3 Watt LED Bike Light Experiments
June, 2008, Rev f
Michael Krabach


Synopsis of Experiments
Prototype 1 - Resistor Controlled Prototype
Prototype 2 - Triple Cree with large heat sink
Prototype 3 - Double Cree with Individual regulators
Prototype 4 - Single Cree and regulator
Prototype 5 - Single Cree and regulator
Prototype 6 - Triple Cree revised heat sink
Prototype 7 - Triple SSC in C Mag-lite Head
Prototype 8 - Triple Cree in D Mag-lite Head
Prototype 9 - Prefabricated 3 LED Mag-lite Head
Prototype 10 - Red flasher with Wide Optic Lens
Prototype 11 - Yellow flasher no Optics
Prototype 12 - Red flasher no Optics
Prototype 13 - Red flasher two LEDs
Prototype 14 - Triple Cree with 8 deg Optics
Prototype 15 - Red flasher two Cree LEDs
Prototype 16 - Red flasher two LEDs Clear Case
Prototype 17 - Auto 12 v clearance light
Light and Beam Measurements
Conclusions, Recommendations, and Further Speculations
Summary Table and Parts Sources

Light and Beam Measurements

It is always difficult to measure the total amount of light from a bike light and almost impossible without expensive lab equipment. Beam assessment is also very subjective depending on the type of riding. To compare the prototypes I have used two crude methods. The first is with a box lined with aluminum foil and with a translucent panel inside to get the light diffused a little bit more. A hole was cut in the front of the box to shine the light through, and another hole to hold the light meter. I used a DX SKU-5100 digital Lux meter. The second method was to use a photo voltaic solar panel measuring the light output with a voltmeter. I placed a shiny #10 tin can over the panel to deflect side light down to the panel, hoping that all the light would be integrated by the solar panel. The top of the tin can also gave me a reference height for all the lights. All values are only relative to each other and are not absolute.

The below data shows that the light box seem to give the most consistent values. That is, three Cree LEDs should give three times a single Cree LED. The solar panel data does not show that. (I will have to look into that phenomenon.) The Mag-lite prototypes 7, 8, with the tri-optics, do not utilize the light as good as the L2 optics for the rest of the lights. Prototype 9 suffers from not being as good LEDs on the prefabricated module. The Q5 vs P4 bin LEDs show an increase of 28% and 21%. Earlier readings showed only 18%. The 8 deg diffuser covers for the L2 optics reduce the light (for all the readings) 12%, 2.5%, 6.5%, and 8.7%. Prototype 4 showed an increase in brightness using the diffuser, obviously trying to confuse me. What ever the value, (I have seen published values of 15%), the beam is greatly improved with the diffusers. The data also shows that my original Cheap 20 Watt Bike Light only puts out about the light as the newer triple LED models.

I have added some flashlights that I have for a comparison with the bike lights. The dimmest is the Cateye bike light, and that is after I modified the original by replacing the original LEDs with 14,000 mcd LEDs. My impression of that modification was that it doubled the brightness.

The Task Force flashlights (Lowes) show that the new Cree version is about twice as bright as the old Luxeon version. But looking at the Ultrafire WF-606A which as a 3 watt Cree LED vs the Cree Task force, which is also 3 watt, I am inclined to think that my light box will bias a broad beam over a narrow one. (When you think how the light reflects and re-radiates inside the box with the translucent panel, that conclusion appears logical.)

Beam patterns are shown for a driveway with milk-white jars to indicate distance. The jars start about 12 feet down a driveway from the bike and spaced about 6 feet apart there after. The orange street glow is typical of full-cutoff high pressure sodium vapor residential streetlights. It gives a nice reference intensity for comparison with the bike lights. The lights were aimed at an optimal distance in front of the bike, which varied depending on the prototype. The camera exposures were all the same at f2.9 at 1/2 sec, ASA 400 for the front bike lights. Camera exposure for the rear flashers on the garage were all the same at f4.5 for 1 sec ASA 100.

The other beam photos on the ceiling show the raw beam pattern with the tape X spanning 20 degrees. The camera exposures were on auto so the brightness is not representative of the lights.

Prototype version

Beam Pattern

Ceiling Beam Pattern
(tape X = 20 deg)

Light Box Lux
(w/diffuser, if any)

Percent of brightest

Light Box Lux
(w/o diffuser)

Solar Panel Voltage

Percent of brightest

Prototype 1 - Resistor limited 10 watt module using Mag-lite reflector.





Prototype 2 - Triple Cree, Buckpuck regulator with variable brightness, surplus power supply heat sink
Photos with and w/o 8 deg diffusers.

(Click above image for various options)






Prototype 3 - Double Cree with individual current regulators and switches.

(Click above image for low/high comparison )

1900 each, 6v lantern
4200 both, 6v lantern
3600 both, 4 AA NiMH


4100, 6v lantern



Prototype 4 - Single Cree Q5 with heat sink, On/off clicky switch.

3120, 6v lantern
2900, 4 AA NiMH


3200, 6v lantern



Prototype 5 - Single Cree P4 with aluminum stock heat sink.

2430, 6v lantern
2380, 4 AA NiMH


2600, 6v lantern



Prototype 6 - Triple Cree, variable brightness, aluminum stock heat sink.
Photos with and w/o 8 deg diffusers.

Same beam pattern as Prototype 2.






20 watt Halogen





Prototype 7 - Triple SSC, Mag-lite C head, bFlex controller.
Photo with DX 4630 tri-optic lens.

(Click above image for comparison of controller levels)





Prototype 8 - Triple Cree, Mag-lite D head, Buckpuck regulator with variable brightness
Photos with DX 1911, 1912, and 1916 tri-optic lens, top to bottom.

(Click above image for comparison of optics)





Prototype 9 Module with three 3 watt LED emitters on plate, Mag-lite C. Buckpuck regulator with variable brightness.
Photo with DX 1912 tri-optic lens.





Cree Ultrafire WF-606a Flashlight 3v C2 battery



910 bare bulb

Mag-lite 2 D cell with 3 watt Luxeon upgrade bulb
Photo with focus tight and opened.




Task Force (Lowes) 2 C cell with 3 watt Luxeon LED Flashlight



Task Force (Lowes) 2 C cell with 3 watt Cree LED Flashlight



MadMax Plus WO Luxeon AA Flashlight conversion

800 NiMH
380 alkaline


Cateye Opticube 5 LED bike light, 4 alkaline batteries. With replacement LEDs.



Prototype 10 Red Flasher Cree red LED with L2 Optics and multi-mode controller.
View video.

(Click here to compare flashers)

Prototype 11 Yellow Flasher, Cree 3 watt white LED, multi-mode controller, no optics.
View video.

Prototype 12 Red Flasher Cree red LED, multi-mode controller, no optics.
View video.

VistaLite VL700 - 7 LED Red Flasher
View video.

VistaLite VL500 - 5 LED Red Flasher
View video.

Five Flashers - Side Visibility views of the five types. Flashers are in steady mode for photo only.

(Click above image for side angle view)

Flasher Size Comparison Prototype 10 at top,
Prototype 12 at middle,
VistaLite VL700 at bottom.

Prototype 13 Red Flasher with two 1 watt LEDs, multi-mode controller. No optics.

(Click here to compare with prototype 12)

Prototype 14 Similar to Prototype 6 but using 8 degree optics.

(Click here to compare with prototype 6)

Prototype 15 Red Flasher with 2 Cree LEDs, multi-mode controller. No optics.

(Click here to compare flashers)

(Click here to compare from rear.)

Prototype 16 Red Flasher with two 1 watt Leds, multi-mode controller. Using off-the-shelf flasher for conversion. Clear cover.

Prototype 17 An adaptation of automotive clearance light for a tail light. Steady light, no flashing.