Why mercury lamps need to cool down before turning on again?

[Bendba]

Hi, this seems to be a simple question but I can't find the right key words to search for the answer.

What is the reason for mercury vapour lamps to have to cool down before you can turn them on again?

The only thing I can think of would be that when still hot, the residual vapours still conduct enough electricity so that the voltage between the electrodes can't get high enough to strike again. Is that why?

[T3sl4co1l]

The opposite, actually: the high pressure gas (not plasma, which would be conductive) requires too much voltage to break down, more than the starter can safely deliver.

See also: why an automotive ignition coil needs so much voltage (20kV+) for a tiny (<1mm) gap.

Tim

[mikeselectricstuff]

The opposite, actually: the high pressure gas (not plasma, which would be conductive) requires too much voltage to break down, more than the starter can safely deliver.

See also: why an automotive ignition coil needs so much voltage (20kV+) for a tiny (<1mm) gap.

Tim

For some lamp types there are special ballasts that allow hot restrike, using higher voltage
 

[Bendba]

Indeed, now that you say it, that makes sense.

Hot = high vapour pressure = high dielectric constant
Cold = liquid mercury = low vapour pressure = low dielectric constant

Then, next silly question, does the same thing happen with ignitrons and mercury arc rectifiers? Do they have to cool down before restriking if you turn them off?

[Zero999]

Indeed, now that you say it, that makes sense.

Hot = high vapour pressure = high dielectric constant
Cold = liquid mercury = low vapour pressure = low dielectric constant

Then, next silly question, does the same thing happen with ignitrons and mercury arc rectifiers? Do they have to cool down before restriking if you turn them off?

I'm not an expert but I don't think they do, because the pressure doesn't increase as much as it does in mercury lamps.

[Bendba]

It certainly would make sense for the ignitron as they are usually water cooled.

I guess then, if one wanted to spend the money on doing so, you could mix the two designs and build a mercury lamp with a third electrodes much closer to one end to reignite it when hot.

I've also been wondering what determines the power rating of a mercury lamp. As far as I can understand, the only limit is how much heat the electrodes and quartz envelope can dissipate. It's that right?

Well, I just found an answer there:
https://youtu.be/-MtdhSopACQ

[SeanB]

Your standard mercury vapour lamp has a starting electrode, which starts it when cold ( or when it has cooled down enough) by simple breakdown of the gap between the pointy end of this wire and the main electrode mass. The reason they do not restrike hot is the gas pressure inside is too high, after it has stopped running and keeping an ionisation path open, for this to arc over.

A metal halide lamp is just a mercury vapour lamp quartz glass body and tungsten electrodes with added metal salts and a low pressure Xenon neon argon gas fill, which needs around 2kV when cold to strike. When hot it might need 30kV to do so so will have to cool down enough. High pressure sodium lamps the same, just they use a small droplet of sodium metal ( often an amalgam of sodium, potassium and mercury) in the gas fill, though the glass used has to be a borosilicate glass so it does not get chemically attacked by the hot sodium vapour.

The size of the active cell is almost constant, it is almost the same in size from a 35W to a 250W lamp, just the electrodes and cooling for the glass and electrodes differ, and the hotter it runs the shorter the life, though it needs to run hot enough to turn the metal halides or metal into vapour with enough pressure to sustain the arc.

Xenon short arc lamps can be hot restruck almost immediately, because the construction of the lamp is such that you can apply up to 100kV across the electrodes to restrike the lamp without it arcing over externally to the lamp itself, simply because the electrodes are on either side of the body, and the lamp cooling is enough that the envelope will not melt in operation ( though it probably will be at red heat right before that), though there you also have a short life of the lamp before you replace it, as the failure mode is the internal pressure building up till the housing detonates. 200 to 1000 hours before you replace the $10k lamp is not unusual, and it will be run in a blast proofed housing, and supplied in one as well.

[Gyro]

I have a bunch of those MBF bulbs in the garage. I must work out what do do with them some day, they are rapidly being overtaken by the LED revolution. 

The Philips UHP short arc Mercury lamps used in projectors have a weird ignition sequence. There are actually two cavities in the quartz envelope - the main one with the discharge electrodes in and a secondary low pressure ignition cavity.

The main cavity is at around 10 bar when cold and something like 200 bar ( ) when hot. They need such a high pressure to get the Mercury to emit a continuous spectrum. This makes it very difficult to strike, even with a HV pulse. That's where the second cavity comes in, it is at low pressure with a sharp edged foil running through it. On the outside there is a coil of wire which acts as an external field electrode. During the starting HV pulse, the high field strength at the edges of the foil in the starter cavity causes UV emission. This UV passes through the quartz lightguide style to the main cavity where it causes sufficient ionization to trigger the main arc.

Clever stuff but they normally still need a short cooling off period before they will restrike. The normal failure mode in these lamps is buildup of tungsten film on the inside of the cavity - the electrodes run very close to their melting point! This causes localised overheating and a big 200 bar bang! Note that they're an explosion risk even at room temperature if mishandled.

Philips UHP lamp paper attached for interest...

[Kjelt]

A hot restrike such as in automotive bulbs are only for safety. If someone turns of their lights by mistake and continues to drive while the bulbs are cooling doen would be a disaster. Or the automatic lights when entering in tunnels.
Normal lampdrivers only goto 3-8kV ignition Voltage but the cardrivers go up to 25kV for safety which makes them more expensive but worse the bulblifetime decreases.
To compensate the carlamps have been extremely designed hence the high price in comparison to same wattage HID shoplight.

For projectors the same, advise is to let the lamp cool down for 15-20 minutes before reigniting, it helps the lifespan of the bulb and another thing the lifetime of a projectionbulb is for a single ignition and then the MTBF of a statistical groupsize of lamps. Using it as a normal user watching one or two movies and shutting down, so using hundred reignitions does severely limit that lifespan. So if you have a break of 15 minutes in between movies, just let the lamp on.

[T3sl4co1l]

On a related note (and possibly relevant here, too), halogen bulbs require a cooldown step: the tungsten vapor reacts with the halogen, preferentially depositing tungsten back onto the filament, but only while it's undergoing thermal cycles IIRC.

This may be relevant to the halide types, give or take the reactivity of the halide salts used.

Tim