Fluorescent trips RCD frequently

[ppppenguin]

I moved a Fluorescent to a circuit protected by a 30mA RCD. The RCD frequently trips when starting.  Seems to be OK when running. The fluorescnt is a dual 5 foot, possibly an old Thorn. 500V megger says insulatbetter than 50M. 

Any ideas why?

[Refugee]

Try giving the tubes a clean..
Also the choke could be a bit iffy so that the mechanical effect of the starting due to the magnetic field adding an intermittent short.
They act like a series tuned circuit generating a fair lick of volts to strike the tubes so that is why the cleaning might help.

[boater sam]

2 chokes and 2 starters? If so its a "pop pack" .
Time to pension it off and go LED, it must be old and tubes will cost more than new fittings with LEDs.

[Nick]

Indeed - just replace it with something more appropriate...

FWIW, I did a couple of replacements on fluorescent strips and the quality of the new kit was generally poor - their failure rate was high

[ppppenguin]

Please excuse typos on OP. Was using a phone to make the post.

I haven't looked inside the fitting. I know it's decades old and scheduled for replacement later this year in a refurbishment. For now it will have to stay on a non RCD circuit.

Still don't understand how an earh leakage fault is happening. Unless insulation in the choke or elsewhere is failing during the starting peak voltage. What sort of voltage is generated when the starter opens?

I did once have a fright when turning off a fluorescent. There was a flash and loud bang, big spark at the switch and blackening on the outside of the switch where my finger had been. The best guess was that switching off the current causes a high voltage across the choke whose insulation broke down. This also jumped the gap in the switch and put a short across the mains. Still surprised that a 6A circuit could give such a loud bang.

[Doodlebug]

I use a dimmer on an old 100w lamp lighting cct.
Last week the bulb went and blew the dimmer !

Doesn't dim now, just full on or full off.

How does that happen?

[boater sam]

Heck of a big spike with the chokes, RCDs are often tripped by the spikes. My arc welder often trips on first connection and on disconnection but not when striking the arc.
We have up to 6 RCDs of various ratings from 30mA to 500mA in the line, it is still pot luck which trips, never understood why.
Some of the main ones are 3 phase, subs are single phase but there is no logic to it.
I am very confused about the currents in 3 phase and neutral cables, is there more energy dissipation in the neutral wire when there are 3 single phase supplies running. In other words, why are the neutral wires not thicker than the phases?
Thorn pop packs were very cheap fittings in their day.

[ppppenguin]

If theearth leakage is greater than the trip current of the highest rated RCD it's pot luck which one in the circuit will trip first. The only way to ensure discrimination is to use time delay RCDs for the higher rated ones.

I've forgotten most of my 3 phase theory but I think the following are true:

1: Equal loads (both resistance and reactance) - neutral current is zero. This is because the peaks on each phase never coincide. The posh way to work it out is to integrate the 3 sinusoids over a cycle period and find that the answer is zero.
2: Unequal resistive loads - neutral current can approach the phase current. Limiting case is one phase fully loaded, others unloaded.
3: Unequal reactive loads - it gets complicated but I think in extreme cases neutral current can exceed a single phase current.

Upstream of the LV distribution transformers all the circuits are delta wired with no neutral at all.

[boater sam]

That's what I understood Jeffrey, using all 3 phases on a 3 phase machine with a small single phase used for controls seems logical.
Where my brain stops is when all 3 phases are used with one neutral as 3 single phase circuits. This came to a head years ago when a health club installed 20 or more single phase sun beds fed by one 4 core swa cable carrying 3 live phases and 1 neutral.
After a while the cable got crazy hot even though it was conservatively rated as far as the loads on the phases was concerned.
The sparky came back and laid a heavy single core swa alongside just for the neutral. I could not get my head around this, I am still convinced that this approach is wrong.

[ppppenguin]

There is no difference between a 3 phase load and 3 equal single phase loads with a common neutral. There may have been some nasty power factor issues with the sunbeds. I've just found this article which says that harmonics are a big problem and can give up to 200% neutral current: http://www.cromptonusa.com/Neutral%20Current.pdf