Loose Stat Earth Tripping RCD?

[Onoff]

Been meaning for years to tidy the wiring centre for the hot water / central heating! 

 

It's a traditional system, oil fired boiler with copper hot water cylinder. Everything as in the controls and boiler is powered via the downstairs ring from an fcu. Of late I've had some nuisance tripping of THE rcd. Consumer unit is fronted by a single 30mA RCD before a row of 9 Crabtree C50s. (Yes, I know, why haven't I changed it! Well, cu position is moving from the porch to inside with an all DP rcbo board but before I do that I need to block up a door, rip up the floors upstairs. All this will happen "eventually"...my eventually... ).

 

Did the usual of narrowing down to which circuit and identified the downstairs ring. This downstairs ring is the original wiring circa 1987 and prior all in T&E. As a quick check I even tested/ramp tested the RCD & the continuity and IR checks even cross connecting the ring ends and testing at each socket (save for two I couldn't get to) fearing rodent damage. All good. Found one (Had to pi$$ about taking the USB socket out as my IR tester doesn't have a 250V setting). Discounted the dw & wm that although in the kitchen are actually on this downstairs ring as opposed to the kitchen ring. Also the big fish tank.

 

That left the central heating / hot water system (no immersion btw). 

 

The system is a bit odd I think in terms of layout and wiring. Downstairs ch is still a 3/4" nominal bore single pipe system. When we bought the place there was no CH upstairs so I added 4 rads. I also added two contactors and two 2-port valves to split the CH from the 3-port valve. Something like the unfinished (photo of the) drawing below:

 

 

 

(As an aside from the thread the drawing might be the subject of another post; started it in Draftsight and working on it now in AutoCAD. Tried plotting as a web JPEG from AutoCAD 2010 but I must be doing something wrong as the quality is terrible compared to on the CAD screen. All I could do was take a photo on the phone. Anyone who wants a go with the .dwg and can tell me where I'm going wrong in the plotting process?)

 

Back to the fault & I narrowed it down to when EITHER room stat was turned and called for heat the rcd would trip. So either calls for heat, they each bring in the upstairs, downstairs or both contactors via the yellows onto A1 of each contactor. Contactor comes in and feeds the brown on the 2 port valve opening it. At the same time, the white from the two contactors, calls the 3-port valve.

 

The ONLY thing I found so far is one earth wire to a room stat had come adrift from the 12-way terminal block. Remade and no more tripping.

 

L,N & E for boiler and pump btw come off of terminals 6, 9 & 12 respectively. A few earths not shown either. The 7-core dives off to a pair of Towerchron QE2 programmers. One does Downstairs CH and HW the other just Upstairs CH. I'll work on the drawing to try and show all this. Apologies again for the poor image, makes it a bitch for anyone looking to help! 

 

Would a loose / detached earth cause this? I know the stat bumpf says it must be earthed "for correct operation".

 

Cheers

 

 

Edited September 5, 2018 by Onoff

[Jeremy Harris]

That's really weird, as a loose earth can't (or shouldn't) be able to cause an imbalance in the L&N.  I'm not even sure the room stats use the earth, do they?  It's only there for the cable protection I thought - at least it is for the room stats in our place - the earth at the stat end is just "parked" and doesn't make an active connection.

 

What could cause either an earth leak, or, more likely, perhaps, a voltage/current phase error spike, that would cause the nuisance trip?  Inductive loads can do it, but there's nothing there that's a big enough inductive load to do it, I'd have thought.  Maybe the contactors were causing some sort of induced spike in the stat cable, and connecting the earth up the cable was enough to reduce it?  Sounds a bit implausible, but stranger things have happened.

[ProDave]

Perhaps you need to call an electrician?

 

(I'll get my coat)

 

Most rcd tripping faults turn out to be a fault on a completely different circuit.  A partial N-E fault will cause an imbalance, and a DP switch that might close the L contact slightly before the N contact, will momentarily exagerate the earth leakage due to the partial N-E fault.

[Onoff]

53 minutes ago, ProDave said:

Perhaps you need to call an electrician?

 

(I'll get my coat)

 

Most rcd tripping faults turn out to be a fault on a completely different circuit.  A partial N-E fault will cause an imbalance, and a DP switch that might close the L contact slightly before the N contact, will momentarily exagerate the earth leakage due to the partial N-E fault.

 

And a plumber, SE, landscape gardener, roofer etc but mainly a TILER!

 

I had all the breakers off except this ring and the trip wouldn't reset whilst either room stat was calling. Everything was unplugged from the rest of the ring. 

 

That downstairs 2 port valve is a bit manky inside, I've just had the cover off. Got a spare luckily. The two 2 ports share N & cpc feeds. 

 

Going to sleep on it!

 

I'll disconnect that earth that was loose tomorrow and see if I can get it to trip again. Sure it's coincidental. 

 

 

[Ed Davies]

16 minutes ago, Onoff said:

I had all the breakers off except this ring

Aren't breakers usually just in the line, not neutral, so an N-E fault in another circuit wouldn't be isolated?

[Onoff]

48 minutes ago, Ed Davies said:

Aren't breakers usually just in the line, not neutral, so an N-E fault in another circuit wouldn't be isolated?

 

Yep unless they're true DP.

 

More testing on the other circuits I think just to be sure. What joy! At least the cu isn't wedged against the ceiling...oh wait it is! At least my eyesight's still good...oh wait it isn't!

[Onoff]

2 hours ago, JSHarris said:

Sounds a bit implausible, but stranger things have happened.

 

I'll give you implausible: Working on a ring once in an empty house. Old Wylex board.Fuse pulled and in my pocket. The biggest electrical storm right overhead the house. TN-S system & v close to the substation. I was changing the face place so had the two legs of the ring exposed and hanging out of the back box. A lightning strike and a foot long blue "spark" came out of the back box next to me. People don't believe me. Dread to think if I'd have been touching it all. Still got a pic of the old board:

 

 

EDIT: My bad just realised that's a MEM!

Edited September 5, 2018 by Onoff

[Temp]

If there is one my money would be on the switch in the motorised valve. The valves leak and water seems to get into the end of travel switch that controls the boiler. So what could happen is the stat calls for heat, the motorised valve moves and operates the switch and trips out the RCD.

 

Thiscisnt your problem but the switches can also fail so they don't make proper contact. In that case the stat clicks but the signal doesn't get to the boiler and the boiler doesn't fire up. 

Edited September 5, 2018 by Temp

[Onoff]

6 hours ago, Temp said:

If there is one my money would be on the switch in the motorised valve. The valves leak and water seems to get into the end of travel switch that controls the boiler. So what could happen is the stat calls for heat, the motorised valve moves and operates the switch and trips out the RCD.

 

Thiscisnt your problem but the switches can also fail so they don't make proper contact. In that case the stat clicks but the signal doesn't get to the boiler and the boiler doesn't fire up. 

 

Boiler fires up every time. 

 

Looking at the drawing I can see if that far left two port is getting wet then I could be bridging the neutral and earth inside. Even if that valve is not getting a call. Tbh I only drew it yesterday and it's a great help.