asdalton
Flashlight Enthusiast
The discussion on this thread generated a lot of argument and not much evidence, so I decided to calculate the efficiency of different 2x123A flashlights by using the metric of their lumen-hours per cell in regulation. Although this number is normalized per cell, the graph below contains only 2-cell lights. One-cell 123A lights will perform more poorly in squeezing lumen-hours out of their single battery.
Note: non-labeled Fenix points are from a regular P3D in all four brightness modes. The P3D Q5 is running in turbo mode.
Sources:
chevrofreak (Fenix and SF P60L data)
chao (Surefire E2L runtime data)
Quickbeam (Wolf Eyes drop-in runtime data)
General trends:
For high outputs, the LEDs are being driven at higher currents, leading to decreasing efficiency from the emitter alone. The batteries are also being taxed harder, limiting the ability of the regulator to squeeze the last bit of charge out of them. The overall trend of decreasing efficiency at very high outputs is therefore expected.
There may be some diminishing returns at very low outputs--below 25 lumens or so. This phenomenon may be due to the power consumed by the regulator circuit.
Notice how low the efficiencies of the Luxeon and incandescent lights are.
Outlier lights:
The Fenix P3D Q5 has the best Cree emitter available now, so it's not surprising that its efficiency is better than the same non-premium Fenix light (both in turbo mode). It is not immune to the trend, however; its efficiency at maximum power is still lower than the non-premium P3D (and SF E2L) running well under 100 lumens.
Compared to the turbo mode of the non-premium P3D, the P3D Q5 has a large advantage in output but a smaller advantage in efficiency. The Q5 emitter likely has a higher Vf, which would consume more power at a given current and therefore act against the efficiency advantage that comes from the higher lumen output. Chevrofreak's data shows the premium version running in regulation for 0.4 hours less, which is consistent with a higher Vf.
:thumbsdow The Surefire P60L lights, which are the LED versions of the 6P and G2 incan models, perform more poorly than expected given the Fenix efficiency vs. output curve. Some of this shortfall may be due to the compromise design as a drop-in. But the Wolf Eyes drop-in performs with similar efficiency at a much higher output. So it appears that criticism of these SF-made retrofits on runtime & output grounds is justified.
:thumbsup: The Surefire E2L, which is the only dedicated 2-cell SF Cree light released so far, matches the Fenix trend.
Future lights:
The upcoming Surefire L5, which is rumored to run for 2 hours at 120 lumens, will be consistent with the Fenix/E2L trend if it follows the usual Surefire rule (well, aside from the P60L) of putting out more light than advertised.
Older data:
Incandescent & Luxeon
Note: non-labeled Fenix points are from a regular P3D in all four brightness modes. The P3D Q5 is running in turbo mode.
Sources:
chevrofreak (Fenix and SF P60L data)
chao (Surefire E2L runtime data)
Quickbeam (Wolf Eyes drop-in runtime data)
General trends:
For high outputs, the LEDs are being driven at higher currents, leading to decreasing efficiency from the emitter alone. The batteries are also being taxed harder, limiting the ability of the regulator to squeeze the last bit of charge out of them. The overall trend of decreasing efficiency at very high outputs is therefore expected.
There may be some diminishing returns at very low outputs--below 25 lumens or so. This phenomenon may be due to the power consumed by the regulator circuit.
Notice how low the efficiencies of the Luxeon and incandescent lights are.
Outlier lights:
The Fenix P3D Q5 has the best Cree emitter available now, so it's not surprising that its efficiency is better than the same non-premium Fenix light (both in turbo mode). It is not immune to the trend, however; its efficiency at maximum power is still lower than the non-premium P3D (and SF E2L) running well under 100 lumens.
Compared to the turbo mode of the non-premium P3D, the P3D Q5 has a large advantage in output but a smaller advantage in efficiency. The Q5 emitter likely has a higher Vf, which would consume more power at a given current and therefore act against the efficiency advantage that comes from the higher lumen output. Chevrofreak's data shows the premium version running in regulation for 0.4 hours less, which is consistent with a higher Vf.
:thumbsdow The Surefire P60L lights, which are the LED versions of the 6P and G2 incan models, perform more poorly than expected given the Fenix efficiency vs. output curve. Some of this shortfall may be due to the compromise design as a drop-in. But the Wolf Eyes drop-in performs with similar efficiency at a much higher output. So it appears that criticism of these SF-made retrofits on runtime & output grounds is justified.
:thumbsup: The Surefire E2L, which is the only dedicated 2-cell SF Cree light released so far, matches the Fenix trend.
Future lights:
The upcoming Surefire L5, which is rumored to run for 2 hours at 120 lumens, will be consistent with the Fenix/E2L trend if it follows the usual Surefire rule (well, aside from the P60L) of putting out more light than advertised.
Older data:
Incandescent & Luxeon
Last edited: