Strange/Unusual Fault

[mattg4321]

So there I was the other day posting on here about me going to solve faults others couldn't find and I've got one myself.

It's been a while since I've been stumped but I am now.

Symptoms - gate supply keeps tripping (when it's raining according to customer). There's a 6mm 3 core SWA running about 3-400m (yes you read that right), to a cabinet by the gates where there is a double socket inside feeding a right mess for the gates. Protected by Hager reduced height 20A RCBO.

Unplugged both sockets, IR tested cable from consumer unit to double socket.

L-N = 140Mohm
L-E = 71Mohm
N-E = 70Mohm

IR tested the plugs for the gate controls. L&N together to earth. One was 12Mohm, the other was 299Mohm. Throughout all the below I never plugged the gates back into the double socket.

With the plugs still unplugged and nothing in the double socked I went to turn on the RCBO. Took 4 goes to get it to stay in. No bang, just as if there was N/E fault there. 1/2 trip and ramp tested RCBO and both returned no result as it tripped so early/quickly. Removed outgoing cables from RCBO and all tested well - 23mA ramp test and turned on without any trouble.

Changed RCBO to rule that out and same symptoms.

Tested IR a few more times to make sure and same results. Tested outgoing cables to socket to all other cables in consumer unit to see if anything was mixed up. All good. Physically traced cables in consumer unit to check same and again all good.

I'm guessing it's something to do with the length of the cable? I can only guess that somehow I'm losing current from the line to the earth as they are run alongside eachother for such a distance? This doesn't really make sense to me though. Someone put me out of my misery!

 

[DPG]

So there I was the other day posting on here about me going to solve faults others couldn't find and I've got one myself.

It's been a while since I've been stumped but I am now.

Symptoms - gate supply keeps tripping (when it's raining according to customer). There's a 6mm 3 core SWA running about 3-400m (yes you read that right), to a cabinet by the gates where there is a double socket inside feeding a right mess for the gates. Protected by Hager reduced height 20A RCBO.

Unplugged both sockets, IR tested cable from consumer unit to double socket.

L-N = 140Mohm
L-E = 71Mohm
N-E = 70Mohm

IR tested the plugs for the gate controls. L&N together to earth. One was 12Mohm, the other was 299Mohm. Throughout all the below I never plugged the gates back into the double socket.

With the plugs still unplugged and nothing in the double socked I went to turn on the RCBO. Took 4 goes to get it to stay in. No bang, just as if there was N/E fault there. 1/2 trip and ramp tested RCBO and both returned no result as it tripped so early/quickly. Removed outgoing cables from RCBO and all tested well - 23mA ramp test and turned on without any trouble.

Changed RCBO to rule that out and same symptoms.

Tested IR a few more times to make sure and same results. Tested outgoing cables to socket to all other cables in consumer unit to see if anything was mixed up. All good. Physically traced cables in consumer unit to check same and again all good.

I'm guessing it's something to do with the length of the cable? I can only guess that somehow I'm losing current from the line to the earth as they are run alongside eachother for such a distance? This doesn't really make sense to me though. Someone put me out of my misery!

I'd be surprised if that could cause enough current to flow to trip an RCBO. 15mA+ just by the coupling of adjacent cables?

 

[Jim_e_Jib]

Did you include the armour in the IR tests?

 

[mattg4321]

Armour was connected to cpc and earth at all times

 

[timhoward]

Is it a metal double socket?
What I don't think you've mentioned is taking front off double socket, but see exactly what you mean about the mix of IR tests and RCBO tripping.
Is it an RCBO with a functional earth connection?

 

[snowhead]

Change the socket, or put the cable ends in a block and power up.

 

[stevethesparks]

I wonder if you may have some capacitance issues over such a run? Just a thought.

 

[Lucien Nunes]

Yes, capacitive leakage is a thing and on long cables it's an important thing. This case looks borderline.

Manufacturers don't usually publish capacitance for small LV power cables in their data, but at some point I clearly managed to get data for 2.5mm SWA as I wrote about it in this thread (although did not include the link mentioned there):

Capacitive Inductance

Hi Everyone, I'm just hoping someone has had this problem and found a way to get over it. I do work for water feature companies, yesterday I came across an awkward problem, they'd installed maybe 70m of 2.5 3 core SWA, down to a pond to run a pump and UV light, load total 7A. The idea was to...

www.electriciansforums.net

My figure of 0.25μF/km was for one core to all the others, whereas with a 3-core on single phase, leakage to the neutral core does not create residual current, so the effective L-E capacitance for our purposes is somewhat lower. However, the cable here at 6mm² has conductors of larger surface area but probably not spaced proportionately further apart, so the capacitance per metre will be higher.

If we took a wild guess at 0.3μF/km L-E effective, the cable will contribute
230 * 2 * pi * 50 * 0.3^e-6 = 22mA of leakge per km

Hence 22 * 350 /1000 = 8mA for a 350m cable (very roughly, since we don't have the actual capacitance for that cable size). That is the leakage at 50Hz, however, when closing the RCBO on the peak of the mains waveform, the fast risetime voltage step might pass a big enough slug of charge to earth to trip the RCBO straight back off. The leakage might be somewhat higher due to harmonics and noise on the supply, too, as the capacitance to earth tends to filter them off. I could imagine a situation where switching large loads on the same supply caused spikes of many mA through the cable capacitance.

If the customer is correct that it trips more when raining, there might be moisture ingress into the cable which, if it surrounds the line core insulation, further increases the capacitance. Alternatively, the tripping when already energised might be a combination of capacitive and resistive leakage (from other causes), while the difficulty resetting is due to the transient voltage across the capacitance.

100mA RCBO at the source for fault protection and 30mA at the destination for additional?

 

[timhoward]

L-N = 140Mohm
L-E = 71Mohm
N-E = 70Mohm

I'm thinking out loud here...
Assuming armour is earthed both ends.
With E and armour earthed you have same effect as a shorter cable and reduced (normal) resistance.
IR is inversely proportional to the cables length.

All that to say shouldn't the IR measurements involving E be higher than the L to N reading?
EDIT - ^^complete rubbish^^ penny has dropped now....

 

[Lucien Nunes]

I don't follow you, but the individual resistances core-core and to SWA are governed by the bulk resistivity of the insulation and filler materials, and the area and thickness of each. With the large area of SWA in contact with the filler, and this not necessarily having high insulation properties, one might expect the resistance from each core to be lower to the SWA than to another isolated core, since only one layer of insulation is involved instead of two. If the CPC core is strapped to the SWA, I would expect results much as the OP has measured.

OTOH, these resistances are all up in the 'ignore' category. They would need to be a thousand times lower to be playing any significant role in the RCBO tripping.

 

[mattg4321]

Great post as usual @Lucien Nunes

It seems I was on the right track then. I had been thinking of 100ma at source and 30ma at gates on the way home as a possible solution.

 

[timhoward]

With the large area of SWA in contact with the filler, and this not necessarily having high insulation properties, one might expect the resistance from each core to be lower to the SWA than to another isolated core

Thanks. That was the main detail I didn't take into account at all. It was also completely illogical to compare normal resistance changes due to csa with the general rule of IR getting shorter as a cable gets shorter. Total brain fail really!

 

[littlespark]

Monday night brain click here, between moving furniture and packing a bag….
(We are going for a few days holiday, she hasn’t chucked me out just yet!)

does the circuit need any rcd at source end at all? Armoured cable

 

[Pete999]

I'd be surprised if that could cause enough current to flow to trip an RCBO. 15mA+ just by the coupling of adjacent cables?

Call someone with the correct test gear to test your installation

 

[DPG]

Call someone with the correct test gear to test your installation

Lol. Not my installation Pete. I was just thinking aloud!

 

[Lucien Nunes]

does the circuit need any rcd at source end at all? Armoured cable

Maybe not. Top of head on a B20 2 cores or XLPE cable with armour alone are out, but 3rd core + armour are OK. Probably doesn't need to be 20A anyway.

 

[timhoward]

Yeah with a ~400m run the Zs will be interesting.
I think it works out around 2.2 ohms
So sitting on the 80% limit of a B16

 

[mattg4321]

I got Zs of 3.96ohms. Thats on the 'no trip' setting though so not 100% reliable.

400m was a bit of a wild guess. Could even be more to be honest. Dependent on route taken too. I know it's a 5 minute brisk walk for me!

 

[Pete999]

Lol. Not my installation Pete. I was just thinking aloud!

Right OK

 

[DefyG]

This is an existing installation right, that has been working satisfactorily for years and just developed a fault?

Cable capacitance is a plausible cause of leakage but if the circuit has been fine for a while/years I'd be looking elsewhere, has there been any work (fence posts!) along its route? Have you tested it while its raining (or just after)?

 

[Lucien Nunes]

It's a valid point, and my hunch in post #8 was that there might be leakage contributions from the capacitance (mainly causing the difficulty resetting) and damp (causing the tripping during rain as mentioned by the customer).

Though I would find it surprising for cable with insulation damage underground to recover from the tens of kilohms needed to cause a trip, all the way to 70MΩ L-E as measured by the OP. Plus the L-E and N-E readings are identical and the L-N double the resistance, which is fairly convincing. The IR of one of the gate systems was much lower at 12MΩ.

The factor that suggests to me that capacitance is at least a significant player here is that both the original and replacement RCBOs tested fine on their own, and the cable tested fine at 70 - 140MΩ, but as soon as the cable was connected to the RCBO it became difficult to reset. I.e. the effect was apparent at the same time as the cable was showing good insulation, regardless of whether the insulation always remains good.

History is increasingly of interest. How old is it? When did it start tripping?

 

[mattg4321]

It was extremely wet here today, so I doubt rainfall will take the IR any lower than that found.

Customer has just moved in a few months ago. It’s tripping intermittently since they moved in. Consumer unit was changed early this year. Very plausible there would’ve been no rcd on the gate circuit beforehand.

I really don’t think there is any other realistic explanation than lucien’s. I was on those lines myself, but wasn’t sure if that was realistic. To be honest I’ve not come across this before. But I’ve also not come across a final circuit 400m plus in length that I can remember!

 

[marconi]

Can one buy a single length of 6mm2 swa 400m long?

 

[telectrix]

might be an idea to use an earthleagage clamp meter ( a decent one). once or twice i've had tripping issues and IR tests have been in the 100s of Megs. sometimes a IR test just will not show up a fault, esp. if it's a slight water/damp ingress. if the SWA does not need RCD protection (regs. wise), feed from a non-RCD way.

 

[James]

Can one buy a single length of 6mm2 swa 400m long?

Yes, most wholesalers will order it in for you.

in general I would not expect any issues until about 1km
if I remember right, you can have about 4 tonne on a large reel, after that it would be more difficult to obtain.

 

[baldelectrician]

I would be inclined to move the RCD to the gate location, even install an IP rated board with 2-3 RCBO's, then each socket would have an individual RCD

I know that they are phasing out RCD sockets but that is also an option at the gate end

 

[DPG]

I would be inclined to move the RCD to the gate location, even install an IP rated board with 2-3 RCBO's, then each socket would have an individual RCD

I know that they are phasing out RCD sockets but that is also an option at the gate end

Are they phasing out RCD sockets?

 

[timhoward]

if the SWA does not need RCD protection (regs. wise), feed from a non-RCD way.

Agreed. I think the issue's going to be ADS for the cable, with Zs of 4 ohms, unless the load is light enough for a B6 breaker, or even 10 amp BS88 fuse, it looks to me like it's going to need an RCD at source for fault protection.

 

[newfutile]

It’s bad practise to add RCD to comply with ADS ,this would not work on a Line-Neutral fault . This would be a departure from the regulations which does not improve safety , its better to run a bigger cable or parallel cables or even an extra R2 (cpc)

 

[timhoward]

It’s bad practise to add RCD to comply with ADS ,this would not work on a Line-Neutral fault . This would be a departure from the regulations which does not improve safety , its better to run a bigger cable or parallel cables or even an extra R2 (cpc)

I agree in ideal circumstances but we are not talking about adding one, we are talking about the effects of removing one. The original design which the OP has inherited is not ideal, and far from easy to change. It’s likely questionable with regard to volt drop too.
In any case I’m fairly sure it is in fact permitted to have an RCD for fault protection and separate overload protection as per reg 411.4.5

 

[Pretty Mouth]

It’s bad practise to add RCD to comply with ADS ,this would not work on a Line-Neutral fault . This would be a departure from the regulations which does not improve safety , its better to run a bigger cable or parallel cables or even an extra R2 (cpc)

RCD for fault protection is perfectly compliant. L to N faults do not need to disconnect within stated disconnection times, only L - E. 411.3.2.1

 

[Mikegh]

When you're IR testing the connected loads at the outlet

How can you be sure you're actually getting everything

Some stuff may not be in-circuit until aftehe supply is energised


Ime SWA faults show on 500v or 1000v , could be wrong though but usually depend on the IR tester to find an issue on fixed wiring

 

[Pete999]

When you're IR testing the connected loads at the outlet

How can you be sure you're actually getting everything

Some stuff may not be in-circuit until aftehe supply is energised


Ime SWA faults show on 500v or 1000v , could be wrong though but usually depend on the IR tester to find an issue on fixed wiring

Are you testing just the installation, or the power consuming equipment? two different testing procedures ime

 

[Mikegh]

Are you testing just the installation, or the power consuming equipment? two different testing procedures ime

I was respkndiu to the first post where he said he had IR tested the SWA.eun and then IR tested.the equipment with L+N connected

I'm saying you may not be effectively checking all the wiring on the connected load as some may not be in-circuit when supply is disconnected

 

[Mikegh]

I was responding to the first post where he said he had IR tested the SWA run and then IR tested.the equipment with L+N connected

I'm saying you may not be effectively checking all the wiring on the connected load as some may not be in-circuit when supply is disconnected

Don't seem to be able to edit anything

 

[Lucien Nunes]

I agree, and it woukd be helpful to confirm the gate operator leakage current when energised. But one of the OP's key points was that it is difficult to get the RCBO to hold with just the cable connected, and the loads completely unplugged. And the cable IR tests very good. I think we're going round in circles now, there's no mystery here IMO.

 

[mattg4321]

Exactly. The rcbo was tripping at a current too low to be registered by my tester with just a cable and a socket on the end of it with nothing plugged in. Sometimes it wouldn’t even reset.

Remove cable and all ok.

 

[Pete999]

I was respkndiu to the first post where he said he had IR tested the SWA.eun and then IR tested.the equipment with L+N connected

I'm saying you may not be effectively checking all the wiring on the connected load as some may not be in-circuit when supply is disconnected

Then if it's that strange carry out two seoerate tests , the wiring and the power consuming equipment \, to try and eliminate one from the other, no good just testing it all in one procedure, any good detective will tell you to eliminate one area at a time just saying, others may call me an idiot, but been called worse. and have a thick skin.

 

[Mikegh]

I agree, and it woukd be helpful to confirm the gate operator leakage current when energised. But one of the OP's key points was that it is difficult to get the RCBO to hold with just the cable connected, and the loads completely unplugged. And the cable IR tests very good. I think we're going round in circles now, there's no mystery here IMO.

Didn't spot that , I usually shoot from the hip anyway

 

[SparkySy]

Sorry I can't remember if you said you had disconnected the socket from the swa as well, it's not moisture in that causing it? I had a gate supply that kept tripping even when it hadn't been raining. Turned out to be the socket, the enclosure it was in was ip rated but was really damp inside, it was sweating and running off the top internally down into the socket causing a N-E short occasionally.
Sy

 

[Mikegh]

Then if it's that strange carry out two seoerate tests , the wiring and the power consuming equipment \, to try and eliminate one from the other, no good just testing it all in one procedure, any good detective will tell you to eliminate one area at a time just saying, others may call me an idiot, but been called worse. and have a thick skin.

Yes obviously 2 separate procedures ?

I wouldn't combine them

Test the cable at a high voltage and equipment separately , I didn't imply differently

 

[Lucien Nunes]

He did say clearly in post #1 that the loads were tested separately and the problem was still there with them disconnected.

Presumably the socket was included in the IR tests that showed 70 / 140 megs? In which case it doesn't seem to be particularly suspect.

 

[Mikegh]

He did say clearly in post #1 that the loads were tested separately and the problem was still there with them disconnected.

Presumably the socket was included in the IR tests that showed 70 / 140 megs? In which case it doesn't seem to be particularly suspect.

He did and I may retire from posting under the influence of alcohol

 

[timbobelfast]

So there I was the other day posting on here about me going to solve faults others couldn't find and I've got one myself.

It's been a while since I've been stumped but I am now.

Symptoms - gate supply keeps tripping (when it's raining according to customer). There's a 6mm 3 core SWA running about 3-400m (yes you read that right), to a cabinet by the gates where there is a double socket inside feeding a right mess for the gates. Protected by Hager reduced height 20A RCBO.

Unplugged both sockets, IR tested cable from consumer unit to double socket.

L-N = 140Mohm
L-E = 71Mohm
N-E = 70Mohm

IR tested the plugs for the gate controls. L&N together to earth. One was 12Mohm, the other was 299Mohm. Throughout all the below I never plugged the gates back into the double socket.

With the plugs still unplugged and nothing in the double socked I went to turn on the RCBO. Took 4 goes to get it to stay in. No bang, just as if there was N/E fault there. 1/2 trip and ramp tested RCBO and both returned no result as it tripped so early/quickly. Removed outgoing cables from RCBO and all tested well - 23mA ramp test and turned on without any trouble.

Changed RCBO to rule that out and same symptoms.

Tested IR a few more times to make sure and same results. Tested outgoing cables to socket to all other cables in consumer unit to see if anything was mixed up. All good. Physically traced cables in consumer unit to check same and again all good.

I'm guessing it's something to do with the length of the cable? I can only guess that somehow I'm losing current from the line to the earth as they are run alongside eachother for such a distance? This doesn't really make sense to me though. Someone put me out of my misery!

Strange set-up, certainly no design to this circuit.

What is the design current for the gates?

You would need a B10 60898 upstream to meet disconnection times.

As it's SWA can you swap out the upstream RCBO to an MCB and fit RCD socket at the gate for the additional protection?

Certainly would not have been a cheap install for 3-400m of 3C 6 XLPE!

 

[mattg4321]

No idea re gates design current. No details present. To be honest I’m trying to stay well away from that side of things. It’s a single ram type arm that opens 1 gate.

Doubt that total load can be much. A B6 mcb would comply and might work fine? I guess the motor should only be a few hundred watts. Anyone like to make a guess?

I forgot I have a picture of the completely typical gate wiring I find whenever I look inside one of these cabinets. Utter disgrace in there.

 

[timbobelfast]

No idea re gates design current. No details present. To be honest I’m trying to stay well away from that side of things. It’s a single ram type arm that opens 1 gate.

Doubt that total load can be much. A B6 mcb would comply and might work fine? I guess the motor should only be a few hundred watts. Anyone like to make a guess?

I forgot I have a picture of the completely typical gate wiring I find whenever I look inside one of these cabinets. Utter disgrace in there.

There wouldn't be much load on the bowl of spaghetti!

You could always try squeeze in an IP rated RCD socket lol!

 

[James]

Now that is unfortunately not unusual for access control installations.

how people go home at the end of an install like that and be happy with themselves, I just don't know!!

 

[timhoward]

I think my approach would be to try and get an accurate Zs, as the length is a bit of an unknown making any attempt to introduce design into this a bit hit and miss. Current bets are it being longer than 400m based on your 4 ohm non-trip Zs.
So either dead tests, or stick any low-ish breaker in and a high current test, hence working out max breaker size up top. Fitting an RCD enclosure inside the gate box could be be the fun part.
It could always be an IP enclosure on the outside. In fact I might prefer the customer not to have to open that can of worms to reset it!