Insulation resistance puzzle

[pete the plug]

I was doing an insulation resistance test today on a newly installed Lighting circuit, (1.5mm 6242y ,hall/landing, and a couple of outside lights)  5 downlights on the hall landing and stairs, and two outside lights, I've used wagoboxes for the interconnections, as the downlights have removable loop in/out plugs and sockets fitted on them. I tested the downlight chain first before connecting in the loop to the outside  lights, all good, resistance of over 199 Mohms, however when I looped in the outside lights the reading fell to 150 Mohms. 

Logic dictated that the lower value was the fault of the outside lights, but doing a test just on that loop also gave a reading of greater than 199 Mohms, I isolated all the cables in that circuit, tested them all individually, removed the  2 light fittings and tested them, all giving a reading of greater tan 199 Mohms.

Connecting either or both of the outside lights to the overall circuit reduced the insulation resistance to around 150 Mohms.

I realise that this is a high and quite acceptable value, and if, like many test instruments that read up to 99  Mohms had been used I would have been totally unaware of any reduction.

The results don't seem to make sense though, and the only solution that I can come up with, is that possibly the resistance value of each leg of the circuit is over 199 Mohms, too high for the meter to read, but when they're both connected together the combined value of the two legs , effectively in parallel is lower that the 199 Mohm limit of the meter?

Though, having said that,  the entire outside light circuit from the junction box including switch drop  is no longer and about 6 metres long, so there should be very little leakage resistance. I've used a CK automatic cable stripper, wondering if anyone's had any problems with the inner conductor insulation being damaged when removing the outer sheath with an auto stripper?

Any comments would be appreciated thanks!

 

[binky]

Test  results are well in spec, so I wouldn't worry about it - I've driven myself potty trying to find a cause of things like you desribe.. IR will reduce with total length of cable ie the more cable / fittings being tested the lower the IR result be, but it does seem a largish drop.

 

[Murdoch]

I use a metrel MFT which has auto ir test .... the results only go up to 30 meg .... as long as the reading is over 2 meg who cares .....

 

[Andy™]

each on their own is over 199. combined they are not. going by your results, id say each is 300, or one might be 250 and the other 350 etc

 

[pete the plug]

Thanks folks for the replies, I was wondering too, if I had some duff cable, shouldn't be as it' was new dated Prysmian, 1.5mm, On the principle of  "never assume anything", wish I'd thought to do an IR test on that new complete cable  roll before using it. ( having once re-plumbed a bathroom without first checking the pipework, only to find when the hot water to the bath didn't flow, there was a crisp packet in the pipe that some kind soul at the builders merchant had inserted! haha)

 

[Tony S]

If you have a bad batch of cable get in touch with BASEC, they will conduct an investigation. It’s in all our interests that substandard cable is taken off the market.

 

[M107]

There is some duff cable doing the rounds, I've returned a roll of 1mm to tool-station in recent weeks the conductors were twisted over each other inside the sheath, found it mid way through the roll.

And last week my parents had some work done by a contractor (done under a grant for my disabled sister so I couldn't do it) when he came to test, the ring had O/C on neutrals.

Fair play he had tested before hand so knew it was his extension to circuit that had caused the issue.

Broke down the extension of ring & on close inspection found a leg between loft joint box & first socket he installed , had a 10mm length of neutral conductor missing inside the sheath.......cable picked up from Screwfix on way to the job.

 

[pete the plug]

Wow, M107, that's pretty bad about the Screwfix cable, though I think there's is Prysmian, which I think used to be Pirelli, so it should be OK, certainly worth doing an end to end resistance test on every roll, ( which I did, but not the IR ) saves a lot of possible hassle later.

I find IR readings a bit erratic, I've got a Kewtech KT63 it's just over a year old and was re-calibrated in January, I find that if I do three readings one after the other, and say, the first one is   600 Mohm, the second  and third readings may well be 100 M ohm higher each time.

Perhaps I'm expecting too much, and should be happy with an IR of 800 M ohm on a lighting leg with 9 downlights and 4 wall lights rather than striving to get "greater than 200 M ohm" reading? Having said that,  the figure of 2M ohm being an acceptable result seems ludicrously low ( I know the maths, but even so!!)

No problems with other  function readings, multiple button presses provide pretty much the same result. The IR reading seems very quick, I think the Megger MFT's take a lot longer, and maybe does a better job? 

 

[SPECIAL LOCATION]

Are you getting this lower reading with some of the lights connected or is everything removed??

If you have some fittings connected, (especially some of these modern low energy jobbies or PIR movement sensors connected on the outside lights), then that may account for some drop....

But , if it is all new wiring, AND everything is disconnected I would suggest you do have some cable damage somewhere...

IMHO in a domestic property it would have to be one helluva long circuit to get any significant IR drop below 200meg+

I can't recall any new circuits I have ever wired with that sort of drop?

Have you double checked no pinched wires in the back of any light switches?

How are your batteries in your meter?

I would probably want to do a second check with a different meter..

(I have at least two meters that can do each of the test readings I need to take....

Not all calibrated BUT whenever I have got a strange reading a second test with reserve meter can clear up a lot of questions.)

It is certainly a bit strange....

:C

 

[binky]

Connecting either or both of the outside lights to the overall circuit reduced the insulation resistance to around 150 Mohms.

 re-read your post - what are the light fittings? LED / PIR / halogen?

 

[pete the plug]

Hi Guys,  thank's so much for the replies, this really is a friendly and helpful forum.

 Sorry,  Looking at my posts I see I've left out a zero on some of the  the readings.  initial post should have read 1500 not 150 M ohms, and 2000 not 200Mohms, sorry about that stupid error!!!.

The outside lights are  GU10 fittings and will have LEDs ( not put in for testing)

No fittings or switches inside as yet, I've used Wago connectors and boxes in lieu of ceiling roses, and temporary Wago connectors for the loop in/out on the downlight fittings ( which will have  Enlite  E6 downlights with loop in/out plugs and removable lights) for the initial testing, I like to test one room or "set" of lights at a time just to be sure, had all individual IR readings of over  two thousand megs, just seems to be a cumulative effect giving a reduced figure.

The  upstairs lighting circuit ( 13 downlights , 4 switch drops, extractor fan wiring with switch drop outside bathroom, and mirror/ shiver socket wiring ) gives an IR reading at the CU of

L-CPC about 1000megohm

N-CPC. about 1000megohm

L-N above 2000 megohm ( I would expect a higher reading with two effective layers of insulation between these conductors?

Would this be considered typical please? and in the "real world" am I expecting too much wanting all my readings to be greater than two thousand megs?

I should point out that I'm not Part P registered, and this work, on my son's house is being done after notifying local building control, who will do the first fix checks and final certification. I do however want to make sure it's all ok and up to spec before that takes place, for my own satisfaction as much as anything else. 

Thanks

 

[Tony S]

1000MΩ? That’s 1GΩ, your meter wouldn’t be able to give a reading above 199MΩ @1000V DC.

You really need to get a competent electrician involved.

 

[pete the plug]

Thank you Tony S for those kind words, I've attached the spec of the KT63's insulation test functions, which would indicate that it reads to 1999Mohms or 2Gigs

 

[kerching]

What voltage are you testing at?

my Fluke 1654B will measure up,to 500M at 1000V

just saying 

 

[pete the plug]

I've tested at 500 and 1000 volts, the accuracy of the tester decreases  with the higher resistance  readings, they only quote accuracy at 500V up to 499Mohms, which is plus or minus 6%, and 999Mohms at 1000V which is also plus or minus 6% , not sure what the accuracy would be at the higher readings, or for that matter why bother to have higher readings if the accuracy is questionable?

 

[Tony S]

I’ll lay money on you’re just giving the meter a quick blip on the test PB.

The higher the voltage the lower the resistance seems to be at the start of the test. Cables are in effect capacitors, they need time to charge.

Learn how to do IR testing.

 

[binky]

500v is plenty for this testing, but if you are getting those sorts of figures you have absolutley nothing to worry about, and as said above hold the test button for a few seconds until the reading stabilises - can't say I hold the button for long, I watch the readings rise on my megger and as soon as they get up to max, I let go - no point going beyond max test figure.

 

[Tony S]

Something I should have mentioned, don’t handle the cable ends until you have discharged them. A long cable packs one hell of a punch.

 

[Phoenix]

I think a lot of modern testers will automatically discharge, Tony, problem comes when you mentally note the figure then pull the probes off before you let go of the button.

Did it once on a lighting board, three phases linked out to neutral, neutral test link out and tested all conductors to earth. I then went to put the neutral link back in I seem to remember I just got a bit off a sizeable 'crack' as it discharged, rather than a belt (I think I've always been carefull enough to manipulate neutral links with insulated tools!)

 

[Tony S]

They do but depending on the amount of charge in the cable it can take a few seconds. At the foundry I had a 5kV Megger, a fully charged cable could take up to 30 seconds to discharge, fortunately it had a warning LED.

It's funny when someone climbs over the “electrical test area” barriers and gets 2.5kV up their arm, they really squeal.