UK High PSCC and strange readings - help much appreciated

[CalB4676]

Hey hows it going,

So was doing two different EICR the other week. Two different locations, same issue and I’m not sure what the best course of action is. Both constructed in 2016, 100% new. Hager metal dbs. TN-C-S (PME)

So for example on one:

Original test sheet Zdb - 0.21ohms, 1.18 kA
(No Ze on teat sheet)

My tests
2 lead test

Ze at Db (earth disconnected) - 0.06ohms, 3.62

Zdb (earth connected) - 0.06, 3.73kA

Now the problem

L-N - 0.01, 20kA

So that would give me a final recordable value of 20kA of a prospective fault current.

Anybody have any ideas on why the reading could be so massive and how to lower it? As it’s not suitable for the 6kA rated MCB’s.

Was looking at regulation 435.5.1, looks like the MCB braking capacity can be ignored if protected by an another protective device that is suitable. The supply fuse is at least 1361 type 2 that has a 33kA breaking. So it might be ok if Im reading it right.

Im using a megger mft 1721,just calibrated three weeks ago. cant null the leads on the these tests, possibly out of range of the tester value. Have not had any issues with other properties.

Any help or info is much appreciated.

Thanks

 

[westward10]

For TN-C-S your L to N should be the same or very similar to L to E.

 

[HappyHippyDad]

Can you actually have an impedance reading between L-N as low as 0.01ohms?

Once we step outside of a residentially building my understanding drops hugely. However, if the reading is 0.01ohms then doesn't that mean the transformer would have to be (approx) 6.88m away. Can you have a transformer that close?

The 6.88m is determined by assuming the tails are 25mm and copper, and then using the known resistance of 25mm copper conductors (i.e 0.727mohms per meter).

edit... the maths is obviously wrong as it's not 25mm tails to the transformer! Sorry.

 

[brianmoooore]

The transformer would have to be (approx) 6.88m away. Can you have a transformer that close?

Transformer feeding my house is that sort of distance from the service head.

 

[CalB4676]

My first thought was that my tester must be reading wrong, checked it against over work and reading fine. Then thought they may of upgraded the local transformer at some point hence the different value from the original test cert however its not like the cable length would of changed. I was expecting the result to be much and such, consumer side L(25mm) to earth (16mm) / L(25mm) to N)(25mm) so some small difference. I am planning back Monday for the odd 1 or 2 remedial so was think I could try using the wander lead and deduct the resistance to make sure that the tester isn't just maxing out. If not, not sure what the course of action is best to lower it to something manageable.

 

[HappyHippyDad]

Transformer feeding my house is that sort of distance from the service head.

Blimey! I wonder what your PSC is?

 

[westward10]

My first thought was that my tester must be reading wrong, checked it against over work and reading fine. Then thought they may of upgraded the local transformer at some point hence the different value from the original test cert however its not like the cable length would of changed. I was expecting the result to be much and such, consumer side L(25mm) to earth (16mm) / L(25mm) to N)(25mm) so some small difference. I am planning back Monday for the odd 1 or 2 remedial so was think I could try using the wander lead and deduct the resistance to make sure that the tester isn't just maxing out. If not, not sure what the course of action is best to lower it to something manageable.

Retest it because for TN-C-S the readings should be the same.

 

[timhoward]

I’d pick up a check box if you don’t have one. They can give a quick sanity check for any tester. One function is to add a known resistance to check the loop test.
The older ones that need a non-rcd protected socket as a source are usually cheap on the 2nd market.

I have a little experience with megger calibration and the number of reference values entered during loop calibration is optional requiring a minimum of 2, so if someone was in an hurry it can end up working better for certain ranges than others.

At the moment my suspicion is with the tester.

 

[UNG]

I’d pick up a check box if you don’t have one. They can give a quick sanity check for any tester. One function is to add a known resistance to check the loop test.
The older ones that need a non-rcd protected socket as a source are usually cheap on the 2nd market.

The check box is a good call even if like the OP's meter it has only been calibrated a few weeks before assuming the meter must be correct in these circumstances can be misleading, again how many ever look through the calibration sheet for minor issues that may skew certain test results out on site, a calibration does not really have a pass or fail so some tests could produce erroneous results

At the moment my suspicion is with the tester.

Or the test lead set or even both
I have had it in the past were the lead set had an intermittant problem that came down to the soldering on one of the 4mm plugs, some days it worked without a problem and other days it gave problems. I know someone who had a problem with a new meter giving odd readings turned out to be the 13A plug on the test that had a couple of poor terminations

It would be interesting to know the results of a 3 lead test when comparing the results of past inspections


Another thought are the 2 properties in question looped supplies and the other properties that have been tested aren't, in the past the nextdoor test for me has thrown up issues the DNO had to sort

 

[Teddy Lop]

Re: using the service fuse as backup is acceptable , I queried it for a 10k scenario with the niceic tech line a few years ago.

But 20k, wowsers! Don't have an old backup machine lying around, borrow one etc?

 

[timhoward]

Or the test lead set or even both

Good shout on the test leads.

It would be interesting to know the results of a 3 lead test when comparing the results of past inspections

I'd assumed the OP is (hopefully) doing a "2 - Hi" high current test. There isn't a "3 - Hi" and I don't think he should confuse matters with low current tests that are inherently flawed anyway.

(One of my dislikes about the Meggers 17 series is that the significant differences between lo and hi tests is under-emphasised by the user interface. Earlier models has a No Trip setting on the mode dial, suggesting (correctly) it's a completely different test method.)

Don't have an old backup machine lying around, borrow one etc?

Another thing well worth having. I've long held the view that any self respecting sparks collector of junk should keep an old tester or two for muddy farmyards, fault finding, and exactly this scenario to give a 2nd opinion. It needn't be expensive or professionally calibrated if you have your own check box.

 

[Marvo]

As per Mr Westy, if it's a PME'd TN C S arrangement the earth fault loop should test identical or almost identical to the short circuit loop. The fact your results are so far apart means either your test equipment is faulty, your test technique is flawed or the E-N bond link in the service head is not effective. I'd retest using a different tester first then go from there.

 

[westward10]

This is definitely being over thought you are essentially testing the same thing where TN-C-S is concerned, L to E or L to N they are the same.

 

[davesparks]

Can you actually have an impedance reading between L-N as low as 0.01ohms?

The short answer is yes, and bigger circuit breakers can have a maximum Zs less than 0.01ohms. high resolution loop testers give readings to an additional decimal place.

But there is also an issue with loop testing when you're close to the supply transformer. I'm not 100% on what is actually happening but basically the inductance of the transformer can cause misleading readings when very close to the transformer.

Can you have a transformer that close?

Yes, certainly in bigger installations anyway. It's not unusual for a transformer to be in a room next door to your switch room, sometimes with just a few metres of busbar between the transformer and the main switch.

Or for a package substation on an industrial estate the LV fuse cabinet can be pretty much bolted to the side of the transformer.

 

[Marvo]

I'm not sure if manual calculation is an option in the UK regs but usually if I get a really low Ze measurement and there's questions about the fault current ratings of the switchgear I'd just manually calculate the PEFC / PSCC by determining the supply cable CSA and measuring or estimating (depending on access) the supply cable length. Obviously it's more work and sometimes it's not even possible but almost always the calculated impedance values are considerably higher than the measured values due to tester inaccuracies at such low loop impedances.