Difficulties finding a fault on lighting...

[user 6351]

Morning All!

I have a problem with a lighting circuit that I am banging my head against the wall a little with and could do with a fresh outlook if anyone has any thoughts...

Customer called to say that they had no lights upstairs, the MCB has tripped and will not reset.

When I get there, I find that the circuit in question supplies the two upper floors of the house plus the loft. Except for the landing lights and loft lights, all the rooms have numerous GU10 downlights.

The downlights are fitted with just enough cable to change the lamps, but not enough to pull them out to have a good look at cables feeding them.

The ceilings are on 3x2 joists and have been filled with insulation as well.

I managed to get a look at a small section of the cable in the bathroom and found that (probably unsurprisingly given the installation!) the cables were toasted in at least on point.

I have replaced all the cable to the downlights, which was not easy as the floors above are all chipboard and there are beams and joists running in both directions, so there were a lot of access holes needing to be cut! However, the MCB still will not reset - when I disconnect the circuit from the CU and stick the tester on it, I get 89 ohms L/N to E on an IR test and 0.16 ohms on a continuity test L to E.

So, I now obviously need to find where the remaining fault(s) are. I have checked the MCB to make sure that the fault doesn't lie there.

The problem I have is mostly that of access... If this were a basic installation with loop in/out at pendants etc, then I could easily isolate sections of the installation and find the fault. However, with everything being downlights; no access; and no clear view of where the cables are routed, I am a bit stuck.

I think it may well be likely that there are other sections of cable which have been fried and need replacing, but as you can imagine, I am loathe to start cutting loads more holes in the floor ( I cannot lift the boards easily as the customer has had stud walls added since the floor was laid).

So, the question is this... does anyone have any techniques that they use in the this sort of situation which could make finding the fault easier??

Thanks a lot.

Ian.

 

[sparkytim]

Sounds like you have L -E fault pretty close to the supply end with that 0.16 ohm reading.

 

[user 6351]

Thanks Sparkytim - I had completely forgotten that I could do the maths on that and get a guesstimate of the location (I was getting a bit stressed at that point as I had hoped the fault would be solely in the bathroom!)

BTW - I forgot to mention earlier - I have advised the customer not to simply just stick all the halogen lamps back in and start the problem again!

 

[SLIPSHOD & SLAPDASH]

Have you checked to see if the circuit is looped at the switches. If so its easy to seperate sections of wiring.

 

[SPECIAL LOCATION]

When I get there, I find that the circuit in question supplies the two upper floors of the house plus the loft. Except for the landing lights and loft lights, all the rooms have numerous GU10 downlights.What is the approximate loading of all these lamps?

How many What ratings? IS THE CIRCUIT OVERLOADED?

- when I disconnect the circuit from the CU and stick the tester on it, I get 89 ohms L/N to E on an IR test and 0.16 ohms on a continuity test L to E.

This would have been the first thing IMHO test the cable to find out what sort of fault you are looking for no?

89 ohms L/N to E on an IR ??

My IR tester would have just shown 0.0Mohms on an 89ohm reading?

Do you mean 89 Meg?

Cant see how you get 0.16ohms between L to E on continuity AND 89ohms/MegOhms between L+N to E on IR??

What size cable?

1.0mm 1.5mm???

I have checked the MCB to make sure that the fault doesn't lie there.

The problem I have is mostly that of access... If this were a basic installation with loop in/out at pendants etc, then I could easily isolate sections of the installation and find the fault. However, with everything being downlights; no access; and no clear view of where the cables are routed, I am a bit stuck.

How old is the wiring, who fitted the downlights have you asked the customer for the previous test certificates?

I think it may well be likely that there are other sections of cable which have been fried and need replacing, but as you can imagine, I am loathe to start cutting loads more holes in the floor ( I cannot lift the boards easily as the customer has had stud walls added since the floor was laid).

If cables have been fried you are going to have to rip stuff out anyway?

So, the question is this... does anyone have any techniques that they use in the this sort of situation which could make finding the fault easier??

Thanks a lot.

Ian.

My comments in BLUE

The bottom line is this....

either the customer wants it fixed or not?

If it has not been correctly installed. (Good workmanship proper materials etc..)

Then someone has got to start going through it section by section testing & verifying in a logical process

An important thing is...

Point out ot the customer that test certificates have been required by BS7671 for a long time..

I guess the lights were fitted in the past 15+ years..

in which case they should have been correctly tested and an EIC MWC issued

So you can pre warn the customer that it may be more complicated and time consuming as the correct paperwork is not available...

i.e.

You should have been able to compare your circuit test results with the original circuit results

SO..

It sound like you are just going to have to slog it out....

Take a note pad and pen with you and start drawing out each section that you identify and test..

readings..

number of cables at any joint you find etc..

As Tim reminds use the resistance values on cable to esitimate how long sections are.

e.g.

10m of 1.0/1.0mm R1+R2=0.36ohms

10m of 1.5/1.0mm R1+R2=0.30ohms

so your 0.16ohms could be approx 5m away..

BUT you may also have other faults further away..

e.g.

four parallel faults 4x the distance away would also give you 0.16ohms

i.e. 4 branches each with a L-E fault 20m away would individually be 0.64ohms

but at the CU 0.64ohms & 0.64ohms & 0.64ohms & 0.64ohms in parallel will give 0.16ohms

Other than that..

take a big flask of coffee and get stuck in.

:C

 

[user 6351]

Have you checked to see if the circuit is looped at the switches. If so its easy to seperate sections of wiring.

Thanks, unfortunately yes, I have checked and no they aren't.

All switches are metal clad unfortunately, so may end up removing them all as well.

 

[Doc Hudson]

Does the MCB trip even if the lights are switched off or only when some are switched on? This would point to either fault on a perm live or a switched live somewhere.

Doc H.

 

[user 6351]

Specs, thank you for a detailed response, I won't quote everything directly to avoid too long a post but answers are:

Approx 20 x 50W GU10 - 600W / 2.6A on a 6A MCB (n.b. All lamps removed and still trips)

89 ohms on IR with my Seaward MFT (although I did check at 100V to start off with and didn't ramp it up once I got the low reading).

Cant see how you get 0.16 between L to E on continuity AND 89/MegOhms between L+N to E on IR??

No, me neither, which is one reason why it's annoying me!

Cable is 1.5/1.0 T&E

Unsure as to age, suspect 10 - 15 years but cannot determine.

Think it may be a DIY job by last owners of the house, and there are no test certs.

Current owners have been in for over two years, and fault has only appeared in the last couple of weeks.

Last work that they had done was the partition walls on the upper floor, completed > 18 months ago.

Husband is banned from using tools in the house and all pictures etc are hung using no more nails...

Yes, agree that if the cables in the other rooms are cooked then I will have to access them forcibly, but with the view I have at the moment, I cannot be sure that they are and want to avoid unnecessary pulling up of carpets and cutting of holes if at all possible.

Whole installation is a bit cowboy to be frank - T&E taken to each D/L, but no earth connections at any point; inaccessible JB's with covers missing and terminals broken; cable routing a little suspect. And that's just on the lights in the bathroom that I've changed!

I agree that it'll be a long old slog... and possibly a lot of disruption too, but yes, the owners would rather like to be able to see when they go upstairs, so it needs fixing!

Will get my book of resistance values out (or just use the one's you've both quoted already) and try to find my most likely starting point and take it from there... will remember to take my sandwiches next time too!

---------- Post Auto-Merged at 11:25 ---------- Previous post was made at 11:20 ----------

Does the MCB trip even if the lights are switched off or only when some are switched on? This would point to either fault on a perm live or a switched live somewhere.Doc H.

I can't tell unfortunately. The switches are mostly push on/off dimmers, and everything is un-lamped, so I don't know the setting of the switches at the moment.

I may well go down the route of removing all the switches though with them being metal, so that will help me identify any switched live issues.

 

[Steve M]

Could it just be a faulty MCB just got a bit knackered over the years like me??

 

[user 6351]

Unfortunately no, I moved the circuit onto another MCB and that tripped too.

 

[batty]

You need to go back to basics dimmers out so you can actually test whether lights are on or off. As it appears to be L-E fault at least that way you can narrow it down to rooms by switching rooms off. If fault is still there you need to track further back. I personally think it may even be in a stud wall.

 

[Electroglow]

Reeeeeeeeeeee wire

 

[steptoe]

LMFAO

first thing you need to do is bypass ALL the dimmers,

take them off and block connector the cables as required,

how on earth can you possibly do an IR with dimmer switches in circuit?

even continuity will be all messed up.

 

[user 6351]

Thanks all, it was getting rather annoying the other day and I was getting stressed which didn't help the thought processes...

Now I've got a bit of reassurance, it's back to it and get the flipping fault located!! Keep it logical and work through it all methodically and I'll get it sussed!

 

[ProDave]

I can't see anyone else stating the obvious, so I will.

You have to find where the lighting circuit loops (at the light fitting, at the switch, or at a junction box) then separate each leg and test each leg individually for IR with both ends of the leg disconnected.

Tedious, but it's the only way to find the fault. Once found, replace the faulty leg.

 

[user 6351]

Thanks Dave,

That's the problem, with the way it's been installed, it's not that easy to find the loops (I'm not saying impossible, just invasive and tricky).

I just noticed Specs comment too about testing the circuit earlier - I didn't do any IR testing etc to start with as I had managed to see one section of the bathroom light circuit and it was quite obviously toast, so there was no point in doing an IR until that part was replaced anyway.

Anyhoo... I shall go back on Tuesday with a clear head and a bucket full of patience and see what a new day brings... will update when I can. (fingers crossed!!)

 

[SPECIAL LOCATION]

Anyhoo... I shall go back on Tuesday with a clear head and a bucket full of patience and see what a new day brings... will update when I can. (fingers crossed!!)

I did hear some wholesalers were offering a free box of good luck with every two buckets of patience purchased!

:slap

When you go back try and get some photos for all of us electrical perverts to look at...

We Luuuuuuuvv toasted cable joints!! ; \

 

[batty]

We have all been there well i have running late on a job then you get a fault. You need to start with a clear head and generally you will get there a lot quicker.

 

[user 6351]

Morning all, back on site yesterday with a full flask and a bag of snap, so here's a quick(ish) update...

First step, try the MCB in the hope that something has miraculously fixed itself... no such luck!

So, all metal fittings removed - all switch plates and three wall lights. There's a flourescent in the loft, so that gets disconnected too.

Test of IR at the CU: I get >200M ohms L-N and N-E and 0 ohms L-E.

Test of continuity gives me 0.51 ohms L-E, so working on 24.2mohms for R1+R2 on 1.5/1.5 T&E I get the fault as being about 21m from the board.

The only other place that I can get to a live and an earth is in the loft switch, and that gave me 0.33ohms, so approx 13.6m from the switch...

There are 9m of cable from the switch to a JB in the corner of the loft, which I could see but not access, so I keep my fingers crossed that the 4m to the fault are in the loft and not in the loop from the floor below and start ripping up the loft boards (it had been done really well, screwed down tight and a couple of coats of varnish, so no chance of raising them neatly

At long last, a junction box!!! Pop the lid and test - 0ohms woohoo!! Pull everything out and test in turn, and find it's the feed to the loft lights that's at fault. Replaced, IR of >200Mohms at the CU, connect everything up and write an invoice!

It was a long day, but a good result, thank you all for your various thoughts. (I didn't get any pictures of toasted stuff as it had all gone in the bin - but I did get some interesting (hopefully) DIY beauties that I'll post elsewhere in a minute!)

 

[ianmacd]

if its 1.5t&e its unlikly a diy job. They wouldn't waste the money over 1.0t&e.

Well done on finding the fault. Fault finding is always the most interesting part of a sparkys job.