Low IR on new radial circuit

[J D Hogg]

Hi All,

I have a very mystifying problem on a very simple circuit.

It is a new radial circuit for a garage power supply, it's wired in 2.5 T&E 16A MCB 30mA RCD. The cable goes directly from the dis board up the wall in trunking to the roof space where it is clipped to the joists before coming down the opposite wall under plastic capping which has been drylined over, feeds two double sockets and one single, then goes back up the wall into the loft before it comes down the adjacent wall in two drops for two double sockets and finally back up the wall into the centre of the ceiling for a garage door motor.

I did the first fix about 6 months ago and the second fix has just been done.

The live to earth IR is not brilliant at about 120 M Ohm but obviously OK, the problem is when I screw any of the sockets into the back boxes the IR goes down to 2 M Ohm. The boxes are just screwed to thermalite blocks and they are not earthed. I have tried different sockets but no difference. If I push the sockets into the boxes but don't screw them in it's OK.

I have carried out end to end tests and all OK, As far as I can I have inspected the cables and cannot see any damage.

It doesn't make sense to me; any thoughts.

Cheers

 

[buzzlightyear]

possibles damp some where on the cable.

 

[pc1966]

When you say "when I screw any of the sockets into the back boxes the IR goes down to 2 M Ohm" have you really tried one at a time?

You say you have tried other sockets, so that would appear to rule that out as a problem, assuming they are not all from the same batch of course.

2 Meg is odd, not a short but very low for any usable insulation and not something to be content with (as you are very ware). I guess the CPC is good (as you seem to be properly testing) but can you try putting a connection between E and the back boxes to see if it is the connection from being screwed in, or (if not) the physical pressure that is a factor.

 

[hasel5]

are they USB sockets?

 

[littlespark]

Got to remove every accessory and test each length of cable.... wayward dry lining screw?

 

[happysteve]

When you say "live to earth IR" can I just check you actually mean "line and neutral tied together, to earth" and you're not just testing line to earth (with the N still in the N bar)?

Because if the latter, then you're measuring the resistance between the wall and your actual earth, which might well be 2MΩ, and isn't a fault.

If the former, then I'll have to ponder some more.

 

[static zap]

If not USB and are metal back boxes .. This makes sense.

Poor 120M is raising damp patch in thermalite live ?
(some damage / damp un earthed connection)..clipped ?
Measure L-> screws in termalite wall .... Thermalite has floated to L not earth !!
.... or am I taking deductive reasoning too far ..

 

[westward10]

USB sockets should be irrelevant as they have isolating transformers like shaver sockets.

 

[freddo]

.... or am I taking deductive reasoning too far ..

Perhaps, the only other way of thinking is the new circuit has a fault to earth, but the CPC isn't continuous for the entire circuit, perhaps a damaged cable. When the sockets connect to the wall boxes that are earthed through the damp wall the true circuit insulation resistance is shown. But I assume r1+r2 was already carried out and a pass?

 

[static zap]

possibles damp some where on the cable.

If Buzz is right with damp /cold ... all the readings are about to change .. next time measured !
... have you had a plasterer in ? ...

 

[happysteve]

I had a fault like this in the past. Here is what I worked out was going on:



There was an old aerial cable buried in the paster near one of the sockets I was testing. When the socket was screwed into the back box, I measured 0.3MΩ (L+N) to E at the socket (and at all other sockets); when socket was disconnected >999MΩ; measured at the DB, 1.6kΩ (N to E) due to the actual fault. There was only one fault - the 0.3MΩ was just the resistance between the back box and the connection to E (probably via the aerial earth).

 

[static zap]

USB sockets should be irrelevant as they have isolating transformers like shaver sockets.

For completeness what voltage was IR test done at 250 / 500 or other ?

 

[Lucien Nunes]

The most likely scenario does seem to be that the line conductor is faulted to the blockwork e.g. by a nail, but the blockwork is not in contact with the CPC e.g. via real earth.

Suppose the the fault gives 2MΩ from line to blockwork but the wall is standing on a good dryish membrane or damp proof course giving 120MΩ between the blockwork and real earth. You'll read the 120MΩ to the CPC when it's in the earth bar of the CU, if that's earthed to real earth and not in good contact with the block. But when a socket is screwed home and the CPC is connected to the back box, the path is now from line through the fault and back to the CPC not through the membrane, so it reads 2MΩ.

 

[Mike Johnson]

Unearthed metal back box's, when you screw in the sockets they try to earth through the blockwork?

 

[pc1966]

Unearthed metal back box's, when you screw in the sockets they try to earth through the blockwork?

If the main CPC is OK I would expect it to go to low-R if it were simply the box coming in contact, but to be checked.
After all it is either related to the box being earthed, or (more likely?) they physical force of being crewed in doing something to the cable.

 

[Lucien Nunes]

or (more likely?) they physical force of being screwed in doing something to the cable.

However, the OP says when any of the sockets are screwed in, the same thing happens, so I'd infer that the change in IR is caused by connecting the CPC via the box to the mass of the wall.

 

[pc1966]

However, the OP says when any of the sockets are screwed in, the same thing happens, so I'd infer that the change in IR is caused by connecting the CPC via the box to the mass of the wall.

Good point, but usually you would be testing to the CPC (already connected to earth)? I.e. box to socket is already earthed.

 

[Lucien Nunes]

Not if the wall is standing on a membrane that provides higher insulation that the blockwork of the wall itself.

My troubleshooting method would be first to disconnect all conductors at the CU and prove that I get the 2MΩ between line and at least one box. Then, split the line at each point and test to the box to find the faulty section.

 

[J D Hogg]

Thanks for all of your replies. I will go back there this week to further investigate but I am thinking the most likely cause is a nail through the cable.

Just to clarify a few points raised; the low IR is between line and CPC with nothing connected in the CU. Some of the sockets are metal faceplate and some are plastic. I have tried screwing each in individually and I get the same result and it is as soon as the screw is screwed into the backbox so nothing to do with pressure on the cables. R1R2 is 0.34 Ohms. Using 500v for the testing. The wall is quite new so is probably still damp.
Cheers

 

[westward10]

The cpc should be connected to earth whilst testing.