RCD tripping

[SUNRAY]

We (A company I sub to) installed the dis board in building 4 about 5 years ago, along with the pillar and ducts with split con to the other 3 outbuildings, until then they were fed with T&E suspended on catenary wires.
Not all of the work was done at the same time.
At the time I commented there were too many RCD's and suggested either RCBO at building 4 or RCD's in 1 to 3 but not both.
Since then there has been a small history of 'false tripping' but no significant.
All summer there has been a significant increase in tripping, including the main 500mA, and multiple visits had found no faults. The earlier visits had not resulted in a proper log but recently:
1) Building 2. - RCD tripped twice, RCBO tripped 3 times but not both - tripping times measure virtually the same on opposite ½ cycles so no surprise
2) Building 1. - 3 times RCD & RCBO & Main tripped.
3) both 1) and 2) of the above combined.
I've been on the last 2 visits and found scenario 1) and 3) on my second visit I added a 3ph RCD to building 1, running P & N through twice to make it more sensitive.

So far, amongst other things, I've checked all RCD's & RCBO tripping times. Ironically the only one which has failed and been replaced with a 2pole isolator is the toilet building 3.
Insulation across the whole system from the C63 MCB shows as ∞ with the 2 PIR's (buildings 3 and 4) disconnected and 5ft fluo unplugged (fluo measured as ∞ to earth).
All of the 'Z' readings are excellent.

Hardly a DIY problem but maybe another idea could help.

EDIT: While writing this I received an email saying the 'new' RCD for building 1 has tripped and reset OK.

 

[Colin Brenton]

If all the wiring tests OK, you may have noise on the mains coming into the building that is tripping random breakers. Is this in an industrial area? I'm thinking things like welders, large motors etc.

Might have to bring in someone with some equipment who can do some power quality analysis... It's not gonna be cheap!

Good luck

 

[SUNRAY]

If all the wiring tests OK, you may have noise on the mains coming into the building that is tripping random breakers. Is this in an industrial area? I'm thinking things like welders, large motors etc.

Might have to bring in someone with some equipment who can do some power quality analysis... It's not gonna be cheap!

Good luck

I'm afraid I was thinking along the same sort of lines but to take out all 3 RCD's in a row (One could be over current) seems unlikely.
This is a soft fruit farm with nothing but a small village and farms all around it. I assume there is an amount of engineering works on the farms but probably not on this current basis.
The sub station (just a HV transformer) is only 50ishM from the intake and I'm assuming it only feeds 2 farms and 3 houses as the next property is about 1/2 mile away in the village right beside a 1000A sub and in the other directions at least 2 miles to a group of properties.

 

[ericmark]

The 200 amp means can't use consumer units so in the control of an instructed person, so you can use a RCD with a delay, seem to remember where I worked the moulded breaker was connected to a RCD with two dials so up to 1 amp and 5 minutes delay, clearly this only fed further distribution or consumer units nothing directly fed from that RCD, and it would feed a few boxes with 100 mA S type which in turn fed portacabins with 30 mA and when some one knocked a nail through the cable in the portacabin it took out all three RCD's so even with time delay you can still loose all supply with no discrimination.

But much is down to who can reset a trip, the above was when 16th Edition was running, and even then all distribution units over 100 amp had to be locked only consumer units can be reset by ordinary persons. Been out of commercial for too long now, don't know who would be permitted to reset? I remember we trained fitters so they were considered as instructed person so could isolate and reset supplies.

 

[SUNRAY]

Well that hopefully is one problem located, by chance this morning someone was in the building 2 when the light went off and he went outside straight away to see if anything had changed. There are a coupe of outdoor DSSO's in polytunnels fed from the 32A socket, one of the workers plugged her Ipad in to listen to music and it seems she regularly does, a colleague has been round and PATed the psu to find a low resistance N-E.

 

[SUNRAY]

The 200 amp means can't use consumer units so in the control of an instructed person, so you can use a RCD with a delay, seem to remember where I worked the moulded breaker was connected to a RCD with two dials so up to 1 amp and 5 minutes delay, clearly this only fed further distribution or consumer units nothing directly fed from that RCD, and it would feed a few boxes with 100 mA S type which in turn fed portacabins with 30 mA and when some one knocked a nail through the cable in the portacabin it took out all three RCD's so even with time delay you can still loose all supply with no discrimination.

But much is down to who can reset a trip, the above was when 16th Edition was running, and even then all distribution units over 100 amp had to be locked only consumer units can be reset by ordinary persons. Been out of commercial for too long now, don't know who would be permitted to reset? I remember we trained fitters so they were considered as instructed person so could isolate and reset supplies.

This 500mA RCD is adjustable almost as you describe (currently set to 500mA and 10 second I believe) and inside a window enclosure with labels for authorised access only. It has to be unscrewed with an ordinary flathead screwdriver. It feeds directly into a 12w 3ph board, which also has authorised access lables, in which the cover over the MCB's is lockable but the key is left in the lock
I have only shown the sub main which appears to be causing the tripping but of course we may be barking up the wrong tree entirely.

 

[ericmark]

one of the workers plugged her Ipad in

I had similar problems, a machine was doing daft things, so banned mobile phones, the problem became worse, found worker hiding phone in electric panel.

 

[SUNRAY]

I had similar problems, a machine was doing daft things, so banned mobile phones, the problem became worse, found worker hiding phone in electric panel.

I've had problems of RF interference from all sorts of radio kit, and particularly phones, into sound systems and BMS circuits over the years, some of which have been a right PITA to identify and resolve and it most certainly doesn't help when people do that sort of thing.

I had a job where the customer kept complaining a BMS panel kept resetting, it took a number of visits and i suspected something like that was happening so I added an additional contact on the door interlocked isolator which operated a sounder in the control room, from that they were able to monitor the guys work times as he opened the enclosure when he got to his work place, before and after each tea and lunch break and at the end of the day, next thing added were a pair of cctv cameras, one inside the enclosure and one in the air vent above him. He was already on written warnings...

In this case it is simply a PSU with an earth pin, failed with E/N fault.

The neutral on this farm does typically have a couple of volts to earth so it is likely to be a bit sensitive to this type of fault

.

 

[Colin Brenton]

Good stuff, I'm glad you found it. Nice to hear what the problem was

 

[SUNRAY]

Good stuff, I'm glad you found it. Nice to hear what the problem was

I don't think we are anywhere nearer the problem in building 1 yet as that person wasn't working yesterday so I'm told.
And building 1 has tripped again today since the PSU has been removed from service.

 

[SUNRAY]

Well it looks like we have got to the bottom of this.

Today 2 of us arrived to find the additional RCD tripped, we insulation tested, looking into Building 1 CU and found around 78MΩ to earth on neutral in damp conditions and we quickly proved it onto one of the lighting circuits.

Despite 'all' of the wiring the in building 1 being inspected and tested several times by different people, this had not been found before, it was quickly proved onto a lighting circuit and we chose to do a finger. We found a RCD FCU on the lighting circuit (Complete with 13A fuse of course) which we didn't know about. It feeds a 500W PIR flood near a gate some 30m from the building and the other side of a row of trees, the plastic bodied light is mounted on a scaff pole which in turn is clamped to a scaff stub concreted into the ground. As such there is no 'earth' connected to the pole and the light works perfectly. We soon established the pole was at neutral potential and 1.12Ω trom neutral in the CU.

So there was a fault where the bolts either side of the fitting for the mounting bracket pass right through the lamp casing and one had punctured the neutral insulation when tightened, connecting neutral to the pole. So the question then was: how was that making the RCD's trip? and why all 4 (including this recently discovered device)?

It astounded me when I tested the 'main switch RCD in the CU' for the unpteenth time that no one, including myself had checked that it was indeed 'off' (open circuit) when tripped, as the neutral pole was found to be a solid 0Ω in either state. Additionally the C32 RCBO and the RCD FCU are both single pole versions which meant the lamp scaff pole remained a steadfast neutral right back to the 500mA main incomer RCD.

Of course every time the building was tested the RCD FCU was tripped and not detected on the phase conductor, I'd not seen the leakage before and as yet I don't know if others have.

So getting back to why had it been tripping was casually found while making the repairs to the lamp. and one of the farm workers opened the metal gate and hooked the chain onto the scaff clamp on the light pole, as they always did. The metal gate meatured 17Ω to wet mud.

I was involved with rewireing the building some 5 years ago or so and knew exactly what was in place then, in the mean time every one is saying there has been no extra work done.

Which just goes to show one must never ignore the probability of people being wrong and additional works being performed without being recorded.

It didn't help that the RCD FCU is in an obscure place and not easily seen where there was originally a loop of slack cable where 3 beams come together.

I hate to think of the man hours spent to find this one but the log shows 14 visits and 23 people and my 4 visits alone amounts to 13 hours.

EDIT: we did restore everything back to original to prove the gate was the cause and demonstrated it to the manager before the final repairs.