Sometimes you come across situation where you wanna convert either incandescent fixture to HID or vice versa. If I have understand right incandescent and discharge lamps do differ littlebit about how they produce the waste heat so I was just wondering how to rate this kind of conversion. If I f.e. have a fixture that is max 75w incandescent, is 80w merc too much? How about 70w hps?

Too hight wattage can ofcourse melt the fixture or catch fire if it's made plastic but too high temperature will probably also shorten lamp life of incandescent. How does HID lamps react if the ambient temperature is high around the lamp?

And to make it simplier, let's asume fixture is remote ballast so we don't have to think ballast adding more heatload to fixture.

Heat is emitted from the lamp by conduction, convection and radiation



Conduction

Affects what is in contact with the lamp, so the socket. This is worst with the GLS lamps, i think for 2 reasons :

 - Since the light from the filament freely reaches the insides of the lamp base, and there hits the Vitrite isolator from inside, we get tht some of the heat emitted initially as IR is converted into heat in the cap itself

 - The lamp surface area of GLS is smaller than of HID

Observation : E27 sockets made of PBT (the common European socket, made of off-white color Polymer called PBT, with mounting by 2 side clips) handle ~60W Max GLS lamp sideways/base down, but not 75W. The same socket handles well 70W SON E sideways, and if not for the eventual UV damage, 125W Merc base down

Ceramic sockets conduct heat better than Plastic sockets, so while they cool off better, they expose more to heat the wiring behind the socket

Copper in the wiring is a good heat conductor too. With 75W+ GLS or 125W+ MBF, sometimes even with smaller lamps, the conduction is sufficient to spread the heat along the wires connected to the socket, and slowly burn the isolation on the wires up to an inch or so back from the socket terminals, i.e. enough to reach the place where the wires come in contact and enter the cable (so make a short circuit) or enter some Metal tube of the luminaire (so make an Earth fault)



Convection

Affects what is above or near the lamp and made of Plastic. Over years affects soot buildup on areas in the luminaire that are above the lamp

 - Since HID lamps emit less IR, there is more heat emitted by convection

 - Since HID lamps are bigger than GLS lamps, for horizontal lamps the heat stream is spread over larger area, so not as concentrated. Allthough for something like MBF lamp standing vertically, the stream of hot air is quite concentrated off the top of the lamp due to its shape (very well visible by the soot pattern in old MBF lanterns)

Observation : HID luminaires made of Plastic tend to degrade in places where the concentrated hot air stream hits the Plastic. For example :

 - Plastic sphere (and cube etc) posttops using >=70W HID lamps normally have a heat shield above the lamp. In the better designs its an Aluminum reflector tower also acting as the heat shield. In the basic designs it is just a Steel disc held on a stick above the lamp. The heat shield breaks and diffuses the stream of hot air rising from the lamp so it won't melt a hole exactly there in the Plastic cover

 - Gaash Mars have a reflector assembly, which is made of folded sheet Steel and Aluminum, and is mounted to stand offs in the Plastic upper body of the lantern. There are knockoff punching lines around the mounting holes. The hot air trapped in the reflector assembly, which escapes in a concentrated stream through the knockout openings, melts the Plastic stand offs (which then causes the reflector assembly with the lamp to drop into the bowl and subsequent melting of a hole through the bowl)



IR Radiation

Affects any surfaces to which the light from the lamp reaches

 - Biggest problem with GLS lamps

Observation : Black Plastic parts of GLS lanterns tend to melt even when there is fair distance, and even when there is a stream of convection separating them from the lamp. The IR can melt Black Plastics even outside the luminaire, if they are too close (observe the "minumum distance from luminaire" signs on the luminaire !). White Plastics are affected too but less, as they dont absorb everything that hits them

Sometimes a Plastic bowl will transmit enough of the IR out to stay intact initially, but as it loses transparency (from marginal overheating, UV degradation, getting dirty) it will start to absorb more IR, which will make it start to melt, which makes it lose transparency, so its positive feedback cycle



i'd guess its roughly something like :

70W HID = below 60W GLS in conduction and IR

70W HID = above 75W GLS in convection



Metal/Glass luminaires would have little problem with either lamp. But beware of things like wiring exposed to the light from the lamp, gasket being near the lamp and so on

Plastic luminaires for GLS usually are already overrated, that is, they will eventually melt with the rated lamp. Use 1..2 sizes smaller GLS lamp than rated if you want it to last and not emit hot Phenolic polymer smell after long work in an enclosed small room

When converting, evaluate the effect of each mechanism of heat transfer separately

And one extra note: Due to the larger surface area of the hot components inside and due to their way lower temperature than incandescents, the HID tend to have higher convection and conduction components than incandescent of the same power input.
The cause is, the high temperature of the incandescent filament means way more is emited by the IR and of that, way more is concentrated around the shorter wavelengths (around 1um, vs about 3..5um in case of HID's), which is easier pass through the normal optic (designed to not attenuate the visible range), so leaves the glass bulb without actually heating it.
The longer wavelengths from the colder running larger HID burners are then quite effectively absorbed by the glass, so all that heat transfer is converted mainly to the convected type.

This is one of the reasons, why incandescent bulbs could be designed way smaller (they are less loaded than with HID's). The second is, the incandescents have limited life anyway, so the envelope does not have to last that long as with HID's (the higher temperature means faster fatigue crack spreading, so the envelope failure comes sooner; but with a 1khour rated incandescent is not a problem, if that happens after 5khours, but the same 5khours bulb envelope life would be quite a life limiting problem for an otherwise 25khour rated HID)

Thanks for both of you. That was quite well explained.

This is one of the reasons, why incandescent bulbs could be designed way smaller (they are less loaded than with HID's). The second is, the incandescents have limited life anyway, so the envelope does not have to last that long as with HID's (the higher temperature means faster fatigue crack spreading, so the envelope failure comes sooner; but with a 1khour rated incandescent is not a problem, if that happens after 5khours, but the same 5khours bulb envelope life would be quite a life limiting problem for an otherwise 25khour rated HID)

If fatigue cracks start to appear, the lamp is allready quite weakened. It may still take long time to reach breaking point if not disturbed. But i'd expect that if there are beginnings of cracks, every EOL GLS that is getting unscrewed from the socket would crack open if some force is applied... I have not noticed EOL lamps being any more fragile than new ones

If fatigue cracks start to appear, the lamp is allready quite weakened. It may still take long time to reach breaking point if not disturbed. But i'd expect that if there are beginnings of cracks, every EOL GLS that is getting unscrewed from the socket would crack open if some force is applied... I have not noticed EOL lamps being any more fragile than new ones

The cracks are not that far yet after the 1khour or so.
But there are always some cracks (in another words microscopic defects), it is impossible to do without. The thing is, how big thy are and how much of them are there.
And the bulb failure does not mean obly shattering, way more often it means mainly seal leak. And these are quite common (not every incandescent develops it, but occassionaly you have one,...)