Put lamp in holder and the power dies... WHY?

[Ladders]

Hi all. So called out to a domestic properety as light fell off ceiling. Put it back together and power all shows good. L/E, L/N power 230v no reading N/E. But no power when connected to fitting? So swapped fitting for 6" pendant and did same? So took the lamp out of pendant and power readings back to normal? Any offers on what is going on? It looks like the lamp is acting as some kind of switch and cutting the power but who knows with out further investigation...

 

[telectrix]

bad connection upstream. with no load you are seeing the 230v. as soon as the load tries to pull current, the bad connection fails, dropping the volts. could be either L or N. easieat way to tell is use a voltstick at the fitting with load applied. if it goes out, the fault is on L. if not, it'son N

 

[KFH]

I had a very similar case last week. Customer complained that external PIR was unreliable. Lens was badly clouded so replaced PIR. Set up PIR and everything worked. Called customer to come and see if the was happy with sensitivity and range. Nothing worked. Suspect bulb in fitting so get long ladder out, bulb looks OK. Take PIR off and test voltage, nothing. Check fuses and lights on same circuit all OK. Customer says any connections are under bedroom floor which is sheet, not boards, with carpet on top and there is too much furniture to move. Can I run some T&E around the house to the light?

Still waiting to hear if he want me to run some proper cable around the outside of the house to it.

 

[telectrix]

and then what do you do about the fault under the floor. wait for a fire?

 

[Marvo]

If you use an analogue tester (a non-digital one that has a needle) you probably won't get the 230v reading with the lamp removed. This would make it easier to trace the fault.

 

[telectrix]

true, think that's because an analogue meter has a lower input impedance than the digital rubbish.

 

[needasparks]

and then what do you do about the fault under the floor. wait for a fire?

Thats the bit that tells you what board to lift. saves all that komplkated testing.

 

[Marvo]

true, think that's because an analogue meter has a lower input impedance than the digital rubbish.

Yep, much lower. For the kind of general electrical testing being done here they're better in so much that you don't get sidetracked by the tester picking up surface voltages that are meaningless.

 

[telectrix]

i used to get the p*ss taken out of my AVO8. not any more. always said these digi.meters were prototytpes for voltsticks.

 

[i=p/u]

maybe.. telectrix but how you gonna test rcd with analogue meter, the voltage tester comes free with rcd tester

 

[KFH]

"and then what do you do about the fault under the floor. wait for a fire?"

I have disconnected the live for the PIR at the switch. So I will have to use the cable tracer to see if I can find a jnc box rather than wait for the fire. Assuming he invites me back to sort it properly. He has paid me for fitting the PIR though.

 

[i=p/u]

if there is rcd in does this not detect anything before a fire could start

 

[telectrix]

as long as the power to the faulty cable/joint/whatever is isolated you can run a new feed external. pref in black flex. it will withstand the weather better than T/E.

 

[Marvo]

Digital testers have a time and a place but 99% of the time you don't need millivolt/milliamp accuracy for general electrical fault finding. Most digital tester users don't realise that with a device that has a very very high input resistance you're going to see a lot of voltages that are induced by other cables running in close proximity and the reading on the screen doesn't always mean what they think it does.

 

[needasparks]

if there is rcd in does this not detect anything before a fire could start

Not for a loose conection, its only when it goes to earth it shorts to earth during the fire that it might.