Metal Halide Lamps

Have you tried a web search? There seems to be quite a lot of information around on them. Starting with:

It's an arc-discharge type of lamp, so I would expect that they would transmit rather well. At EOL they blink, so add in some modulation and make a mess of the spectrum. ;-)

Overkill. A .22 should do it.

EOL blinking is a long cycle (minutes) so won't appreciably generate modulation. The failure mode is that the voltage for ionization rises (with temperature) above the available voltage from the transformer, and the arc extinguishes. The cycle begins anew after the lamp cools sufficiently to re-strike the arc, a process that can take as long as 3 to 5 minutes.

Easier to hit with a shotgun...

I use halide lamps in my shop, and have no appreciable noise problems from them. New bulbs which may well help. A first step to reducing noise would be a line noise filter.

Also it may be a good idea to look into the differences between probe start (old school) and pulse start (new designs) as far as noise goes. I suspect that pulse start may generate more noise when starting up, as they must generate an arc over a greater distance than probe start bulbs.

te:

ation

ransmit

mess of

Maybe I should restate the problem. I often work as a consultant for large wireless carriers to go identify (and squelch) external interference their cell sites can "hear". We occasionally run across metal halide parking lot lights.

This is a top-20 market, so there are tens of thousands of these lamp fixtures already in-service. Filtering is not a reasonable approach. At least, not as a prophylaxis. -- And there's a word that doesn't get nearly enough use. :)

I also want to clarify that I'm not talking about lamps that come on for a few minutes, then turn off (like a street light you might see with a busted photocell or something). No, these lights can burn steady and bright -- and all the while, spew out broadband noise, often with significant amplitude (say -60 dbM, as measured at street level with a "omni" handheld antenna).

If the bulb intensity were fluctuating (and perhaps it is and our eyes just can't detect it?), then I would think: A-ha! Bad bulb. But more often than not, replacing the bulb alone does nothing to fix the noise problem. Changing the other stuff does. (Though, I've never encountered a situation where the bulb was not also replaced simultaneously with the rest..)

Thanks KRW for the Wiki link. I will be sure to check that out.

There's a joke in here somewhere.... "How many engineers does it take to repair a metal halide...." nevermind..

- mpm

OK, I'm getting an idea, here. Probably the ballasts have an integral line filter to reduce conducted EMI. It may be that lightning or other electrical disturbances can blow out the filter without killing the whole fixture. These line filter caps are designed to fail open for obvious safety reasons.

Jon

Someone else told me that over time, the windings start to short and arc (bad or melted varhish, I guess?), and that it's the arcing the causes the noise. Seems reasonable. Of course, a lot of times, when you look at the replaced windings, they look fine. No worse for wear. But maybe a visual inspection isn't enough to see the root cause.

I used to do a bit of work with arc lamps. When the street lamps come on, warm up and go off (then repeat), it's usually end of life for the lamp. As the lamp ages, the electrodes burn back and increase the effective arc spacing. This in turn increases the required voltage to run the arc. Most older lamps ran on a fairly simple magnetic ballast. As the lamp aged, it required more voltage than the ballast could provide and lamp goes out. After a cold lamp initially strikes with a high voltage pulse, a low voltage arc discharge forms. As the lamp warms up and the additives vaporize (halides, Hg...) and pressure increases, the terminal voltage rises to steady state. Arc lamps are current run devices.

The Xenon lamps I used were a wonderful white light source. A nice even output across the entire visible spectrum. Unfortunately, as a 'continuum' source, the emissions went down to RF and an unshielded/unfiltered lamphouse and socket would wipe out reception on any nearby radio. RF spectrum shifts somewhat with running current.

The Mercury arc lamps were generally very quiet and required no RF measures. Haven't really worked in the nitty gritty of metal halides.

Ballasts are now increasingly electronic types that achieve better power factor.