After an agonizing three month wait (quality budget machinists are busy machinists and never send anything you greatly anticipate through regular mail from Australia to Canada), the parts for the first Nº58A prototype arrived. It was back in February that I first posted about
designing a new classically styled bicycle headlight and the prototype was ordered in March.
While I was waiting, I designed and ordered the PCBs for the LED board (with low loss FET rectifier) and stand light board. When they arrived I assembled and tested them. The circuit is very similar to the
dynamo standlight I designed for the
Sturmey Archer headlamp LED upgrade. It took me a while to find the time to solder the LED to the heat sink, connect the LED and stand light boards and assemble the whole thing into the housing. After fawning over its shininess for a few days, I took a few photos with my lousy point and shoot:
 |
| Nº58A |
 |
| Second hole is for a taillight connection |
Right off the bat, there were a couple of problems. The first was that the machinist's supplier substituted brass for the C101/C110 pure copper that was ordered for the heat sink This is a pretty sloppy and annoying transgression for which I was assured the supplier was sufficiently chastised (though it would've been nice if the operator had noticed that he was working with brass when pure copper was specified). Brass is fine for soldering LEDs to but its thermal properties make it unacceptable as an efficient heat sink. Consequently, it took a lot longer to solder the LED as brass isn't nearly as good a conductor of heat as pure copper. For the prototype it should be fine, but I'll need to make sure the heat sink is copper for the next revision.
The second problem is that a slight widening at the opening of the lamp body that was specified in the CAD files wasn't machined. Due to confusion about the female threading it was omitted on the final drawings by mistake and I didn't notice. This widening was to accomodate an o-ring for weather-proofing and without it the o-ring doesn't fit. Again, the next revision will need to have this step cut for the o-ring.
I've had several people suggest making a hanging version that can be mounted on a front rack. This, which will likely be christened the Nº58B, is in the works. In the meantime, the current version is designed to be mounted over the fork crown. Its weight dictates that a heavy duty bracket be used, something like the
4 mm wire bracket offered by Schmidt. Using this bracket, I mounted it on one of my bikes (enjoy the reflections of my living room):
 |
| LED needs centring in reflector |
Full set is available on
Flickr.
Overall, I'm quite pleased with the results. Everything fits together as it should, with the exception of the bezel's o-ring. I fretted about the location of the switch, but I don't think it's an unacceptable interruption to the lamp's shape. I might try a full switch boot instead of the half boot pictured or perhaps a switch with a shorter actuator.
With a few modifications, this could be ready for a production run if there's enough interest. Before that, I'd like to get the front rack hanging version made and log some serious hours to flush out any design flaws. An asymmetrical version would be nice as the round, flashlight-like beam of the current reflector certainly won't meet the
StVZO rules for bicycle lighting. However, for small volume production it seems unlikely that a custom engineered asymmetrical reflector with cutoff will be viable.
Beam shots? Not yet. I personally don't find them very informative, so they are a lower priority. Plus, the LED tested out OK after assembling the lamp, but a few days later it was dead, so I need to replace it which requires waiting for a revised heat sink to get machined. Hopefully this was just an unlucky failure due to the excessive heat required to solder it to the brass heat sink and not because of a systemic design problem. Only testing on the bike will tell...
