Motor Tripping RCD but it's not even connected to it

[JD Hogg]

@JD Hogg If it is a TT system how can you have circuits not connected to an RCD?[/QUOTE]


There is a 100mA RCD before this board but this isn't affected.

 

[Wilko]

Just trying to understand the set up - has this board itself got its own means of earthing, or is it derived from the main board ?

 

[Pete999]

It is a TT system needs rcd for fault protection.

Yes I appreciate that Westy, but the OP has mentioned in his post that the cb for the motor is fed from the main switch, or I think that's what he means#10.
Streamer, the reason I asked about the high integrity board was the OP mentions the Motor cb is fed from the MS, I was asking mainly because the OP may have adapted the board to suit his needs, it has been known to happen .
If he has then the borrowed neutral theory does hold up despite the OPs assurance that there is not a borrowed N, OP I'm not having a dig, just can't get my head around the problem.

 

[Pete999]

Is it possible to move the problem circuit to between the Main switch and the 1st RCD? might stop the heavy load affecting the RCDs. Or possibly fit another small non RCD ccu just for that circuit?

Loading should not trip an RCD.

 

[Billsays]

Loading should not trip an RCD.

True but RCDs can be twitchy things when motors are around. It was just a suggestion. Your borrowed N idea doesn't hold up as the OP said the RCD can be reset once the motor is running. Its just the start up that trips it.

 

[westward10]

Do you know the exact type of split phase motor it is.

 

[marconi]

My early thinking is along the following lines:

Worth checking the motor IR and its supply IR to confirm all is well even though it is a new motor and new installation.

1. DOL motor start - large transient current through L and N.
2. TT system so N not hard wired to E in installation so N can change in potential with respect to E/cpcs.
3. 1 above means all L conductors throughout installation will experience a transient drop in potential with respect to E/cpc. This will cause a current to flow through distributed capacitance between L conductors and E/cpcs and earth mass.
4. 1 above also means all N conductors will experience a transient rise in potential with respect to E/cpcs and body or earth. Again this will cause a current to flow through distributed capacitance between all N conductors and E/cpcs and earth mass.
5. The net result of 3 and 4 is an unbalanced current through the RCDs, and in the case of one of them sufficient for it to trip. That RCD may also be 'biased' already towards tripping because of earth leakages in the final-circuits it protects.

What to do?

a. Can you discover insulation resistance faults on the tripping RCD's final circuits? Or damp, dirt, dust, detritus near live conductors which can be cleaned away.
b. Check all L and N connections are well made and low resistance. Poor connections can create transients only noticeable when high currents flow through them.
b. Can the installation be converted to TNC-S to tie N and E/cpcs?

 

[SparkyChick]

Just out of interest... is the MCB for the motor that causes the tripping next to the RCD that trips?

 

[Marvo]

Some of the cheaper RCD's can be twitchy with phase angle change from non-linear loads like motors.

Another possibility is the motor is causing a voltage spike that's being sunk to earth by surge arrestors on a completely different circuit which can get RCD's all emotional especially if they're earth referenced.

Before you get complicated, the first thing is to to do the usual barrage of tests on the RCD, IR all the circuits fed from the RCD, check for borrowed /crossed neutrals etc. Also IR test the motor @ 500v and spin the motor by hand whilst your IR testing it.

 

[westward10]

I have no idea why people think this is a cross neutral issue.

 

[SparkyChick]

I don't... I think it could be as simple as the quick trip magnetic coil in the MCB interfering with the toroidal inductor in the RCD. Hence my question about placement of devices

But I'm happy to be wrong as then I'll probably learn something

 

[westward10]

I don't... I think it could be as simple as the quick trip magnetic coil in the MCB interfering with the toroidal inductor in the RCD. Hence my question about placement of devices

But I'm happy to be wrong as then I'll probably learn something

Not impossible I have had rcbos tripping the adjacent device on Memshield 2 boards, rcds seem to defy logic.

 

[Wilko]

Agree Marconi idea re TT vs TNCS - tempted to have a "never have I ever" test with rod in place etc.

 

[Strima]

I don't... I think it could be as simple as the quick trip magnetic coil in the MCB interfering with the toroidal inductor in the RCD. Hence my question about placement of devices

But I'm happy to be wrong as then I'll probably learn something

You stole my idea from post #6...

 

[JD Hogg]

I don't... I think it could be as simple as the quick trip magnetic coil in the MCB interfering with the toroidal inductor in the RCD. Hence my question about placement of devices

But I'm happy to be wrong as then I'll probably learn something

When I originally installed it the motor MCB was next to the RCD that trips and was in fact fed from the RCD. When we started the motor and the tripping issue arose I moved the MCB away from the RCD and as stated previously I have fed it from the main switch.

It is an MK metal clad board so not particularly cheap and nasty. The motor is made by Wonder which I think is a chinese mfr and the starting current is 110 Amps. Also as previously stated there isn't a crossed neutral and I have IR tested all of the circuits.

 

[JD Hogg]

Also I think it is very unlikely that PME will be available.

 

[SparkyChick]

You stole my idea from post #6...

Well it was the reason I asked for a picture of the board in post #3

Seems this may have just been ruled out by the OP though so we're both probably wrong

 

[Lucien Nunes]

Um, have we had an answer yet to the most important and obvious question, which if I have read the thread right only appeared with Marconi's post #27:

a. Can you discover insulation resistance faults on the tripping RCD's final circuits?

The starting current of the motor may be the highest peak load on the system, hence the only thing that shows up an existing N-E fault on a final circuit protected by the RCD that trips. It is just possible that a surge protector on one of those circuits will cause the same effect. I would have no hesitation in making tests on the tripping RCD's circuits, not the motor circuit, my next task...

 

[Leesparkykent]

Is the upfront 100mA RCD time delayed?

 

[marconi]

I don't have a photo of it but from left to right it is main switch, RCD 1 63A 30mA, 5 MCB's RCD protected, RCD 2 63A 30mA (this is the one that trips), 3 MCB's RCD protected, 32A MCB feeding motor, 16A MCB feeding different motor both 32A and 16A MCB's fed directly from main switch. Mainly old installation, new board, installation fully tested all OK, 2 new circuits for the motors. No borrowed neutrals, definite separation between busbars. Even if the three circuits breakers fed from RCD 2 are switched off the RCD still trips when the bigger motor starts, nothing trips when smaller motor starts. TT system. Brand new motors.

Opening the 3 mcbs will leave the Ns and Es connected so some electrical characteristics of the final circuits remain in situ fed from RCD2.

I suggest:

a. you completely disconnect the LNEs of the final circuits fed by RCD2 and then observe the behaviour of RCD 2 when you start the motor.
b. check the IR of these circuits one by one.
c. connect one final circuit back - operate the motor - observe what happens. Then completely disconnect it.
d. repeat c) for each of the other final circuits.

This little method might help you discover if it is one final circuit fed from RCD2 which is interacting with the motor to trip the RCD.

As Marvo and Lucien Nunes said - are there any transient filter sockets, computers, washing machines and the like which contain filters connected across LNE? If there are see what happens when you completely disconnect them.