Light output of dimmed lamps

[nicksanspam]

Somehow that seems unlikely, since incandescents are so non-linear.

That looks like the rms value of the rectified sine wave. If 120 = sqr(2V^2), V = 85.85 for half the waveform.

Nick

[nicksanspam]

I don't see that there. I can imagine it could go either way. A lower temp filament will evaporate less, so the bulb might last longer. How long would it last at 10% power? At 1%? At 0.0001%? :-)

Every black body above absolute zero emits some visible light.

m Ransley wrote:

Easy, with a grease spot photometer. Move a 3x5 card with a grease spot between dimmed and undimmed bulbs until the spot disappears, when it has equal illumination on both sides and outputs are inversely proportional to the square of the distance from each bulb.

Nick

[Joseph Meehan]

Yes it evaporates less, but far more of it ends up on the glass while when the cycle is working it just goes back to the filament. In real life it does work in such a way that the lamps do not last as long.

That said, my experience with projector bulbs is that used at half power they seem to last about the same time as full. While the older non-halogen lamps lasted a lot longer at half power. My 60W lamps by my sofa don't seem to be much effected, but the little peanut lamps under the counter last longer at full power maybe 50% longer.

Difficult to measure if true. I fear my college physics is about 40 years old now and while I am sure the answers have not changed on that one, my memory is a little weak.

That is what I would use, but there are other methods. I doubt if my spot meter could measure the light from a filament cooled to 3º K. :-)

[Calvin Henry-Cotnam]

Another thing to consider when dimming an incandescent bulb (and I believe halogens apply as well), is the spectral shift.

If you were to dim a 100 watt bulb so that it is only drawing 60 watts, not only will its light output not be the same as the 60 watt bulb burning fully, but the spectrum it puts out will be different.

What happens is that the lower temperature of the filament doesn't simply lower the light output, but shifts some of it into the non-visible part of the spectrum (the infra-red part: heat).

That 100 watt bulb that is dimmed so that it is only drawing 60 watts of power will appear less bright and more red/orange in color than a 60 watt bulb drawing its rated 60 watts.

[Jeff Wisnia]

Yep, and that's why I mounted a "hidden" dimmer with it's control shaft (sans knob)sticking down through the top of the medicine cabinet in our master bathroom. It's set a little bit down from full on and makes the light from the bunch of 25 watt frosted globe bulbs around the mirror redder than at full line voltage.

SWMBO likes it better for judging the appearance of the paint and stuff the puts on her face, and I find it makes my own mug somewhat easier to take on tough mornings.

As a plus, we've been here almost 20 years since I put that dimmer in and I swear I haven't had to replace even one of the bulbs it's controlling yet.

Happy Holidays!

Jeff

[Mark Lloyd]

It would emit EM radiation (because of this, an electric heater is NOT

100% efficient). Would it always be in the limited frequency range this is visible to humans?

[Mark Lloyd]

You might be able to see something at a lower voltage, with a small video camera that responds to infrared. Infrared is just below visible light in frequency, and will be emitted by a filament that's not quite hot enough to emit visible light.

[m Ransley]

Alot of the radiation from incandesants is invisable to the eye, it is in various areas of the spectrum and changes with voltage. Ever notice how incandesants photograph orangeish, cameras have different sensitivities to different wavelents of light. An incandesant outputs only apx 10% of its energy as usefull light, hence it is actualy an electric heater that emits light, where as a flourescent does better. Incandesant output 17-19 LPW, lumen per watt, T8 flourescent up to 110 LPW with 45-60 for CFLs being average.

[nicksanspam]

Every black body above absolute zero emits visible light.

Nick

[Edwin Pawlowski]

Yeah, but you just can't see it.

[Wayne Whitney]

More precisely, every black body above absolute zero emits light in the visible portion of the spectrum, often at an intensity too low for the human eye to see.

Cheers, Wayne

[m Ransley]

Well nick I find lights kind of usless unless they illuminate things I see.

[nicksanspam]

There would always be some visible light, but it's very faint at low temps :-)

20 PI=4*ATN(1) 30 LAMBDAU=7.8E-07'upper visible limit (meters) 40 LAMBDAL=3.8E-07'lower visible limit (meters) 50 FOR LT=2 TO 4'log F temp 60 TF=10^LT'temp (F) 70 TK=(TF+460)/1.8'temp (K) 80 GAM=.0143879/(LAMBDAU*TK)'Pivovonsky and Nagel (1961) polynomial term 90 CFRACU=0'initialize upper cumulative fraction 100 FOR M=1 TO 3 110 TERM=15/(PI^4)*(((M*GAM+3)*M*GAM+6)*M*GAM+6) 120 CFRACU=CFRACU+EXP(-M*GAM)/(M^4)*TERM 130 NEXT M 140 GAM=.0143879/(LAMBDAL*TK) 150 CFRACL=0 160 FOR M=1 TO 3 170 TERM=15/(PI^4)*(((M*GAM+3)*M*GAM+6)*M*GAM+6) 180 CFRACL=CFRACL+EXP(-M*GAM)/(M^4)*TERM 190 NEXT M 200 P=5.6697E-08*(TK^4)'total emissive power (W/m^2) 210 EFRAC=CFRACU-CFRACL'fraction of emissive power in visible spectrum 220 PRINT TF,EFRAC,P,EFRAC*P 230 NEXT

temp visible fraction total power visible power

100 F 6.010933E-22 531.1561 W/m^2 3.192744E-19 W/m^2 1000 2.753038E-07 24540.4 6.756065E-03 10000 .4668266 6.465423E+07 3.018231E+07

Nick

[CJT]

Wouldn't it be enough to periodically heat them up (i.e. isn't the damage reversible to some extent, as long as the bulb isn't too badly damaged)?

[CJT]

Ahh ... that's what I suspected.

[Don Klipstein]

If turned down enough that when turning back up, the light suddenly comes on, then no power was being sent to the light. If the light fades on in a manner like it faded off (except reverse), then the light probably was being sent power.

As for how much power a usual dimmer consumes when it is not sending power to anything: Not zero, but generally a small fraction of a watt.

- Don Klipstein ( [e-mail])

[nicksanspam]

So far, we seem to have no evidence of any damage.

Nick

[Don Klipstein]

The way I saw it on a rather technical book (25 years ago) on this, the halogen cycle slows down less than filament evaporation does when the bulb is underpowered. This is, ideally!

However, the halogen cycle does attack the cooler ends of the filament. You don't gain much when you dim below the point at which the ends of the filament get eaten away faster than the filament gets damaged by evaporation. Yes, the halogen cycle is only a little good at protecting a filament from evaporation - the evaporation is not perfectly even and the halogen cycle is not that good at returning tungsten to the parts of the filament that need it the most. Halogen lamps last longer despite higher filament temperatures not as much from the halogen cycle but because the fill gas is at a higher pressure and because the fill gas often has premium gases such as krypton - whose larger atoms "bounce" evaporated tungsten atoms back to the filament more than argon atoms do. The small size of halogen bulbs/"capsules" makes premium fill gases more economical. Still another advantage of halogen lamps is that the small capsule size slows convection movement of the fill gas, and with the fill gas more stagnant the filament evaporates more slowly. These reasons rather than the halogen cycle are mostly why a halogen lamp lasts longer than a non-halogen one of the same filament temperature. Of course, an advantage of the halogen cycle is keepin the inside surface of the bulb clean - which becomes more necessary when the bulb has a smaller inside surface where tungsten deposit would be more concentrated and more opaque.

Sometimes when a halogen bulb is dimmed severely, things go really awry

- depending on the quality of the bulb. Traces of water vapor or oxygen could be doing the opposite of what the halogen cycle is trying to do, and these bad effects do not slow down as much as the halogen cycle does when the bulb is dimmed, and the bulb needs a certain filament and bulb temperature in order for the halogen to outrun the negative effects of contaminants.

Well, a blackbody does emit some visible light at any temperature, but it appears to me that you need about 700 Kelvin to see it with a dark-adapted eye. In any case, I would think something would have to be far above body temperature to squeeze more light from a filament through your pupil onto any point of your retina than that point receives from the remainder of your retina.

- Don Klipstein ( [e-mail])

[Don Klipstein]

I have heard - sometimes even in literature from halogen lamp manufacturers - that that actually works.

If the bulb is badly damaged from the halogen cycle failing to outrun effects of water vapor or oxygen, then I would say only maybe. And the "repair" may not be good - the tungsten could have come mainly from one region of the filament and it could go mainly to a different region, and uneven filament thickness is bad since it tends to reinforce itself (thinner parts get hotter and evaporate faster). Also, the bulb material gets hotter than normal if the bulb is operated full-blast while blackened. Some halogen bulbs operate under enough stress as it is...

- Don Klipstein ( [e-mail])

[Don Klipstein]

Most 120V incandescent lamps glow dimly visible to a dark-adapted eye in a dark room at 6 volts. This varies from one lamp to another. For example, the lowest wattages that get a gas fill have the heat conductivity of the gas affecting things more, and need more voltage to glow, and brightness varies a little more with voltage than with high wattage lamps and vacuum ones.

With a tric type dimmer, you need a true RMS meter to get an accurate meter reading of the voltage when dimmed.

- Don Klipstein ( [e-mail])