This time for sure, bicycle light design
January 22, 2012
Not to be confused with this post.
Trying two new things, and one of them certainly worked in a prototype.
Thing #1, is to make the printed circuit board fit in a 1-inch tube. That means I can stuff it either in my steerer tube, or in my seat tube, out of sight and mostly protected from the weather.
Thing #2, which is a little bit trickier, but made a little bit easier by the amazing op-amps we can buy nowadays, is to build a current monitor that works from a few-millivolt signal above ground. This means that no matter what is attached in the way of lights, it will pull “the right amount” of power from a hub dynamo.
There’s a microprocessor in there, so that if you wanted to do something interesting, maybe you could. The microprocessor has the option to control a voltage doubler (PD2), to turn on a battery for a standlight (PD3), to count wheel spins (PD6 and PD7), to monitor voltage (PC5) and current (PC0), and to control both the steady light level (PWM on PB2), or to jam it hard on to prevent overvoltage (PD1).
I did a mess of simulation first, then built a simple test circuit, and it seemed to work; in particular, it pulled a steady current across a change in the load, and it did not put a big AC signal on the power control (i.e., it was not oscillating; that’s what the little feedback capacitor on the op-amp is for).
The constraints of stuffing it in a tube gave me some incentive to get all the wires at one end, but I couldn’t get them all there. Spare microprocessor pins couldn’t all fit at the end, and there’s a regulated pin that shares current with the LED supply. You could use this to run another supply, say a 5V supply to charge a USB device. The lights would dim as necessary to maintain a healthy current and voltage from the hub.