolator (main switch) and circuit breaker (no RCD) at the incoming supply conected to a consumer unit in the workshop with second isolator ,a RCBO for each machine and additional RCBOs for ring main and lighting.

Robert.

[Chris Pearson 1Participant @chrispearson1]

Posted by Peter Cook 6 on 24/09/2023 11:58:56:

Sounds like individually the invertors don't have enough leakage to trip the RCD, but when on together the sum total is big enough to so do.

Agreed.

Unfortunately, sufficiently sensitive clamp meters which could measure the current in the CPC are rather expensive.

And yes, you could put a AVO meter in series with the CPC, but there are risks attached!

[Robert Atkinson 2Participant @robertatkinson2]

You don't actually measure thr current in the CPC ("earth" conductor). This is becuse the leakage culd be finding a different path to ground. You do what an RCD does, measure the difference in current between the line and neutral. A clamp meterto do this has to go around the line and neutral.

Robert.

[not done it yetParticipant @notdoneityet]

It is almost certainly ‘the straw that breaks the camel’s back’ scenario.

It might be surprising but the allowable earth leakage is about 7W – about the same power as a typical lowish powered LED lamp!

However the trip could be occurring at as much as 14W (if each machine is providing enough leakage to ‘almost’ trip the RCD.

I might suspect the VFDs are cheaper examples (from China?), but am not qualified to make any definitive judgement on that score.

As I see it, leakage to earth is not actually going to be a danger – until the earth path resistance makes it preferable for the current to go through you! Of course, far better – every time – to trip than the alternative of possible electrocution.

I do wonder how many actually check the trip, on their main RCD, on a regular basis?

[Robert Atkinson 2Participant @robertatkinson2]

Strangely enough a "cheap chinese" VFD is, barring faults, likely to have less leakage current than a good quaility one. This is becase the cheap ones tend to leave out the EMC fiter capacitors that cause the leakage current.

[Mike PooleParticipant @mikepoole82104]

An friend was moving into another office in the factory and they were having trouble with getting more than 3 crt monitors powered up before the RCD tripped, I found that for some reason a 10mA RCD had been fitted rather than the typical 30mA device.
Mike

[Chris Pearson 1Participant @chrispearson1]

Posted by Robert Atkinson 2 on 24/09/2023 19:19:24:

You don't actually measure thr current in the CPC ("earth" conductor). This is becuse the leakage culd be finding a different path to ground. You do what an RCD does, measure the difference in current between the line and neutral. A clamp meterto do this has to go around the line and neutral.

You can put your clamp meter around line and neutral together, or CPC, but it doesn't make any difference.

You could even put two AVOs in the line and neutral and calculate the difference.

[Master of noneParticipant @masterofnone]

Posted by Chris Pearson 1 on 24/09/2023 20:58:47:

Posted by Robert Atkinson 2 on 24/09/2023 19:19:24:

You don't actually measure thr current in the CPC ("earth" conductor). This is becuse the leakage culd be finding a different path to ground. You do what an RCD does, measure the difference in current between the line and neutral. A clamp meterto do this has to go around the line and neutral.

You can put your clamp meter around line and neutral together, or CPC, but it doesn't make any difference.

You could even put two AVOs in the line and neutral and calculate the difference.

[Master of noneParticipant @masterofnone]

In discussions with Meggar, they recommended using the clamp meter around line and neutral together. A parallel path to earth through an extraneous conductive part (e.g. water pipe, gas pipe or structural steelwork) could reduce the current in the earthing conductor and give a misleading measurement of the difference between line and neutral currents. The clamp meter needs to cover the range 0-30mA.

[Robert Atkinson 2Participant @robertatkinson2]

Wot I said….

[Chris Pearson 1Participant @chrispearson1]

Posted by Master of none on 24/09/2023 23:01:37:

In discussions with Meggar, they recommended using the clamp meter around line and neutral together. A parallel path to earth through an extraneous conductive part (e.g. water pipe, gas pipe or structural steelwork) could reduce the current in the earthing conductor and give a misleading measurement of the difference between line and neutral currents. The clamp meter needs to cover the range 0-30mA.

Only: (1) if there are any extraneous CPs (which is seldom the case in modern homes); and (2) the device is attached to an ECP and the fault or leakage is to that.

[SillyOldDufferModerator @sillyoldduffer]

Posted by Robert Atkinson 2 on 25/09/2023 07:47:01:

Wot I said….

Yes, but I'm glad of all the clarifications. My understanding wasn't good enough! I assumed current in the earth cable should be measured, now I know better.

Dave

[Martin Shaw 1Participant @martinshaw1]

As a matter of interest since I have a clamp on ammeter I thought to measure the earth leakage current in my house, by no means a scientific and accurate measurement and more of a look and see. The house is nearly 100 years old and of typical interwar construction so not much leakage via extraneous conductive parts, plastic water pipe and bonded gas supply and wiring that is all fairly modern which has all been either replaced or tested satisfactorily for leakage.

I turned as much as possible on, milling machine and lathe both with DC electronic motors, single phase 2.5hp air compressor, several electronic fluorescent and LED lights, hifi, fridge, about the only high current device not on was the washing machine which may well prove to be the worst case.

The ammeter reads to the nearest 10mA and the displayed leakage was exactly 0, both on the main earthing conductor and the phase and neutral conductors, so less than 5mA total. Entirely possible that I have good quality equipment that doesn't leak much to earth but I wonder on this basis how much domestic plant as opposed to poor wiring or miswired plugs actually causes RCDs to trip.

Regards

Martin

[Chris Pearson 1Participant @chrispearson1]

Posted by SillyOldDuffer on 25/09/2023 12:28:45:

Posted by Robert Atkinson 2 on 25/09/2023 07:47:01:

Wot I said….

Yes, but I'm glad of all the clarifications. My understanding wasn't good enough! I assumed current in the earth cable should be measured, now I know better.

Hang on – this is all getting a bit muddled.

If you think that there is high earth leakage, you need to measure or calculate it, which requires access to the conductors individually. I see no advantage in testing L & N together as opposed to the CPC. If the device is bonded and you test L & N together, you do not know which route any missing electrons are taking.

[Chris Pearson 1Participant @chrispearson1]

Posted by Martin Shaw 1 on 25/09/2023 15:50:03:

The ammeter reads to the nearest 10mA and the displayed leakage was exactly 0, both on the main earthing conductor and the phase and neutral conductors, so less than 5mA total. Entirely possible that I have good quality equipment that doesn't leak much to earth but I wonder on this basis how much domestic plant as opposed to poor wiring or miswired plugs actually causes RCDs to trip.

I am afraid that it is not quite as simple as that.

What type of earthing do you have please: TN-S, TN-C-S, or TT?

(FWIW, the current in my earthing conductor when I started writing this was 0.17 A, but Mrs P. has just put on the tumble drier and it has gone up to 3.5 A.)

[john fletcher 1Participant @johnfletcher1]

Back again, many thanks chaps, plenty to read and ideas to act upon. I've print out the text and will give to my friend, I'll come back some time later to let you all know the result. John

[Robert Atkinson 2Participant @robertatkinson2]

Posted by Chris Pearson 1 on 25/09/2023 12:03:24:

Posted by Master of none on 24/09/2023 23:01:37:

In discussions with Meggar, they recommended using the clamp meter around line and neutral together. A parallel path to earth through an extraneous conductive part (e.g. water pipe, gas pipe or structural steelwork) could reduce the current in the earthing conductor and give a misleading measurement of the difference between line and neutral currents. The clamp meter needs to cover the range 0-30mA.

Only: (1) if there are any extraneous CPs (which is seldom the case in modern homes); and (2) the device is attached to an ECP and the fault or leakage is to that.

It does not have to be leakage to a conductive (metal) part extraneous or otherwise. An old junnction box screwed to wood and exposed to water will cause leakage current. The RCD senses line / neutral differential current so that is what you need to measure when testing.
A mill bolted to a even slightly damp concrete floor would provide a parallel path to earth.

[SillyOldDufferModerator @sillyoldduffer]

Posted by Chris Pearson 1 on 25/09/2023 16:57:16:

Posted by SillyOldDuffer on 25/09/2023 12:28:45:

Posted by Robert Atkinson 2 on 25/09/2023 07:47:01:

Wot I said….

Yes, but I'm glad of all the clarifications. My understanding wasn't good enough! I assumed current in the earth cable should be measured, now I know better.

Hang on – this is all getting a bit muddled.

If you think that there is high earth leakage, you need to measure or calculate it, which requires access to the conductors individually. I see no advantage in testing L & N together as opposed to the CPC. If the device is bonded and you test L & N together, you do not know which route any missing electrons are taking.

Always happy to be educated!

In this Fluke diagram showing how to measure AC current, it's explained you can't put the clamp over Live and Neutral:

This I understand to be because the AC current flows one way in the Line wire and in the opposite direction in Neutral. The two generate magnetic fields that cancel out, so the clamp meter reads zero. Highly misleading.

However Live leaking to earth is a different case. The current flowing in Live is not equal to the Neutral current, so the magnetic fields don't quite cancel. A sensitive clamp meter will measure the difference, which is equal to the earth leak current.

Chris has me worried though: if the clamp is put over LN and E and the leak is due to a suppressor, the clamp won't detect it because the 3 magnet fields balance. It will detect a leak due to faulty insulation and a damp floor, because then not all the leak current goes back along the earth wire. Is that right?

Can anyone design a circuit that would reliably show how much current a device was leaking under all conditions?

I don't believe the debate alters the advice given to John.

Dave

[Chris Pearson 1Participant @chrispearson1]

Posted by Robert Atkinson 2 on 25/09/2023 17:55:43:

A mill bolted to an even slightly damp concrete floor would provide a parallel path to earth.

Yes, but a really good earth rod in damp soil might manage 20 Ohms and the same at the transformer. By contrast, the earth path through all the green and yellow might be, say 0.5 Ohm. So assuming that any earth leakage is to the bed or case of the machine, the vast majority (99%+) will go down the green and yellow.

The insulation of your junction box should be sound. Around one megaohm is about the minimum acceptable resistance, but a new machine in good order should achieve one gigaohm.

[Author]

Posts