Mini Dynamo Light

[syc]

I uploaded some photos of a tiny dynamo light that I built up as a secondary for my main commuter light. Its your minimal light with only a rectifier, LED (XPG R4) and optic (Ledil Lisa XP). I had some surface mount schottky bridge rectifier IC's from Mouser and used them to make the build as small as possible.

The primary light is a B+M D'Lumotec - its got a nice beam pattern that put a horizontal strip of light on the road in front of you, as well as a standlight. But the emitter in the unit is fairly old and inefficient, and there isn't much spill if you keep the light low and on the road it should be. So I built up a secondary to provide additional light and wired it in series to the primary.

Size comparisons against an incan bulb from a dynamo light:

The guts of the light:

The parts in unassembled form (wired for testing)

Mounted on the bike's CETMA rack:

The wiring into the dynamo is only temporary, until I pick up some more heat shrink and maybe some connectors. But it works very nicely, and with only 2 LED, there's adequate light at walking speed and it should be more than adequate even on slow uphill grinds.

 

[shy_an]

Any chance of a circuit diagram for newbies? :twothumbs

It looks clean with minimal circuitry. Some of the ones in this forum is too much for my head to wrap around.

 

[BrianMc]

Nice. Theft resistant. (Hard to see it, let alone recognize it is a light.)
:thumbsup:

 

[syc]

Any chance of a circuit diagram for newbies? :twothumbs

It looks clean with minimal circuitry. Some of the ones in this forum is too much for my head to wrap around.

The circuit is basically identical to circuit #1 on Martin's oft referenced site:



The main difference is that D1-D4 are all inside of the one little IC. I don't have the original receipt/order anymore, but it is basically identical to this part in the DigiKey catalog.

In this diagram from the spec sheet, you see how the 4 diodes are wired up just like D1-D4 in Martin's schematic:

So you basically connect each of the ~ pins to the dynamo inputs, connect the + and - pins on the rectifier to the respective pads on your star.

The only hard part is how small the rectifier chip is - soldering wires to the pins takes some concentration and planning.

 

[shy_an]

Thanks. Don't you worry about burning the led on high speed runs?

I did a trial with bicycle on a stand and turning the cranks by hand I already burnt a led, but that was a simple red 5mm led. :p Did not dare try with a expensive power led

Any spec of the led used?

 

[znomit]

Thanks. Don't you worry about burning the led on high speed runs?

I did a trial with bicycle on a stand and turning the cranks by hand I already burnt a led, but that was a simple red 5mm led. :p Did not dare try with a expensive power led

Any spec of the led used?

Dynamo puts out 500mA.

5mm LED max current usually 20-50mA. :poof:

Power LED (Cree XRE XPE XPG, SSC P4) max current > 500mA :thumbsup:

 

[Steve K]

Thanks. Don't you worry about burning the led on high speed runs?

It does seem like heatsinking would be an issue. You might maintain the small size by mounting the light to an aluminum plate mounted to the underside of the rack. Or, if the rack was aluminum, use a heavy aluminum clip or bracket to transfer heat from the LED to the rack.

regards,
Steve K.

 

[kananga]

I think it should be okay for heat, it's only pulling 1.5 watts. The body and clamp alone should be plenty to take care of the less than one watt's worth of heat it'll generate.

I was thinking about something like this when I saw the 10mm Lisa optics. I need more throw, but the wow-factor alone of tucking a light this bright into a tiny scrap of 1/2" tube is nearly worth it!

 

[Christexan]

I see I was beat to the answer as I was replying!

But yep, 1.5w (max at speed) isn't much in moving air with an all-aluminum housing, and that's only at max output, when moving slower, the output will obviously be less, so really, as long as the LED base-to-outer-housing thermal path is good, shouldn't be any issue at all.

 

[Steve K]

I see I was beat to the answer as I was replying!

But yep, 1.5w (max at speed) isn't much in moving air with an all-aluminum housing, and that's only at max output, when moving slower, the output will obviously be less, so really, as long as the LED base-to-outer-housing thermal path is good, shouldn't be any issue at all.

For those of us who weren't involved with the design and build, it's hard to tell if aluminum was used in the construction, or how good the thermal path is. Steel is a lousy conductor of heat, and should be avoided for this sort of purpose.

The best approach is to use all aluminum, and use large surface areas at any interfaces. If in doubt, measure the temperature near the LED itself. Checking the temperature of the outside of the package won't tell you how warm the LED die is. The LED can be very hot and still work. It just won't live very long.

regards,
Steve K.

 

[syc]

Hi Steve,

For those of us who weren't involved with the design and build, it's hard to tell if aluminum was used in the construction, or how good the thermal path is. Steel is a lousy conductor of heat, and should be avoided for this sort of purpose.

The best approach is to use all aluminum, and use large surface areas at any interfaces. If in doubt, measure the temperature near the LED itself. Checking the temperature of the outside of the package won't tell you how warm the LED die is. The LED can be very hot and still work. It just won't live very long.

Thanks for the comments - I was lazy and didn't elaborate on the build in the posting, but it is mostly laid out in the flickr photo set, with an inventory of parts in this photo:



The emitter is thermally epoxied to an aluminum screw post that extends down the length of the housing (which is also aluminum - you might recognize that it is the head of a maglite solitaire). The rolled in bezel of the maglite is actually the rear of my light. This bezel is pinched firmly between the base of the screw post (you are looking at the base of the screw post in this photo) and another screw which is screwed in from the outside. There's also some thermal epoxy in this gap as well to maximize the thermal conductivity of the interface between the screw post and the housing and hold everything in contact.

So basically, you have an aluminum axle that runs down the length of the light (which isn't very long to begin with) and connects with the aluminum housing at the rear of the light, with thermal epoxy at all the interfaces. Its a short, all aluminum thermal path. There's not a lot of mass in the screw post, but its probably no worse than many 20mm stars.

The clamp that holds it to the bike is steel, but there's a lot of contact area, and lots of airflow.

Another option would have been to leave off the tape I used to wrap the rectifier/optic assembly and just pot the whole hollow area around the rectifier with thermal compound, so that there is a thermal path from the star/screw head directly to the housing. The thermal epoxy could have been injected in from the wire port on the side.

Though I think the path to the interface at the rear of the housing may be adequate. I'm counting on the short thermal path and airflow to keep it cool. If I ever make another one, I may go the route of potting the guts in thermal epoxy to get a more direct path - but I would wait for a tighter optic to come out before doing another one. This optic suffices because its a secondary to fill out the tight beam of the B+M.

Steve

 

[znomit]

... you might recognize that it is the head of a maglite solitaire

Kai buckle AAA lights might make good hosts for this.
http://www.kaidomain.com/ProductDetails.aspx?ProductId=5628