IET logo
 
IET
Decrease font size
Increase font size
Topic Title: RCD trips - SON lighting feedback?
Topic Summary:
Created On: 26 November 2015 07:45 PM
Status: Read Only
Linear : Threading : Single : Branch
Search Topic Search Topic
Topic Tools Topic Tools
View similar topics View similar topics
View topic in raw text format. Print this topic.
 26 November 2015 07:45 PM
User is offline View Users Profile Print this message



alanblaby

Posts: 747
Joined: 09 March 2012

I was called to a strange one today, I was told the cable to outside SON lamps needed replacing as it was 'tripping'.
I had a look at the cable, it was old, but didnt look too bad, all conduit boxes were dry inside.I went inside the pub, to ask to turn off the power, and ask what happens.
'This one goes down' - pointing to the RCD.
OK, I isolate, and do some IR testing, which all came out fine - the lowest reading was 30Mohms.
Both fittings were new, so I wouldnt think it was them.
Ok, reconnect, turn on the lights, which came on fine, and leave them for 15 minutes - no tripping.
Go back in and the landlord was there, "they are working fine now - all tested and OK I said.
He then tells me the RCD trips when you turn off the outside lights.
Damn. I turned them off, and the RCD tripped.

It is a poor design, in that the CU has an RCD as the main switch, and all the lighting (7 circuits) on this one RCD.

OK, rather than going into a lot of testing, I said we'll get a separate RCBO for that circuit.

Now, I would like to have some views on what could cause the tripping when the lighting is turned off?
I changed the 2 pole grid switch, just in case. That made no difference.

So, possible tipping point of earth leakage from the multiple circuits?
Or, some backfeed/interference from the SON fittings when they are turned off?
But, why does it happen when they are turned off, rather than turned on?

Thanks
Alan.
 26 November 2015 08:09 PM
User is offline View Users Profile Print this message



perspicacious

Posts: 7854
Joined: 18 April 2006

Both fittings were new

Classic case of the installer not issuing a MEIWC

It is a poor design, in that the CU has an RCD as the main switch, and all the lighting (7 circuits) on this one RCD.

It could be something as simple as the circuit wiring being in close proximity to the RCD in the CU.

Regards

BOD
 26 November 2015 08:26 PM
User is offline View Users Profile Print this message



mapj1

Posts: 9387
Joined: 22 July 2004

I presume you have eliminated the possibility that it is a single pole light switch in the neutral, or a DP switch with the live side stuck, and its something much less obvious.

It may be that the RCD is nearly at trip point any way, and the extra few mA just tips it over.

-------------------------
regards Mike
 26 November 2015 08:36 PM
User is offline View Users Profile Print this message



alanblaby

Posts: 747
Joined: 09 March 2012

First thought was the switch, so it was changed, and didnt make any difference.
Yes, cumulative earth leakage would be close to the top of causes, but why would the tripping occur on disconnection?
I could understand when switched on, but I'm having trouble fathoming out the cause.
 26 November 2015 08:44 PM
User is offline View Users Profile Print this message



broadgage

Posts: 2389
Joined: 07 August 2007

I suspect that DP pole switch opens the neutral first, even if only by a minute fraction of a second.
If one of the light fittings has a high resistance fault between neutral and CPC then it wont trip the RCD when lit due to the low voltage between neutral and earth.
When full line voltage is applied, albeit very briefly at the moment of turning off, then the RCD trips.
To eliminate this as a cause, swap live and neutral over at the DP switch.

Another possibility, is that the interruption of one pole only of the supply is producing a substantial back EMF due to breaking an inductive load. This can trip the RCD due to capacitive or resistance leakage currents that would be negligible at normal voltage.
 26 November 2015 11:36 PM
User is offline View Users Profile Print this message



mapj1

Posts: 9387
Joined: 22 July 2004

If the lamps are inductivly ballasted, at the instant of switch off, the current will try and keep flowing (that being what inductors do), and to do so a large voltage will appear at the opening contacts (the back EMF) this will have a very fast rise time comapred to the normal 50Hz cycle, and will last until the magnetic field in the ballasts has collapsed to zero.
This sharp voltage spike may cause the contacts to fizzle a bit, but assuming the fittings have some EMC filtering, which normally involves a delta of capacitors between LE and NE, this spike will force current to ground via one or other of these capacitors. Generally these spikes are more on one pole than the other, so any RCD sees an imbalance between N and L. It is aggraveated by neutral side switching, as all the capacitors that are NE and were more or less flat, suddenly get charged to near line voltage as well. This alone can easily exceed 30mA..
The fix is a mixture of single pole live side switching, voltage clamping perhaps with a small clamp in each fitting accross the load side, and allowing plenty of margin on the RCD. In some cases a full blown Mains filter will be needed to dissipate the sudden switching transient into a much gentler voltage bounce of longer duration.
But try it on its own RCBO first..

-------------------------
regards Mike
 27 November 2015 08:16 PM
User is offline View Users Profile Print this message



alanblaby

Posts: 747
Joined: 09 March 2012

Thanks all, especially mapj1 - thats a good explanation.
Thanks
Alan.
 28 November 2015 08:21 AM
User is offline View Users Profile Print this message



aligarjon

Posts: 3854
Joined: 09 September 2005

I had a similar problem a few years back, i put it down to there not being a proper 3mm gap on the switch breaking properly. I linked the neutral as mapj1 suggests above and the problem was solved.

Gary

-------------------------
Specialised Subject. The Bleedin Obvious. John Cleese
 29 November 2015 05:33 PM
User is offline View Users Profile Print this message



jravenscroft

Posts: 73
Joined: 07 November 2015

Can anyone explain what is meant by voltage clamping? Searching the term brings up a lot of articles unrelated to this context.

How would you connect one of these varistors?
 29 November 2015 11:04 PM
User is offline View Users Profile Print this message



mapj1

Posts: 9387
Joined: 22 July 2004

The last time I answered a similar question for din rail SPDs, whixh work the same way, but are boskier.

here

It was perhaps over detailed.
To use them, without needing to know the physics, they connect accross the place where an overvoltage may occur.
Personally I use wire ended ones, and if inside a fitting or a metal back box, then I sleeve a cm or so of the legs and then push them into chock bloc, or push terminals.
Unlike the inrush devices that go in series, the voltage clamps go accross L-N at the load side of the switches and pass no current during normal voltage conditions,so they run cool. However, don't position it any where where if it blew up in a thunderstorm the bits would not be safely contained. Equally don't put it on an unfused supply. If a really big spike comes down the wire from outside, they can occasionally go dead short - and that is very dissapointing if some ADS does not come in quick enough.
It is important to look hard at the data, and select a device that does not conduct at all during normal use - so for 230V mains, I'd select a unit rated 250RMS, or maybe even 275RMS - which may be expressed in some circles as an AC peak or DC rating of 350-400.

Also look at the surge rating - it needs to be good for many times the worst case kick it may receive.

So for example, a 1 henry inductor passing 5 amps has a stored energy
of (half I squared, times L) 12 joules or so. I'd reach for perhaps a 50 or 100J unit for that load.
Something like a 100W lamp ballast may drop about half the mains voltage and pass an amp - so 100j ohms in very rounded numbers.
at 50Hz (as impedance is 2.Pi*F*L) thats about 1/3 of a henry.
so 1/3 of a joule to dump. Although a 1 joule unit would probably be fine, I'd be very tempted tofit a 10J or more unit in any case, as spikes that big can come in from outside.
Apart from cost and space, there is no penalty for always fitting the bigger ones, and it saves stocking loads of sizes.
note that an insulation meter will see the partial turn on as an insualtion fault, so mark where they are for the benefit of anyone who follows .

-------------------------
regards Mike
Statistics

New here?


See Also:



FuseTalk Standard Edition v3.2 - © 1999-2017 FuseTalk Inc. All rights reserved.

..