Lumins per watt - LED vs CFL for incandescent replacements

[FloridaGuy]

I was at Sam's Club today and was pleased to see that they now carry a number of LED lights that are intended as screw in replacements for several common-sized incandescents and even halogen track lights with the non-screw-in base. I was surprised and disappointed to see that the lumens per watt (in the 30's) was only about half that of the CFLs (in the 60's) that they had.

Is it fair to say that the CFLs still have that much advantage over the LEDs from a lumens per watt perspective or is there another way to look at their relative efficiencies?

 

[Ragiska]

the bulbs they use are LOUSY. modern led's can manage >150lm/W

 

[kramer5150]

I would question the accuracy of the published CFL Lumens / Watt efficiency, how it was measured and under what conditions... cold start, after warming, fresh out of the box... etc. Its possible that the two lights were rated under different conditions.

FWIW... Using BCs well loved 10.5" integrating sphere and my own tailcap measurements, (Watt is the product of VxA measured at the tailcap) here's what some of my lights do on a Lumens / Watt basis. Note that these numbers are based on OTF Lumens and tailcap wattage consumption, so they factor in optical losses as well as electrical loss through the host and DC-DC circuits.

Xtar TZ-20, 4.2Vin:
47.6 Lu/W

Malkoff M60 in a Stock SF-6P, 8.4Vin:
42.93 Lu/W

Zebralight H501, 1.2Vin:
HI = 50 Lu/W
MED = 80 Lu/W
LO = 30 Lu/W

Surefire A2 incan, 8.4Vin:
10.7 Lu/W :candle:

Surefire E2L-KX2, 8.4Vin:
HI = 79.7 Lu/W

Fenix MC-10, 1.2Vin:
LOW = 80 Lu/W
MED = 97 Lu /W
HI = 32 Lu/W ( I had to double check my measurements, but the MC10 on HI draws a WHOPPING 3A at the tailcap!)

Clearwater P7, 4.2Vin
HI = 44.6 Lu/W

If that particular LED bulb is truthfully doing 30Lu/W, then its on par with my least efficient LED torches. Based on my own measurements, LED efficiency drops significantly when driven at higher currents. Which makes sense, given how much heat they radiate.

 

[vvs]

Highest Lumens per Watt on emitter has Cree XP-G R5 bin (around 150) with 350mA - peak efficacy. More current, less Lu/W.

 

[LEDninja]

FloridaGuy
Please read this thread before handing over your hard earned cash.
Don't buy Lights of America 120V LED Bulbs
http://www.candlepowerforums.com/vb/showthread.php?t=220441

-

the bulbs they use are LOUSY. modern led's can manage >150lm/W

Highest Lumens per Watt on emitter has Cree XP-G R5 bin (around 150) with 350mA - peak efficacy. More current, less Lu/W.

You guys are talking raw LED. In a household light bulb a lot of inefficient things come into play.

115VAC gives you 171VDC peak. A 171V to 3.5V buck converter is less efficient than a 6V to 3.5V buck regulator.

Big C tests his flashlights for 3 minutes occasionally up to 5 minutes. Even then you can see output drops of 10-20% as the LEDs heat up. A household bulb runs for hours and the adherence to incandescent bulb sizes means insufficient heat sinking in the small bulbs. The LEDs run hot and is much less efficient.

Bulbs sold in the US has to meet UL standards. The plastic front cover has to be fire retardant. The designer do not have his choice of the best optical material.

Bulbs sold in the US has to meet FCC standards. Extra filtering is required to prevent radiation interference with radio and TV reception nor mess up readings of medical instruments. This increases the inefficiency of the driver.

All these things combined reduces the lm/w numbers for an LED light bulb by a significant amount.

SSC current rating for a C bin P7 is 700-800 lumens. 800+ lumens is available in limited quantities as the D bin. Cree rates the MCE- at 700+ lumens. Your P7/MCE flashlight has to be minimum 700 lumens right.
http://www.candlepowerforums.com/vb/showthread.php?t=260659
From Big-C's testing the TK40 is down to 553 lumens at the 3 minute mark AND STILL DROPPING. At the 1 hour mark it will be much lower. (The DOE expects a light bulb to be used 3 hours/day.) Most single P7/MCE flashlights test out between 400 and 600 lumens. This shows raw LED performance numbers should not be used in equipment whether it is a flashlight or household bulb.

 

[Stress_Test]

Not to drag this too much off topic, but here are output/runtime results for the TK40 by Light-Reviews

http://www.light-reviews.com/fenix_tk40/

This is apples and oranges comparison though for flashlights vs. fixed lighting.