breaking light bulb trips other circuits

[tawraste]

hi all,
any idea why a breaking light bulb would trip other circuits mcbs?
system is TT, entire installation in an mk cu under one 80A 30ma RCD
bonding etc is correct and connected, pfc, psc etc all ok.
everything looks normal except:
all the circuits which tripped at the same time as the light are ring circuits
on 16A mcbs.
there is some difficulty in resetting at least one of these.
the light bulb that causes this is less than 1.5m from the cu.
i reckon the mcbs are tripping because they are rated too low/old and faulty.
but i do not understand why a fault on one circuit would affect another.

any ideas?

 

[malcolmsanford]

Is this a halogen downlighter??

An halogen downlighter does sometimes short itself on failure, especially cheap ones. My next door neighbour in the UK as them in the kitchen and when one lamp blew it took out the BS 3036 rewireable fuse, but only to the light circuit.

Now I have know lamps to take the RCD out, but not a row of MCB's, and to be honest can not think why a light circuit would take out socket ring circuit, even if it protected by a 16amp MCB.

I think your going to have to investigate this a little more to be honest. I would do the normal battery of tests IR, Continuity, R1+R2 etc and see what you find.

Has the ring finals as you call them, been split onto 2 16amp radials??

 

[telectrix]

could there be a borrowed neutral from the socket circuit to the lighting circuit?

 

[nickblake]

i have had the vibration of a tripping MCB trip others some of the weaker MCB like MK can do this , took me ages to discover this ,also if they are next to each other it could be the levers of the MCB catching had this too

 

[telectrix]

for what they cost, it might be an idea to replace the breakers.

 

[tawraste]

thanks for your helpful replies folks, i'm back there on monday and will post how i get on.
to clear up previous questions - its a normal bulb not a downlighter, and the ring circuits are not joined radials.
the mcbs themselves are not even next to each other, they are spaced out along the row.
so here's my plan of attack:
start with the basics, test everything to infinity.
replace questionable mcbs - one of them has trouble latching anyway.
look closely for borrowed neutral.

if the tests show nothing and there's no borrowed radial i shall put it down to old/faulty mcbs.

 

[tawraste]

hi all,
got up there today and took a look. unfortunately time was a factor for the client and i only managed to look at the lighting circuit.
the whole house is on one lighting circuit. polarity and continuity and IR tests came back ok. linked the cpc and live and got good readings for the (R1+R2).
there is definately no borrowed neutral from the ring circuits or anywhere else.
only thing i noticed inspecting all the connections was a slightly loose screw in the neutral bar of the CU that had caused a little softening of the insulation on the lighting circuit, but as above, IR results were good. i changed the mcb for a new one anyway.
i'm getting back up there soon to look over every other circuit, because whilst the IR was good for the lighting circuit, the entire installation came back at 0.01Mohm...... so there is another problem somewhere else...

but this did get me thinking about maximum Zs for MCBs. taking the (R1+R2) value of 1.34 and adding the 21ohms for TT gives you 22.34ohms, which does not follow table 41.3. i assume the compliance for disconnection time is simply satisfied by the RCD?

cheers, will keep posting what i find.

 

[simonspark]

I have found that some mcb's when they trip causes others to go because of the electro-magnetic surge affecting them. Sometimes when a lamp is near the end of it's life, the large amount of tungsten deposits in the gas sort of turns into a super conductor for a split second. This is enough to trip the breaker.
I once had a 100 watt lamp in a bathroom blow up the pull cord and take out the main service fuse

 

[somersetsparks]

your 0.01 Mohm reading is almost certainly an aplliance of some description plugged in somewere or the fly lead of an rcbo still connected. I always find the old favorites for bad readings are glade plug-ins or equivalent, or cookers.
God knows how the glade thingys do it but they even give bad readings L-E and N-E even though they just have a plastic earth pin.

 

[tawraste]

i reckon you may be right, even though i asked them to unplug everything i didn;t check with my own eyes as it quickly became obvious i'd only have time for the lights. nice one!

 

[malcolmsanford]

hi all,
got up there today and took a look. unfortunately time was a factor for the client and i only managed to look at the lighting circuit.
the whole house is on one lighting circuit. polarity and continuity and IR tests came back ok. linked the cpc and live and got good readings for the (R1+R2).
there is definately no borrowed neutral from the ring circuits or anywhere else.
only thing i noticed inspecting all the connections was a slightly loose screw in the neutral bar of the CU that had caused a little softening of the insulation on the lighting circuit, but as above, IR results were good. i changed the mcb for a new one anyway.
i'm getting back up there soon to look over every other circuit, because whilst the IR was good for the lighting circuit, the entire installation came back at 0.01Mohm...... so there is another problem somewhere else...

but this did get me thinking about maximum Zs for MCBs. taking the (R1+R2) value of 1.34 and adding the 21ohms for TT gives you 22.34ohms, which does not follow table 41.3. i assume the compliance for disconnection time is simply satisfied by the RCD?

cheers, will keep posting what i find.

Son that is not how you should be calculating for Zs unless you actually got the magical 21ohm Ra. Your fault finding on the lights is good, but not sure if your getting yourself muddled with Zs.

The 21 ohms is a value the network must maintain at their earth point at the transformer, we must for a TT system take the Ra reading of the earth electrode we install, and that is the one we use for calculating Zs.

 

[tawraste]

thanks malcolm,
yes the zs is getting me muddled - my tutor used the 21ohm in examples, explaining it away with the RCD disconnect times; hence me putting it up here to see what others think.
yet again my understanding is advancing thanks to the help of this forum and its members.