Compact fluorescents in enclosed fixtures

[dhinds]

My understanding is that use of compact fluorescent bulbs in enclosed and recessed fixtures is discouraged because the bulbs can overheat. I'm not entirely clear on whether the problem is primarily reduced bulb life, or if there is also a safety issue due to overheating the fixtures. I've heard the claim that CF bulbs heat up a fixture more than an incandescent bulb of higher wattage. How that is possible is a little mysterious to me since it seems to me that the CF bulb is producing a much smaller percentage of a much smaller wattage as waste heat. But maybe the CF heat is somehow better captured.

So... throwing caution to the wind, I accidentally left a CFL bulb (Feit 23w spiral, 100w equivalent, available all over in California for $0.25 due to PG&E credits) on overnight in an enclosed, recessed ceiling fixture. This morning I tentatively touched the glass cover plate... a little warm, not hot. I opened it up, felt around inside, the reflector and casing are only warm. I grab the bulb itself by the ballast and by the spiral itself... warmer, but not too hot to hold.

I know from experience that these fixtures (unvented metal boxes) get VERY hot with a standard 75w incandescent bulb. Can I conclude from these observations that the CF bulbs are fine? What sort of reduction in life should I expect from what seems like a very modest increase in operating temperature?

-- Dave

[Thomas Paterson]

Hi Dave,

It's going to depend from product to product, but it basically comes down to a simple question - what temperature was the lamp designed to operate at. Although there is no real way a 23W lamp is going to heat an enclosure to the same extent than a 75W lamp (especially when more energy is leaving as light), that doesn't mean that the lamp's operating temperature won't be exceded. Throw in a thermometer which records a maximum, and see what temperature the lamp reaches. Of course, it's difficult to get the thermometer bulb right in there, but pressing it against the control gear enclosure will come somewhat close. A thermocouple would be even better if you have one.

Then read the fixture's tech specs for its operating range. That's about the only way to answer the question.

Thomas.

[Don Klipstein]

In article , Thomas Paterson wrote in part:

I know this is true, but a 42 watt CFL can heat a fixture more than a 60 watt incandescent does. I experimented with an 8 inch glass globe, a 60 watt Sylvania "Soft White", a 42 watt Commercial Electric spiral, and a Raytek non-contact thermometer. The 42 watt spiral heated the globe very slightly more than the 60 watt A19, and not just at the top.

More light *and* greater fixture temperature rise with less input power

- how? The incandescent produces a lot of infrared, much of which escapes the fixture the same way that visible light does. With a CFL, nearly all energy that does not get converted to visible light becomes convected and conducted heat.

- Don Klipstein ( [e-mail])

[Victor Roberts]

There is no safety issue. The real issue is shorter life than you would have in free air.

The CFL ballast is composed of electronic components and all electronic systems have higher failure rates as the temperature increases. One rule of thumb is a 50% reduction in life for each 10C rise in temperature.

The part of an incandescent lamp that fails, the filament, is already operating at a very high temperature. Therefore, increasing the temperature by 10 or 20C does not significantly reduce its life.

As Don has said, most of the "waste heat" in an incandescent lamps is IR radiation at wavelengths just longer than visible light. It can escape the fixture just as the light does.

All of the waste heat from a CFL is very long wavelength IR that tends to get trapped in the fixture. But, as has been stated by others, the lower power of the CFL tends to offset this issue so any increased fixture temperature is small for equal amounts of light. If you used equal input power for the CFL and incandescent sources, then the fixture with the CFL would get hotter.

But you are not a semiconductor component.

Well, one lamp I know about was designed for 15,000 hours life in free air at 25C so it would reach 10,000 hours in a recessed can.

[Richard J Kinch]

I suspect the problem is the small electrolytic capacitors. They dry out and fail faster when run warm. Same thing that typically goes in a computer monitor or PC power supply.

[Don Klipstein]

I would like to add that for equal light output, a CFL typically has so much less power input that a CFL heats the fixture less than an incandescent does despite CFL being more efficient at fixture heating than incandescent.

I have posted one data point saying that CFLs are about 50% more efficient at heating a fixture than incandescents are. (This will vary, because different incandescents have different percentage loss by heat conduction/convection by the fill gas or lack thereof. The loss is highest in the lowest current models that use a fill gas.) CFLs are generally 3-4 times as efficacious at producing visible light as incandescents are. So a CFL should be able to produce somewhere around or over twice as much light as an incandescent that causes equal fixture heating.

On the other hand, half as much fixture heating may still be too much for CFLs since CFLs have a more restrictive upper limit on operating temperature than incandescents have. CFL life may be shortened. Also, excessively high temperature can cause the light output of a CFL to decrease, color to shift, and color rendering index to decrease due to above-optimum concentration of mercury vapor.

Based on how much of a heat hellhole recessed cans often are, I suspect the electronic ballast in that lamp would have to have a life expectancy of a goodly 50,000-100,000 hours, probably closer to or maybe even above

100,000 hours, in free air in order to have that low an impact on life expectancy of a CFL in a recessed can. In a recessed can, air warmed by the lamp rises and the ballast will get hotter than anything else. And hot air around the ballast may be at least somewhat trapped there, and the ballast often gets pretty toasty warm to put it mildly. That is true even if the tip end of the tubing is not all that hot.

- Don Klipstein ( [e-mail])

[Andrew Gabriel]

IME, the electronic failures in CFLs are due to the switching transistor burning out (often explosively). I have had the electrolytic storage capacitor fail when the control gear has been reused for much longer than its design life, but that doesn't stop the lamp working -- in the ones I've had, it just goes dimmer and takes longer starting.

[Victor Roberts]

Don, I agree. I was thinking last night that my post on this subject was incorrect on this point. In spite of the higher fraction of non-radiated power, a CFL will create lower in-fixture temperatures than an incandescent lamp of the same light output.

The issue that got those of us working on CFLs interested in this subject was not an equal light comparison, but an equal power comparison. At one point in time, we naively thought we could predict the in-fixture ambient temperature of a CFL by installing an incandescent lamp of the same POWER in the fixture and measuring the ambient temperature. We then realized the now obvious fact that due to the high fraction of "short wavelength" IR radiation from an incandescent lamp, a CFL would produce higher in-fixture ambient temperatures than an incandescent lamp of the SAME POWER.

I am guilty once again. I should have pointed out that screw-base CFL lifetime is function of both the life of the "wire lamp" and the ballast, with most failures in free air being caused by failure of the "wire lamp." The "wire lamp" failure rate is not a significant function of ambient temperature.

The term "wire lamp" is GE-jargon for the fully functional lamp portion of the CFL excluding the ballast, ballast housing and base. Perhaps others use the same jargon. I would love to have a better name for this part of a CFL, but the term "lamp" is already taken since it refers to the whole, fully-functional CFL and the term "bulb" is already used to refer to the formed glassware before processing into a functional lamp.

[Richard]

My long term experience is that open fixture compact Florescents tend to last longer. I typically use lower wattage compacts in small enclosed fixtures and higher wattage compacts in open fixtures. Never had a significant performance problem; and far superior to edison bulbs that run much hotter in either fixture.

Richard.

[dhinds]

It may be that the recessed fixtures we've got are not as bad as most. In these cans (cubes, more or less), the bulb is mounted horizontally roughly in the middle, and there's a curved metallic reflector at the top. So the ballast is definitely not at the highest, hottest point in the fixture. Again, after running all night, the ballast was warm, but I could hold it comfortably.

The box these bulbs came in says their "optimal" operating range goes up to +140F. I think the ballast was less than 140F, maybe closer to

120F, though I haven't checked with a thermometer. I think 140F would be too hot to touch, never mind hold comfortably?

I would be happy with a 50% reduction in operating life, given the usual life expectancy of a CFL. Given the bulbs are $0.25 each, cost isn't even an issue, I mainly want to avoid the environmental impact of going through lots of bulbs.

-- Dave

[Dave Martindale]

That's a very interesting result. However, it's worth pointing out that the 42 W CFL was probably putting out about 3 times as much light as the

60 W incandescent. Most people replace an incandescent with a CFL of similar light output, probably 13 or 15 W to replace a 60 W. In that case, the fixture would be much cooler with the 15 W CFL than the 60 W incandescent.

Dave

[Don Klipstein]

In , Victor Roberts wrote in part:

I was not aware of "wire lamp" being GE-specific jargon. I saw that term a bit in this newsgroup, and knew that it referred to the lamp minus its base and any integral ballast. As in being the bulb, everything within the bulb, and wires coming out of the bulb waiting for a base to be attached.

I just have in my mind non-regular people taking a look at this newsgroup, with not especially high chance of knowing what a "wire lamp" is, and should they be told that "the bulb is the glass part of the lamp" may still think of as a "bulb" the thing that most Americans that are engineers or other "technical types" outside the automotive industry call a lamp.

- Don Klipstein ( [e-mail])