On Sep 12, 8:47 pm, SMS <
[email protected]> wrote:
> Andy M-S wrote:
> > Generators, hub and/or bottle, are an excellent way to light the road;
> > but they've been handicapped by inefficient lights--incandescent
> > lights, including halogens, honestly work better with batteries. With
> > LEDs, generators come into their own, and no complex circuits are
> > needed to make things work.
>
> Yes, it seems that with the Solidlight, or equivalent homebrew LED you
> can finally get a decent amount of light with a dynamo system.
>
> What is the voltage out of the bridge rectifier? Do you use a capacitor
> to smooth the peaks or not bother? Are the LEDs regulated at all? There
> must be a series resistor either built in or external.
>
> The difficult issue with the high power LED lights is dissipating all
> the heat from the semiconductor junction. It's a common misconception
> that LED lights generate no heat, while in fact the high power LEDs
> generate significant amounts of heat.
The rectifier I'm using is rated at 400v, 1.5A--reasonably
conservative; it *might* eat itself if both LEDs shorted. But (A)
that's an unlikely failure mode, and (B) honestly, in that case, who
cares about the bridge? As long as the LEDs are working, all should
be well. I got the bridge here:
http://www.allelectronics.com/cgi-b..._AMP_400_PIV_FULL_WAVE_BRIDGE_RECTIFIER_.html
There are *no* other electronics. I do not use a capacitor; I played
with a 1F 5.5v goldcap across one of the LEDs, but the stand light
effect didn't last long enough, not was it bright enough, to seem to
me to be worth including the cap. Nor is any flicker apparent except
at what I consider unusually low speeds--and at that point, the
flicker is kind of nice to have as an attention-drawing device, and in
any event I'm nearly stationary and in be-seen rather than see-by mode
with regard to my lighting.
No resistor is used; the bridge provides a small amount of resistance,
since it's not a high-efficiency unit, but that's all. Remember that
these LEDs are rated to produce full output at around 700ma, which is
more than the (Shimano model '70) is able to produce. Fortunately,
LEDs are generally efficient devices and light up extremely well long
before you reach the rated current. My initial test runs with a
regulated bench supply knocked my socks off!
Heatsinking is simply not as big an issue as one might think. My
first K2 light was wired the same as the above, but with a single LED
housed in a section of steerer from a 1-1/8" threadless fork. From
someone else on the web, I got the idea of mounting the LED star to a
star-fangled nut (nice coincidence, eh?), via a small cylinder of
aluminum. This too was extremely bright, and flickered/died only at
about very low speeds indeed (probably because it was easier to
achieve a voltage at which the LED would light), However, I was very
concerned about heat, hence the nice heavy aluminum casing.
I do not know what the junction temperature was, but the casing only
ever grew very slightly warm--certainly nothing like the casing on a
5- or 10-watt halogen/reflector light. I suspect this was because (A)
being fed from a pulsing DC source (recitifier output) the duty cycle
of the LED was low, and (B) whenever the LED is energized, the bike
was moving and the case was exposed to moving, not static air, and so
cooling was not merely convection. Forced-air cooling, all the time.
For the second light, both LEDs are mounted via thermal epoxy to a
steel washer (the nice thing about the K2 star is that it is
electrically isolated, so I could do this and not worry about
insulation) which is mounted (again via thermal epoxy) to an old
aluminum bullet-light housing. Once the epoxy was set, it was
reinforced with silicon caulk. At the section of casing where the
washer meets it, the temperature immediately after stopping has never
been more than a fraction of a degree over that of, say, the fender
beneath. So, unscientifically, I'd say that the light stays
reasonably cool.
Both lights used lenses and lens holders like these:
http://www.luxeonstar.com/item.php?id=1777&link_str=327&partno=L2OP015
http://www.luxeonstar.com/item.php?id=1778&link_str=327&partno=L2OP025
http://www.luxeonstar.com/item.php?id=1782&link_str=327&partno=L2OHS35-WH
For the first light, I used a 15-degree lens, which was excellent (I
also played with a narrower one, which was OK, but I wasn't that
impressed). For the second light, I put on 15-degree and one wide-
angle lens to use; that allows me a beam as well as a widef spill of
light, both of which are useful for visiblity to traffic and for use
on unlighted trails (part of my daily commute takes me on a moderately
curvy 2 mile MUT across a well-treed and unlit marsh, far from any
artificial lighting, occasionally frequented by iPod-equipped joggers
who think solid black is a fashionable color and reflectors are for
wimps. I like being able to see them in advance.
Neither light was planned all that thoroughly, and I was quite willing
to accept burnt-out LEDs along the way. In fact, nothing like that
has happened, which leads me to suspect that. for bike-lighting
purposes, doing much more than treating one or two K2s and a bridge
just as you would a light bulb is a waste of time. As long as you
follow half-decent wiring practices and use a metal casing to provide
heat conductance to the moving air, you have a light that blows away
the competition.
