Most effective or correct way to fault find?

[PEG]

I am just itching for a fault to come my way now!! It's not going to stand a chance!

...3 storey victorian multi-occupancy dwelling,all floors on one circuit,laminate in every room,joiner,plasterer,decorator all left this morning...enjoy :conehead:

 

[Lucien Nunes]

If it's a hard fault like a screw penetration, you can find it without any opening or splitting. Ensure everything is disconnected, separate the ends at the DB and connect a small regulated current source, say 1A, across the faulted cores of one end only. Go round the points reading the voltage across these cores. As you move towards the fault from the energised end the voltage will drop, when you reach the fault it will be near zero and all points beyond it will be identical.

In effect you are doing manually what a continuity tester does, but because you are using a four-lead connection to the fault (current injected through the circuit conductors, voltage read via the outlet) you eliminate the relatively high and variable contact resistance and get a much more precise indication.

If the CPC isn't picking up any parallel paths, the drop for 1A will be about 20mV per metre of 2.5 T+E, so with an ordinary DMM that will read in mV, you can localise to within a few inches.

Alternatively, if you pump in an audio frequency tone, provided the wiring is not in metallic containment you can find the fault with a search coil, if it happens to be buried in the wall rather than at a point. Standard practice in comms wiring.

 

[davesparks]

Now there's a good idea !
And constant current sources are pretty commonly available these days in the form of LED drivers. Might have to use 750 mA rather than 1A but that shouldn't make a huge difference I wouldn't have thought?

Or just knock up a 1A supply with a LM317 or similar

 

[Lucien Nunes]

Precisely. Unfortunately you can't just use a modern SMPSU and put it into current limit as it will fold back to next to nothing or go into burst mode.

Of course, if the fault is much more than an ohm or so it will tend to get lost amongst the cable resistance unless the circuit is very long, making it harder to locate. Also, if the test supply has much voltage compliance, you could dissipate significant heat and make the fault worse. Limiting it to a couple of volts is a good idea, or use a lower current like 100mA if you don't need pinpoint accuracy.

 

[PEG]

If it's a hard fault like a screw penetration, you can find it without any opening or splitting. Ensure everything is disconnected, separate the ends at the DB and connect a small regulated current source, say 1A, across the faulted cores of one end only. Go round the points reading the voltage across these cores. As you move towards the fault from the energised end the voltage will drop, when you reach the fault it will be near zero and all points beyond it will be identical.

In effect you are doing manually what a continuity tester does, but because you are using a four-lead connection to the fault (current injected through the circuit conductors, voltage read via the outlet) you eliminate the relatively high and variable contact resistance and get a much more precise indication.

If the CPC isn't picking up any parallel paths, the drop for 1A will be about 20mV per metre of 2.5 T+E, so with an ordinary DMM that will read in mV, you can localise to within a few inches.

Alternatively, if you pump in an audio frequency tone, provided the wiring is not in metallic containment you can find the fault with a search coil, if it happens to be buried in the wall rather than at a point. Standard practice in comms wiring.

I have an older type BT Tone Generator,that i have played with,although i have to admit,not having actually diagnosing a fault using it. It is in a box with some Yankee gear that we experimented with years ago,in an attempt to apply a traceable signal to an underground water main. If i remember,its' success rated alongside that of a capable divining rod exponent...:conehead:

 

[davesparks]

Divining rods work pretty well so I assume that was successful?

 

[ferg]

I recently did an EICR on a small 3 bed property where it had been DIY rewired at some point.

The socket rings were a bit awkward to fault find.

1 socket in lounge, 1 in kitchen, 1 in bed 1, 1 in bed 2 ...


Next ring similar.

Lighting similar.

Immersion cct also fed a socket in living room and 1 in kitchen.

You get the picture.

A nightmare to map out never mind fault find.

In the end my best friends were pen, paper and patience.

 

[SirKit Breaker]

Before answering this, i want more info, such as How has the fault come to light, what are the symptoms, and what are the protective device arrangements. I would then decide how to find the fault.

A recent example i had.

Lighting circuit trips RCD when any light on circuit is switched on, no fault with lights switched off, it has a 6a mCB and 30mA RCD. It is wired in 1.5mm T&E, and is brand new rewired circuit loop in/out. So straight away i know switch wires are OK, so i IR each leg of the loop and find that bed 3 to bathroom loop has 0 meg N-E, Replaced cable job done.

Cheers.............Howard

 

[TJ Anderson]

Before answering this, i want more info, such as How has the fault come to light, what are the symptoms, and what are the protective device arrangements. I would then decide how to find the fault.

A recent example i had.

Lighting circuit trips RCD when any light on circuit is switched on, no fault with lights switched off, it has a 6a mCB and 30mA RCD. It is wired in 1.5mm T&E, and is brand new rewired circuit loop in/out. So straight away i know switch wires are OK, so i IR each leg of the loop and find that bed 3 to bathroom loop has 0 meg N-E, Replaced cable job done.

Cheers.............Howard

That's bad on a newly wired circuit. What had happened to the cable on the Bedroom to Bathroom Leg to give to N-E fault?

 

[David M]

That's bad on a newly wired circuit. What had happened to the cable on the Bedroom to Bathroom Leg to give to N-E fault?

Not sure if any of you guys have come across this, but a few years back while doing a rewire during the final test we found a similar fault. It turned out to be a length of faulty t&e. I took the cable back to the suppliers as it was a manufacturing fault.

 

[Prat P]

Evening all...

I'm in two minds whether to post this but i've thought about it for a while and was hopeing someone could give me a very clear idea of how they would find the fault.

Lets say we have a ring. We test between N-E at the CU and find a poor IR reading, lets say 0Mohms, ie socket screw has gone through the neutral.

I break the ring somewhere in the house and then test both neutrals to see which side the fault is on.

I then break the ring between the CU and the already broken point (on the fault side) etc etc until I have found the fault.

This sounds very straight forward, but in practise it can be very difficult to know which sockets/FCU are actually on the faulty side once you have broken the ring as you have to completely guess at the layout of the ring.

The only way I could think of doing it would be to have a good idea of where all the sockets fed by the ring are in the first place, then break the ring, find the faulty side and then reconnect just the faulty side in the CU so that you know that any socket that is dead is not on the faulty side and can be discounted. Any socket that is live could be the culprit, and slowly break it down this way which will also give you a bit of a picture of the layout of the ring.

How do others do this?

Cheers..

I always think the Murray loop test is the simplest and quickest test to locate a fault.

 

[shanky887614]

I always think the Murray loop test is the simplest and quickest test to locate a fault.

thanks for that i haven't heard of that one

electrical4u.com/blavier-test-murray-loop-test-varley-loop-test/

 

[andyelectric]

Half Split Method...

(This is from general fault finding on equipments)

Unless a fault is a usual suspect (which can lead you down a wrong path if you're not careful) always try and work methodically. Narrow down on a 50-50 basis unless there are reasons to say otherwise. Often these reasonings will be gut from experience but usually they will be deduced.

But first, check the ffing fuses.

 

[marc8]

Hi all
did have a problem on some ground lights outside years ago friend of mine I was helping for the day, we started making resin kits up splitting cable down the drive ( a long way) about 20 joints total split cable terminate join light cable anyway on the next day it was being back filled & Mike was testing when he found N/E fault on IR test, found fault quite quick by doing r1 r2 at each light starting at beginning working his way Down readings started going down till go to low place then started from other end got the same low reading at that point, dug back up & hey presto that was the one cut in new joint bigger swa kit job done.
Could work on ring disconnect from DB join ends & go round do a r1-r2 looking for lowest reading at one point or between 2, split legs at board & repeat process, should have a low reading at one or two points, this way no sockets off can be done in any combination l/n n/e. On continuity for obvious reasons not IR, low ohms should do the trick, just a thought

 

[shanky887614]

Hi all
did have a problem on some ground lights outside years ago friend of mine I was helping for the day, we started making resin kits up splitting cable down the drive ( a long way) about 20 joints total split cable terminate join light cable anyway on the next day it was being back filled & Mike was testing when he found N/E fault on IR test, found fault quite quick by doing r1 r2 at each light starting at beginning working his way Down readings started going down till go to low place then started from other end got the same low reading at that point, dug back up & hey presto that was the one cut in new joint bigger swa kit job done.
Could work on ring disconnect from DB join ends & go round do a r1-r2 looking for lowest reading at one point or between 2, split legs at board & repeat process, should have a low reading at one or two points, this way no sockets off can be done in any combination l/n n/e. On continuity for obvious reasons not IR, low ohms should do the trick, just a thought

on ring.

link N1-E1 and L2-E2 at board

then go to where you think the middle is then check continuity on n-e and l-e this will tell you which leg is down.

connect it back up and pick another socket but remember the last resistance value, some people dont install rings the way you think.

me and a mate had a nightmare trying to find a fault on a ring circuit becauss of the way the ring was wired, split the ring in half and 1/2 came on in each room with some still off.

had to resort to drawing it out where we though ring went while trying to split it and work upto fault

 

[marc8]

Hi shanky
what I meant was you don't need to know the middle you are looking for the lowest reading on the ring on any of the pairs or between the pairs, this could be fault location, ie if you had say 20,19,17,45,19,0.1, 21 ect then surely lowest reading could be location of fault on low continuity, forget it being a ring just look for the lowest reading, no sockets of no split only at board just to check split legs at board & go round again if same location lowest reading should be the fault. Just a thought without splitting down the ring straight away, next time I get call for similar fault I will give it a quick go should only take 15 mins depending on size of ring

 

[TJ Anderson]

Hi shanky
what I meant was you don't need to know the middle you are looking for the lowest reading on the ring on any of the pairs or between the pairs, this could be fault location, ie if you had say 20,19,17,45,19,0.1, 21 ect then surely lowest reading could be location of fault on low continuity, forget it being a ring just look for the lowest reading, no sockets of no split only at board just to check split legs at board & go round again if same location lowest reading should be the fault. Just a thought without splitting down the ring straight away, next time I get call for similar fault I will give it a quick go should only take 15 mins depending on size of ring

That works when you have a low impedance fault such as a short. the lowest reading obtained will be closest to the fault. For insulation faults in M/Ohm range it wont be accurate enough and will show the same reading throughout the circuit.

 

[marc8]

Hi that's what I meant if the fault was low enough it could show easy to check if it shows up on a continuity check at board then should be able to use it to find it, if it is to high then I would agree it won't work & would have to user IR testing.