Capacitive Inductance

[Arewehere]

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 spur off a DSSO in the main house via an RCD spur and provide two outlets by the pond for the pump and UV.
Pretty straightforward !!
All connected, power up and RCD will not stay in, sometimes tripping the main house RCD, I know two 30mA trips in line is not ideal but it was specified. So OK I think I've made a mistake, take it apart piece by piece and check, using the insulation tester as I go.
I finally am back where I started with every end open and everything showing as close to infinity as makes no difference. I spent ten years of my career on T&R contracts with a meter in my hand all day so I'm reasonably confident with it, but this had me checking and rechecking, but the figures remained the same for cable and appliances.
All put back together the pump and light will run all the time the earth of the SWA and Sheath are not connected, as soon as they are it will trip the RCD spur as if the is a fault, but the meter will show none at 500v.
It can only be capacitive inductance, the system needs an earth as it's a buried SWA and the pump and UV need to be protected for users safety obviously.
This leaves us in an awkward situation as the cable has been buried and landscaped over some time ago, but the feature has to work.
Has anyone come across the problem of capacitive inductance and found a safe and effective way round it.

 

[stevethesparks]

You say you have two RCD's (in-series) is it the same RCD that trips every-time or both?
If just the one, have you tried replacing the RCD?
In my experience, a ramp test does not always identify a faulty RCD.

 

[stevethesparks]

By the way, who specified two RCD's? Have you tried removing the second RCD at all?

 

[Arewehere]

PS I think Grommit you may be onto something with the different earth potentials, one of the things I'm taking with me is a stake and once I've checked for all other faults, I had thought just to take L+N from the house, use two RCD spurs and they can run as a TT system down by the pool.
Lucien seems to have written out the capacitance idea, grabbing at straws really, so two different potentials does seem the most logical way to look at it.

 

[Arewehere]

PS I think Grommit you may be onto something with the different earth potentials, one of the things I'm taking with me is a stake and once I've checked for all other faults, I had thought just to take L+N from the house, use two RCD spurs and they can run as a TT system down by the pool.
Lucien seems to have written out the capacitance idea, grabbing at straws really, so two different potentials does seem the most logical way to look at it.

 

[telectrix]

start off by removing the RCD FCU. or supply from a non-RCD way in the house CU. still tripping? got to be the SWA. as previous post, a fault may not show up on a D.C., low current IR test, but bites your arse when you stick a higher current capable A.C. voltage.

 

[Octopus]

PS I think Grommit you may be onto something with the different earth potentials, one of the things I'm taking with me is a stake and once I've checked for all other faults, I had thought just to take L+N from the house, use two RCD spurs and they can run as a TT system down by the pool.
Lucien seems to have written out the capacitance idea, grabbing at straws really, so two different potentials does seem the most logical way to look at it.

Did you check the continuity of ALL the cores and the wire armour before doing the IR?

Too many people do R1 + R2 , but few seem to do Rn + R2 ...... which can be very revealing

 

[Lucien Nunes]

onto something with the different earth potentials

Seems unlikely. The RCD can't sense earth potentials, only the imbalance between line and neutral currents, two conductors that you claim are effectively insulated from earth.

If there is a difference, there is leakage, but it matters not where the leakage takes place or where it goes; The circuit's own CPC, another earth near neutral potential, a rogue 'earth' at some other potential, another circuit's neutral. It can be inward leakage from another circuit's line to the protected circuit's neutral, or even into its line from another line on a different phase. Any or all of these combine to give a particular figure for imbalance which is what the RCD detects. A few volts difference between one leakage 'destination' and another won't affect that total more than a gnat's.

 

[Shoei]

As Lucien explains, the only way the RCD could trip is if there is a current path from L or N to earth, that's all the RCD can detect. The problem is investigating the fault when the RCD trips right away.

From your first post, all is well with the CPC connected with sheath isolated, but the RCD trips when the sheath is connected to earth, is that correct? It's weird that should happen while you can't find any leakage with the meter at (say) 500V sheath-L, but that must be where the current is going (also possibly sheath-N).

You could try connecting the sheath to the CPC through a 10k
Ω resistor, value isn't critical but it should be rated at 500V+ and be large enough to avoid tripping the RCD. That would let you check the circuit under fault conditions without losing the supply. You may get an indication of what's going on by measuring the voltage across the resistor. 230V would mean a low resistance sheath-L. A small voltage could indicate low resistance sheath-N which could be confirmed by disconnecting N from the supply, when the voltage would drop to zero.

 

[Lucien Nunes]

I have to say I am very much with Murdoch's line of reasoning at the moment. Have you with 100% certainty identified the conductors throughout, and proven that the core connected to the armour is indeed the CPC and not actually the neutral due to a discrepancy in labelling at the two ends?

 

[Intoelectrics]

@Arewehere,

please do post your findings, even if it turns out to be an error on your part. We all make mistakes so don't worry about any ridicule.

But it would be interesting to us if it turns out to be some gremlin in the system.

Look forward to your findings