1. 120V Incandescent lamps, according to a Raytek brand non-contact thermometer, aimed at top of bulb from 2 mm away, hottest reading:
25W GE green party bulb A19, base down in free air, 26 C ambient: 70 C.NOTE The clear one of these has the usual argon-nitrogen gas fill and the "party bulb" appears to have a vacuum.
25 watt GE clear A19, base down in free air, 26 C ambient: 144 C.
40 watt GE clear A19, base down in free air, 23 C ambient: 160 C. 40W Sylvania softwhite A19, base down in free air, 23 C ambient: 155 C. 60W Sylvania softwhite A19, base down in free air, 23 C ambient: 200 C. 100 watt Sylvania clear A19, base down in free air, 23 C ambient: 217 C. 100W GE Extra Soft White A19 base down in free air, 23 C ambient: 230 C. 100W GE Extra Soft White base down in 20 cm globe, 23 C ambient: 274 C.2. 120V incandescents, other readings with the Raytek non-contact thermometer:
100W Sylvania clear, base down free air 26 C ambient, side of bulb: 120 C.3. Compact fluorescents, readings with the Raytek non-contact thermometer:
(Raytek was 12-13 cm from bulb, "mental average" of several readings)
13W PL/twintube, base down, free air, 26 C ambient, mid-bulb: 70 C. 13W PL/twintube, base down, free air, 26 C ambient, filament region: 107 C. 19W spiral, base down free air 26 C ambient, "mid-bulb" tubing: 84 C. 19W spiral, base down free air 26 C ambient, filament region: 98 C. 19W spiral, base down free air 26 C ambient, ballast housing: 49 C.
Temperature of the top of an 8-inch glass globe in 21-23 C ambient, as measured with a Raytek non-contact thermometer after non-noted amounts of time presumably sufficiently long for temperatures to stabilize:It is to be noted that a 42 watt spiral compact fluorescent heated the globe slightly more than a 60 watt incandescent did. The explanation is that incandescents produce a significant amount of infrared in the 700-2000 nm range that passes through glass without heating it, while CFLs produce little radiation other than visible light. CFLs produce more non-radiant heat than incandescents of the same wattage do.
41 C with a 40 watt T10 (vacuum) 59 C with a 40 watt A19 (gas filled) 69 C with a 60 watt A19 (gas filled) 82 C with a 100 watt A19 (gas filled)
50 C with a 20 watt spiral 57 C with a 25 watt Philips SLS 70 C with a 23 watt Sylvania Dulux EL (probably from having its top closer to the top of the globe) 70 C with a 42 watt spiral, (presumably with the sides of the globe hotter than with the 23 watt Sylvania Dulux EL)
However, all radiation that does not escape the room/building that the lamp is operated in does become heat - although not necessarily at the lamp or the fixture.
Equal power per unit wavelength from 377.5-762.5 nm - 189.6
Equal power per unit wavelength from 380 to 760 nm - 192 lumens/watt
Equal power per unit wavelength from 400 to 700 nm - 243 lumens/watt
Equal power per unit wavelength from 422.5 to 692.5 nm,which has color of a 5270 K blackbody: 264.5 lumens/watt.
400-700 nm portions of blackbody radiation at a few color temperatures:
3000 K: 256 lumens/watt
3400 K: 262 lumens/watt
3600-4100 K: 263 lumens/watt, within 1 lumen/watt.
4500 K: 261 lumens/watt
5000 K: 258 lumens/watt
5500 K: 254 lumens/watt
5800 K: 251 lumens/watt, supposed as a maximum by the linked paper.
6500 K: 246 lumens/watt
Trichromatic white / "white" achieved with 611 nm, 544 nm, and 450 nm:
4100K approximation - 397.5 lumens/watt
3500K approximation - 401.4 lumens/watt
Dichromatic "white" (with extreme low color rendering index):
4100K approximation with 450 nm blue and 573 nm greenish yellow: 478 lumens/watt.
3500K approximation with 450 nm blue and 576 nm yellow: 495 lumens/watt.
For some reference and in some cases some fine tuning and small discrepancies between use of the 1924 and 1988 photopic functions, look in the relevant discussions (mostly around 1997) here. Some figures are by Vic Roberts, a frequent poster to the Usenet newsgroup sci.engr.lighting, and some are calculated by me.
UPDATE 3/11/2009: More figures for various forms / definitions of "white light" having color rendering index at least 77, ranging from 154 to 366 lumens for 1 watt of such "white light", can be found on page 105 and pages 112-116 of:
Even higher figures are stated for cases when CRI is low, especially on pages 114 and 116.
As for what kind of non-white / non-whitish light has maximum possible lumens per watt of radiated light, that would be monochromatic yellowish green light with a wavelength of 555-556 nm, and such light has 683 lumens per radiated watt.
UPDATE 12/21/2006: Lumens in a watt of "white LED light": The answer is 331, at least for a particular Nichia one of correlated color temp. of 5450 K. The LED here was a prototype noted in This LEDsMagazine news item mainly for being a high power one that achieved an overall luminous efficacy of 91.7 lm/W. I derived 331 luminous efficacy of lm/W efficacy of emitted light by dividing 91.7 lm/w by the 27.7% "wallplug efficiency" (electrical energy to light energy conversion efficiency) mentioned in that article.
White LEDs of color temperature around 3500-5000 K will often have higher luminous efficacy of emitted light, while white LEDs of color temperature outside this range will tend to have lower luminous efficacy of emitted light. Also, higher color rendering index generally makes the lumens per radiated watt lower, by requiring significant amounts of red wavelengths that the human eye is less sensitive to than to most other visible wavelengths.
Copyright (C) 2006, 2009, 2015 Donald L. Klipstein (firstname.lastname@example.org)
Please read my Copyright and authorship info.
Please read my Disclaimer.