How to test RCD and earthing of a temporary supply.

[epsilonGreedy]

My static caravan earthing does not follow best practice as posted by @ProDave

 

I do not know if the safety performance of an RCD is dependent on a reliable earth so how can I test my set up without putting myself in danger? 

[ProDave]

To be honest, get an electrician to come and test it. You won't have the test equipment needed.

[Jeremy Harris]

As @ProDave says, without proper test gear, and the knowledge of how to use it, it's not something you can DIY.  Apart from testing the installation, it must also be inspected to ensure that it physically complies with the regs, which means ensuring that the earth electrode is properly installed, that the CPC is of the correct gauge and properly terminated, that the terminal at the top of the electrode is protected by an appropriate housing, that the CPC is protected in conduit from that housing to the earth block, etc, etc.

 

It's not a big job to do, the actual inspection and testing shouldn't take more than half an hour at most, if all has been installed as it should be.

[Onoff]

Wire in a 230V old fashioned light bulb with one side on live and the other on the rod. The better the earth the brighter the bulb! 

 

I can't condone this as a test method btw...

[Carrerahill]

1 hour ago, epsilonGreedy said:

My static caravan earthing does not follow best practice as posted by @ProDave

 

I do not know if the safety performance of an RCD is dependent on a reliable earth so how can I test my set up without putting myself in danger? 

 

Basically to test an RCD you need to "leak" some current in a safe manner to ground (i.e. it's going AWOL from the circuit - this could be via a person or animal hence it will trip) this creates an imbalance because current flows one way, but not the other, so it is "lost". A tester can set the "leak" current to a predetermined level to test the responsiveness of an RCD, so a 30mA RCD should trip out within 40ms at 150mA. 

 

There should be a test switch on your RCD - this creates a small leak locally and will trip it - if this works, and your sockets work, then it will work as RCD's will work regardless of a good earth - in fact, that is sort of the point of them is to detect remote leaks to earth, i.e. not via the earth cable because a remote leak suggests it is going via someone or something direct to ground - dangerous - wear wellies! Obviously it will also protect against leaks to the ground cable.

 

I had an RCD constantly trip out on a garage circuit, it was because because the circuit, in a low lying electrical wiring connector in a freezer was damp, it was leaking current to earth via the damp. So really all you need is a damp extension cable lying in a puddle and an RCD would trip. 

 

Another way to trip them is to use an RCD socket, like the ones used for lawnmowers and things, or even use something like a lawnmower or external extension lead with one built in, hit the test button and it should take the main RCD with it at the distro board etc. 

 

 

Edited March 19, 2019 by Carrerahill

[Onoff]

6 hours ago, Carrerahill said:

Another way to trip them is to use an RCD socket, like the ones used for lawnmowers and things, or even use something like a lawnmower or external extension lead with one built in, hit the test button and it should take the main RCD with it at the distro board etc. 

 

Not necessarily. If both 30mA then it's likely one or the other might go if selection (formerly discrimination) rules haven't been considered. 

[Jeremy Harris]

7 hours ago, Carrerahill said:

 

Basically to test an RCD you need to "leak" some current in a safe manner to ground (i.e. it's going AWOL from the circuit - this could be via a person or animal hence it will trip) this creates an imbalance because current flows one way, but not the other, so it is "lost". A tester can set the "leak" current to a predetermined level to test the responsiveness of an RCD, so a 30mA RCD should trip out within 40ms at 150mA. 

 

There should be a test switch on your RCD - this creates a small leak locally and will trip it - if this works, and your sockets work, then it will work as RCD's will work regardless of a good earth - in fact, that is sort of the point of them is to detect remote leaks to earth, i.e. not via the earth cable because a remote leak suggests it is going via someone or something direct to ground - dangerous - wear wellies! Obviously it will also protect against leaks to the ground cable.

 

I had an RCD constantly trip out on a garage circuit, it was because because the circuit, in a low lying electrical wiring connector in a freezer was damp, it was leaking current to earth via the damp. So really all you need is a damp extension cable lying in a puddle and an RCD would trip. 

 

Another way to trip them is to use an RCD socket, like the ones used for lawnmowers and things, or even use something like a lawnmower or external extension lead with one built in, hit the test button and it should take the main RCD with it at the distro board etc. 

 

 

 

FWIW, to test Ra, a multitester will usually use the same function as used for testing Ze, and that pulses a fairly high current (around 20 A) from line to the earth electrode CPC in order to measure the resistance.  IIRC mine uses a pulse of current of about 40ms duration for this test.  Testing the trip current and time of the RCD is a separate test, only done once the electrode has tested out OK (Ra lower than 200Ω, ideally lower than about 100Ω).

 

A 30 mA RCD only needs an Ra of around 2,333Ω  to trip, whilst staying within the maximum safe voltage of 70 V on the CPC from the electrode, which is way too high a resistance to be acceptable for an earth electrode.  For this reason alone it's extremely unwise to rely on an RCD test as a way to guess that the earth electrode is installed correctly and has a low enough Ra.

[dpmiller]

I acquired a pukka (but chinesium) tester to do Ra on my recently-installed spike as my MFT doesn't have that capability.  It's a 50V test. I was well pleased with the sub-70ohm value achieved and glad not to find rock as the spike went in, more to the point...

It was interesting just how little effect the position of the test electrodes made. I tried them in a number of different substrates and positions accross the rear of the house, between it and the pole/transformer our supply will come from. Hardcore, soil, puddle, every test came back between 65 and 68 ohms. Pretty repeatable I thought.

[Jeremy Harris]

I think repeatabilty is more important than the absolute value, as that probably gives a good indication of stability, so, with luck, the value will be consistent through the year.  Although the blue book suggests 200Ω as the upper limit, I believe that one or two of the Part P approval bodies set a maximum acceptable limit of 100Ω, on the basis that this provides assurance that Ra won't increase above 200Ω through seasonal variation.

 

I've still got an old 4 terminal tester, and although that's a tedious way to test an earth electrode, and I haven't used it in a long time.  That does have the advantage that it's more likely to detect patchy ground conductivity issues, though.

[MJNewton]

On 19/03/2019 at 23:46, Carrerahill said:

 

Another way to trip them is to use an RCD socket, like the ones used for lawnmowers and things, or even use something like a lawnmower or external extension lead with one built in, hit the test button and it should take the main RCD with it at the distro board etc. 

 

 

That won't happen as the upstream RCD doesn't 'see' the imbalance created by the downstream RCD test button. All the test button does is shunt a small current between post-coil Live and pre-coil Neutral and the effect of this is therefore only ever seen locally. It doesn't leak current to Earth like a 'real' RCD tester does.

Edited March 28, 2019 by MJNewton

[Carrerahill]

13 hours ago, MJNewton said:

 

That won't happen as the upstream RCD doesn't 'see' the imbalance created by the downstream RCD test button. All the test button does is shunt a small current between post-coil Live and pre-coil Neutral and the effect of this is therefore only ever seen locally. It doesn't leak current to Earth like a 'real' RCD tester does.

I didn't know that is how the plug versions work, DIN mounted RCD's such as in our DB's use two methods, either a post coil live and neutral but the test button operates a small additional winding via a resistor on the core opposite the relay winding, thus this creates an imbalance. Merlin Gerin RCD's (the rep used to come in and offer to have designed all our panels on projects and would talk about specs of their kit) used to have a test button with a small resistor to the earth from post coil live, it would dump 20-25mA - whatever it was to earth. They were famed for claiming this made their RCD more robust as it was not relying on a secondary winding to create the imbalance but rather the real circuit was going to detect the imbalance. This also proved the coils were not degraded.

 

Anyway it appears now there are potentially 3 circuit options! I only suggested it as when my wife inadvertently presses the test on the lawnmower it would trip the outside power RCD on the DB and I was always under the impression the cheapo plug RCD's were the resistor to earth option as they are cheaper to make. 

 

 

[Jeremy Harris]

FWIW, my home made test plug (used to do "quick and dirty" checks when I can't be arsed to get the multitester out) just has a resistor to earth, switched via a push button, to create around 30 mA of leakage.  It's a pretty foolproof way of testing a low current RCD, IMHO.  Good enough for an "is it still working?" test, anyway.  If it fails then it's time to do a proper test.

 

Interesting to see that the chap from SSE who came around to change our meter "tested" the installation afterwards with what looked like one of the cheapo eBay test plugs, the ones with three LEDs on the front.

 

I've often wondered what the failure rate is on RCDs and RCBOs now, as I've had two DP RCBOs fail within 5 years.  Both failed the same way, leaving the circuit on with an earth fault present, which is the opposite to the way I remember older RCDs tending to fail (my experience of older TT installation RCCB/RCD failures was them often failing to latch back on after they'd tripped).  

 

 

[Onoff]

34 minutes ago, Carrerahill said:

I didn't know that is how the plug versions work, DIN mounted RCD's such as in our DB's use two methods, either a post coil live and neutral but the test button operates a small additional winding via a resistor on the core opposite the relay winding, thus this creates an imbalance. Merlin Gerin RCD's (the rep used to come in and offer to have designed all our panels on projects and would talk about specs of their kit) used to have a test button with a small resistor to the earth from post coil live, it would dump 20-25mA - whatever it was to earth. They were famed for claiming this made their RCD more robust as it was not relying on a secondary winding to create the imbalance but rather the real circuit was going to detect the imbalance. This also proved the coils were not degraded.

 

Anyway it appears now there are potentially 3 circuit options! I only suggested it as when my wife inadvertently presses the test on the lawnmower it would trip the outside power RCD on the DB and I was always under the impression the cheapo plug RCD's were the resistor to earth option as they are cheaper to make. 

 

 

 

Can't beat a Multi 9 "si" for problem free protection especially with vsds.

[ProDave]

23 minutes ago, JSHarris said:

Interesting to see that the chap from SSE who came around to change our meter "tested" the installation afterwards with what looked like one of the cheapo eBay test plugs, the ones with three LEDs on the front.

 

My opinion of SSE Linesman testing techniques is VERY low following an incident where a care home was connected to a 3 phase generator when the power lines were down after a storm.  I got called to the care home as stuff was going pop all over the place,  I quickly identified incorrect wiring from the generator putting 400V on one phase so I shut the genny down and called SSE.  When he got there, he started the genny, waved a volt stick about and said "what's the problem guv"?

[Carrerahill]

1 hour ago, Onoff said:

 

Can't beat a Multi 9 "si" for problem free protection especially with vsds.

Rolls Royce of switchgear!

 

[Carrerahill]

1 hour ago, JSHarris said:

FWIW, my home made test plug (used to do "quick and dirty" checks when I can't be arsed to get the multitester out) just has a resistor to earth, switched via a push button, to create around 30 mA of leakage.  It's a pretty foolproof way of testing a low current RCD, IMHO.  Good enough for an "is it still working?" test, anyway.  If it fails then it's time to do a proper test.

 

Interesting to see that the chap from SSE who came around to change our meter "tested" the installation afterwards with what looked like one of the cheapo eBay test plugs, the ones with three LEDs on the front.

 

I've often wondered what the failure rate is on RCDs and RCBOs now, as I've had two DP RCBOs fail within 5 years.  Both failed the same way, leaving the circuit on with an earth fault present, which is the opposite to the way I remember older RCDs tending to fail (my experience of older TT installation RCCB/RCD failures was them often failing to latch back on after they'd tripped).  

 

 

Jeremy, that is how the Merlin Gerin ones worked, albeit built into the RCD, and I thought many others did too now. Down and dirty testing is often pretty good! 

 

The thing with RCD's is that they are essentially a relay, with the coil current coming from a fault imbalance, it is like a light, one stage rifle trigger, it's a hairspring and a small current should create enough of a field to trip the latching relay. If they are not used for long periods and are exposed to damp, i.e. a garage install, damp basement, it is entirely possible the mechanics of it can gum up. Now this is from a theoretical report, but from what you have said it sounds like that could be the mode of failure.

 

It is your experience which makes me generally dislike 18th edition boards.

 

 

 

 

[Carrerahill]

1 hour ago, ProDave said:

My opinion of SSE Linesman testing techniques is VERY low following an incident where a care home was connected to a 3 phase generator when the power lines were down after a storm.  I got called to the care home as stuff was going pop all over the place,  I quickly identified incorrect wiring from the generator putting 400V on one phase so I shut the genny down and called SSE.  When he got there, he started the genny, waved a volt stick about and said "what's the problem guv"?

Let me guess, he ballsed up new colours to an older red, blue and yellow installation?

 

 

[MJNewton]

28 minutes ago, Carrerahill said:

Jeremy, that is how the Merlin Gerin ones worked, albeit built into the RCD, and I thought many others did too now. Down and dirty testing is often pretty good!

 

It is definitely the best method for testing the RCD protection of a single RCD or a complete circuit (ideally with timing measurement of course) but the only way to test a downstream RCD such as the plug-in types without indiscriminate tripping of an upstream RCD is to force an inbalance through that one coil and not them both as would be the case if leaking to Earth.

 

Incidentally, my neon socket tester has a built-in RCD test function with variable leakage currents and whilst it doesn't perform trip timings (or detect neutral-earth reversal) it is nevertheless a handy cheap tool to have. I don't see many of this type around though for some reason.

 

 

(Apologies for the incorrect rotation - the original is landscape though)

Edited March 29, 2019 by MJNewton

[Jeremy Harris]

5 minutes ago, MJNewton said:

 

It is definitely the best method for testing the RCD protection of a single RCD or a complete circuit (ideally with timing measurement of course) but the only way to test a downstream RCD such as the plug-in types without indiscriminate tripping of an upstream RCD is to force an inbalance through that one coil and not them both as would be the case if leaking to Earth.

 

Incidentally, my neon plug tester has a built-in RCD test function with variable leakage currents and whilst it doesn't perform trip timings (or detect neutral-earth reversal) it is nevertheless a handy cheap tool to have. I don't see this type much around though for some reason.

 

 

(Apologies for the incorrect rotation - the original is landscape though)

 

That's a posh version of my tatty home made one!  Mine just has a single resistor that gives around 30 mA or so, as we've not lived anywhere with a TT installation for years now.

[MJNewton]

I was trying to see if I could find mine anywhere as people often ask where it was from but it seems it has long since been discontinued. I did just stumble across this one though (a DI-LOG DL1092 in case the link ever rots) which for £17.99 might be worthwhile for those not wanting to spend too much or go down the DIY route.

 

Edited March 29, 2019 by MJNewton

[ProDave]

For not a great deal of money you can get an old, but perfectly good tester, e.g https://www.ebay.co.uk/itm/Newlec-NL175N-RCD-Tester-mint-condition/123705826588?hash=item1ccd711d1c:g:tOkAAOSwCqZcYJyV

[ProDave]

40 minutes ago, Carrerahill said:

Let me guess, he ballsed up new colours to an older red, blue and yellow installation?

 

 

No it was worse than that.  A pre made cable plugged into a socket at the generator.  It had black, red, yellow and blue cables exiting the end.  Unfortunately the black was NOT neutral.  So either the plug was connected wrong or the socket in the generator was connected wrong.

 

The point was they connected it (it can't possibly be wrong can it) and then just waved a volt stick about and declared it "good"  I had to show them with a proper volt meter what the problem was, and even then they just said "oh"  and didn't seem that concerned.

 

They paid a lot of money to the care home for new computers, fire alarm, cctv etc all cooked by their incompetence.

[Temp]

On ‎19‎/‎03‎/‎2019 at 23:46, Carrerahill said:

There should be a test switch on your RCD - this creates a small leak locally and will trip it - if this works, and your sockets work, then it will work as RCD's will work regardless of a good earth - in fact, that is sort of the point of them is to detect remote leaks to earth, i.e. not via the earth cable because a remote leak suggests it is going via someone or something direct to ground - dangerous - wear wellies! Obviously it will also protect against leaks to the ground cable.

 

Ideally they should trip when they detect a short between live and earth not wait until the remote leak is via a human!

 

[Jeremy Harris]

1 hour ago, ProDave said:

For not a great deal of money you can get an old, but perfectly good tester, e.g https://www.ebay.co.uk/itm/Newlec-NL175N-RCD-Tester-mint-condition/123705826588?hash=item1ccd711d1c:g:tOkAAOSwCqZcYJyV

 

A while ago I saw the same model of multitester as I have going on eBay for £70 or best offer.  It's old, but perfectly OK, and will do just about every test needed.  Not as easy to use as some of the newer models (the test button seems designed for someone with 8" long thumbs) but for occasional use it's fine (I'd not want to use it every day to earn a living, though).  Mine's out of cal now, but I have a couple of big, known value, resistors that I can use to check the low resistance bit and the insulation resistance, plus I can cross check the voltage readings against a couple of other multimeters; it always seems to read the same from one year to the next.  Errors are more likely to be from leads not being nulled out, anyway.  Every time I've scratched my head over an odd reading it's been because I haven't nulled the leads properly (something made more awkward because the ergonomics of the tester are so dire - you really need three hands to use it at times).

[Carrerahill]

51 minutes ago, Temp said:

 

Ideally they should trip when they detect a short between live and earth not wait until the remote leak is via a human!

 

Ah, be careful, you are getting mixed up with MCB's - This is an RCD - a fault/danger can exist that will NOT cause any tripping until something, i.e. a human with wet hands touches an exposed live on say a grazed electrical cord. 

 

Short circuit protection is quite different.