Fluorescent Lighting Surge

[MrDuck]

Well I'm getting closer! The theory about metal cased capacitors is a non-starter, as mine are all plastic dating from 1987.

However, by a process of elimination involving taking out starters I have been able to identify which tube and ballast are causing the unwanted tripping. I then discovered that the ballast had been badly installed, trapping three single core wires between it and the metal chassis of the light. The insulation on these wires had overheated badly over the years from the heat of the ballast, and become very thin. I couldn't see bare copper through the thinned insulation, but a breakdown causing an earth leakage fault under start conditions seemed entirely possible. So I reinsulated the damaged wires and reinstalled the ballast, with high optimism of a fix.

Some hope - the tripping started again immediately! My next step was to remove the ballast again and leave it suspended by its wires in free space with the cover off but the lamps and starters installed. Now all is well and tripping no longer seems to occur. So the ballast is twice cursed, once for damaging the wiring and now, it seems, for having an earth leakage fault under start conditions.

Is a defective ballast a reasonable possibility? I can't measure any leakage with test meter, but don't have Megger to do a high voltage test.

 

[JohnW2]

Well I'm getting closer! ... However, by a process of elimination involving taking out starters I have been able to identify which tube and ballast are causing the unwanted tripping. ... Is a defective ballast a reasonable possibility?

There really is very little to a simple fluorescent fitting - so I would think that, if the wiring is now OK, there's not much left to point a finger at other than the ballast.

Kind Regards, John

 

[MrDuck]

For now I've 'fixed' the problem by reversing the connections to the ballast so that the end that was earth leaking is now connected to 230v and vice versa. How long this will last I don't know! I will pick up a spare ballast in due course before they are finally obsolete.

Many thanks to everyone who has contributed to solving what has been a very confusing fault. I am most grateful.