400w metal halide cycling.

[stummish]

Hi all.

Yesterday I replaced 6x 400w metal halide fittings and lamps in a volvo dealership workshop. The existing fittings were tatty and old, so it was a straight swap old for new. After switching on all was well, until today when I got a call saying one had turned off, I asked my customer to keep me informed about if it came back on after it had cooled down. Sure enough it came back on, although apparently slightly brighter than the others. Since then 2 others have done the same, at one point all 3 turned off at once.

Does anyone have any ideas why? They are split between 2 circuits, the 3 problem fittings are on the same circuit, but I can't see any reason for this to be a reason. I've heard various reasons, lamps having a bedding in period, a faulty capacitor allowing too much current to the lamp.. and the standard faulty lamp and faulty fitting reasons, just a 50% fail rate seems unlikely.

Any input is welcome

 

[Darkwood]

Have you been supplied the correct lamps to go with the ballasts?.... Wholesalers sometimes mess up.
Are you getting a volts drop for some reason...bad joint etc?

 

[stummish]

I'm there again on Thursday to have a look, and I'll take a look at the invoice tomorrow to get an idea if they are the right lamps. As for volt drop I'll just have to take a reading on site on Thursday. I just assumed the new accessories were the problem, and the wholesalers would supply the correct gear.

Assume - Makes an a$$ of u and me.

 

[UKMeterman]

Sounds like a car dealership, do the lights go off when the car ramp starts up? DOL starter causing a big voltage drop, causes the arc to extinguish... Same thing happened at a large uni, very time the compressor started up on the library aircon chiller the street light went out for 5 mins...

 

[stummish]

There are 5 ramps and a compressor, I was hoping for a nice easy job, swap light fittings collect money and go. Oh well, I'll know by Thursday.

 

[stummish]

Update:

So I re terminated all the kliks, nipped up the conductors in the board, replaced the lamps. No effect, same fault 2 hours later, all 3 on the circuit cycled again. 2 circuits of 3 lights each, 1 circuit good, one bad. This only left 2 variables, fittings on circuit a were lower in the pitch of the roof than circuit b, and they are on different phases. So, in a fit of madness, I swapped the line from circuit a to mcb b and vise versa. Lo and behold, 2 hours later the same fault but on the other circuit. I'm there to swap the mcb tomorrow, this is the only conclusion I can make from this, aside from some problem with the L1 phase I don't understand.

Any more input would be great if you can think of it. Thanks for the help so far.

 

[Darkwood]

Update:

So I re terminated all the kliks, nipped up the conductors in the board, replaced the lamps. No effect, same fault 2 hours later, all 3 on the circuit cycled again. 2 circuits of 3 lights each, 1 circuit good, one bad. This only left 2 variables, fittings on circuit a were lower in the pitch of the roof than circuit b, and they are on different phases. So, in a fit of madness, I swapped the line from circuit a to mcb b and vise versa. Lo and behold, 2 hours later the same fault but on the other circuit. I'm there to swap the mcb tomorrow, this is the only conclusion I can make from this, aside from some problem with the L1 phase I don't understand.

Any more input would be great if you can think of it. Thanks for the help so far.

Or there is a single phase possibly motor starting dragging the mains down on that phase

 

[Richard Burns]

Possibly the cable termination is dirty on that MCB or probably the MCB is just missing the bus bar and making poor contact and slowly overheating.

 

[stummish]

Termination of the mcb on the conductor and bus bar is good. The fittings cycle within 5 mins of each other, and after roughly 2h15min each time. My thoughts were a faulty mcb? Never come across this myself, but a plastic connection inside the mcb is possible?

 

[Lucien Nunes]

If there were an intermittent high resistance on the lamp circuit, whether due to a loose termination, poor contact in the MCB or whatever, with maybe 1500 circuit watts I would usually expect to see some heating taking place because it's unlikely to remain at a stable low resistance for all the time the lamps are alight. I would also expect to be able to provoke it by tapping / wriggling at the board. Have you got a max/min logging function on your DMM that you could use to test Darkwood's theory, that something is pulling that phase down? If you connect to another DB way (so that nothing on your own circuit can affect the reading) and wait for the lights to extinguish, then check the minimum...

 

[Rigpig]

Hello stummish,

Had this same problem at our work.
load fluctuation was causing the bus to dip below 11kV and 60 Hz . This lead to floodlight s out all over the bloody ship.
As a relatively painless fix we purchased dual strike lamps which meant that we didn't have to wait for the lamp to cool down before turning them on and off which restarted them.
As of last hitch we have abandoned all our glamox floodlighting and are now changing them out for LED's instead......

It might be worth getting the mechanics to start any ramps or similar while you are present to see if this is causing the lights to go out.
cheers
Rigpig

 

[stummish]

The terminations are all a-ok. Mcb, kliks, fittings all checked and any even slightly suspect ones re terminated. I've checked for volt drop at the end of the circuit, and forced the compressor to start running, and 2 of the 6 ramps to be starting to see if there is volt drop due to load, the most it dropped was a few volts. That and the fact that the cycling period seems to be a fairly regular 2h15min makes me suspect the mcb as it's the only thing left as I can see it.

 

[Darkwood]

Failing lamps cycle at regular intervals too you may have good a bad batch, the VD may be due to external issues so be wary of that, other industries tapped to the same distribution network transformer may be causing your problems - the fitting you put in may not have the same level of tolerance as the old ones.

The easy cheap option is to replace you mcb - faulty ones are not unknown but are not common.

 

[stummish]

I've just left site and replaced the mcb, and moved it from L1 to L3. I guess I'll get a call in a couple of hours to say if it's worked or not. Can't wait for that call!