Metal Halide Lamp Flicker

[gafferted]

Hi all

 

 

I'm doing up some old Futurelight MH640 wash fixtures and have just put a new Sylvania BA250/2 lamp in one. It works fine but there is a noticeable flicker at, I presume 25 or 50 Hz.. I've not noticed this an any of the other fixtures, although I do have a Martin Pro518 on which I've noticed the same thing occasionally.

 

Before I start swapping ingnitors/ballasts etc with other fixtures I though I'd ask if anyone knows any common causes for this. I have to say I've also not run the unit for any significant length of time to see if it goes away when the thing is up to temperature.

 

Maybe it's quite normal for them to flicker slightly when cold?

 

Any advice appreciated.

 

Eddie

[greenalien]

No, it's not normal - can be an indication of ballast on its last legs.

[niclights]

In my experience this can be normal for new lamps. Personally I'd give it a while before I started to worry.

[greenalien]

The problem with 'waiting to see if it goes away' is that, if the ballast is faulty, the lamp can fail prematurely, as it's the ballast that 'balances' the current feed to the lamp. Not cheap if you waste a lamp. Swapping ballast need only take a few minutes - you don't need to remove anything, just put another similar fixture alongside the suspect one, disconnect the suspect ballast and wire in the ballast from the known good one - if the flickering goes away, it's the ballast that was faulty; if not, it's something else...

[gafferted]

Thanks for the tips guys.

 

Is the ballast simply a coil? If so I suppose the only thing that can go wrong with it is that it could over heat and some of the windings could short out, resulting in too much lamp current.

 

Am I thinking along the right lines? Is it worth measuring the lamp current and voltage to check they are correct for the lamp? From the datasheet the Sylvania lamp is rated at 95V / 3A. If those are correct is it safe to assume the ballast is OK?

[KevinE]

The easiest thing is the swap the lamp, they do sometimes flicker 'til they've bedded down, it sometimes happens to MH industrial floodlights too. I woudn't worry too much.

[greenalien]

Is the ballast simply a coil?

 

There are 2 types of ballast - reactive, which is basically a big coil of wire around an iron armature, like a one-sided transformer, and electronic, which is a bunch of solid-state circuitry. Reactive ballast may also have a capacitor across it to improve its power factor and this is also something that can fail.

[gafferted]

Thanks for the tips one and all

I've measured the ballast with an (admittedly cheap) inductance meter and it's around the 155mH mark, as is another one in another unit which works fine.

From my A-level physics (a long time ago!) this give a reactance of 48 Ohms at 50Hz, so is about right for dropping 146v at 3A. If anything the lamp voltage is likely to be slightly lower than the spec'd 95v.

I haven't measured the PF correction capacitor. Would a failure affect lamp operation? I'm happy to stand corrected but I'm not sure it would - surely it would just give a nasty power factor?

As suggested, I'll leave it for a while and see if it settles.

Will post updates, especially if I'm clearing up after an explosion in a few weeks!

[greenalien]

I haven't measured the PF correction capacitor. Would a failure affect lamp operation? I'm happy to stand corrected but I'm not sure it would - surely it would just give a nasty power factor?

 

Depends on the failure mode - open-circuit, not likely to be a problem, as you suggest. However, failure in short-circuit, or intermittent short-circuit mode could be bad news. Intermittent failure, where the dielectric insulator is just starting to break down, could cause the lamp flicker you describe, so something else to try swapping.

 

Another thing to check - some metal halide arc lamp circuits also have a 'mains power conditioner' - these can also fail. Brown fluid leaking from the can is a subtle hint that all is not well with them...

[Brian]

MH640 Lamp Schematic...

 

[greenalien]

That's useful - the power conditioner is part 'FS981' shown at top left.

[gafferted]

 

From my A-level physics (a long time ago!) this give a reactance of 48 Ohms at 50Hz, so is about right for dropping 146v at 3A. If anything the lamp voltage is likely to be slightly lower than the spec'd 95v.

 

 

 

Just in case anyone believes this I should point out that it's nonsense since I've completely neglected to take phase into account!

 

Since the ballast and lamp are in series the current through them will be the same, but the voltages will be roughly 90 degrees out of phase.

 

The measured DC resistance of the choke is 1.6 Ohms so about 5V at 3 Amps.

 

I've just sparked one up and taken a few readings:

237v mains supply

218v across the ballast

76v across the lamp

 

Using vector addition this would indeed suggest a phase difference of around 90 degrees between the lamp and ballast voltages.

 

Also - the current draw for the whole unit was about .33A with the lamp off and 2.5 on initially sparking it up, falling to around 2A after 5 mins or so.

 

I didn't measure the actual lamp current - I guess it will be more than the 1.7A suggested by simply taking .3 of the total, as the PF correction capacitor will reduce it.

 

If the ballast really does have a reactance of around 50 Ohms, then 218V would suggest a lamp current in the region of 4A if I've got it right, which is obviously too much.

 

I'm just confused now!

 

Off the ABTT for some light relief!

[adam2]

Voltage and current readings in discharge lamp circuits can be perplexing and non intuitive.

Especialy, the lamp current multiplied by the lamp voltage will not normally equal the lamp wattage.

And this applies to both the nominal and the actual measured values.

The voltage and current will have odd waveforms, that may or may not be in phase.

Different types of instrument will react differently to these odd waveforms.

 

Measurements are therefore of relatively little use, unless they can be compared with similar measurements taken with the same type of instrument, on a good unit.

[KevinE]

I wouldn't think too hard about it.

If the PFC cap is bad it will either go bang and blow the fuse (s/c or leaky) or simply not correct for power factor (o/c) leaving the lamp operation untouched.

[bigclive]

Can I just mention that there's a bit of rocket science going on here. In various parts of the entertainment industry the term "rocket scientist" is used to describe a technician who tries to fix things with lots of maths, scientific theory and ill conceived logic. They're notorious for turning a small fault into a big one by doing "special" repairs, or indeed just getting so complex about a fault that the bit of equipment ends up in bits in a corner.

 

The lamp circuit is not complicated. If the ballast is open circuit then the lamp will not light. If it's short circuit then the fuse may blow and the lamp may fail abruptly. If it's the igniter then the lamp will probably not strike reliably. If it's the lampholder or a wiring issue then the lamp may keep cutting out despite not having many hours on it. If it's the lamp (and they do make the odd off-tolerance lamp) then it may keep extinguishing or exhibit a rather jittery discharge, although some new lamps also jitter for a while until any residual impurities have been absorbed into the getter, the metal vapour components have stabilised and the coating on the electrodes has reached a balanced state.

 

You can occasionally get odd effects if a power factor capacitor fails, since they have an internal mechanism that will disconnect the capacitor internally if pressure builds up inside its case. In most instances the loss of the capacitor won't make any difference, but if the wiring is already under rated then the extra perceived load due to bad power factor may have odd results.

 

I'd say to let the lamp run for a while and see if it stabilises. Alternatively swap in a known good lamp from another fixture.