Originally Posted by highseastravler
yes ive thought of that too one method was a controlled 360 to 355 degree delay line
thattook a sample of engine rpm and pre calculated the timing for the next revolution "a way of getting a pulse before the pulse came along " but programming a pic controller isnt in everybodies toolkit
PICs are the most important tool in my toolkit.
You read my mind, Keith, that's exactly the method I was considering. I have a range of 20MHz PICs here and all the junk to go with them, along with PCB design software etc to etch my own boards.
I'm getting more interested, at this point.
the other would be getting a 3 wire cdi that has electronic advance circutry built in and hoping that there was enough energy "voltage current and timing"
but i have been very dissappointed at the exciter coil output power "possibly due to that weak magnet they use
I thought about fitting a hall-effect sensor and using a real CDI, but as you point out, not enough energy without a battery and charge system.
I'm in the midst of putting together exactly that for my bike - a 100W DC motor (generator) into a regulator and 12V 5Ah battery. I'm powering the generator from my shift-kit jackshaft.
The drawback is that the battery carries all of the load when the bike is stationary.
It'll still give me enough for good
lighting and accessories, (like a real
so i was back at trying to trick the factory cdi into firing at a different time
and comming to the conclusion that it must fire at the zero crossing point to not be affected much by varying rpm magnet gap coil voltage etc"
the bias voltage on the ground return wire can be kept in check by back to back zeeners
and filtered with a non polar electrolytic with the applied bias voltage comming from a half wave rectifier from the lighting circut with a transistors bias controlled by a pulse with circut
but frist test would be to just put a low voltage battery in series on the cdi ground return line and see how much and if its consistant throught the rpm range the timing change
then a graph can bemade so a controller circut can be built
it should only have 3 connections 1 frame ground 2 lighting coil and 3 the ground return from the cdi
the exciter wire can be left alone anyway theres no need to have to midify that wavefourm
well off to testing hope we can beat the little black box lol
ps i think theres only a rectifier cap darlington pair and coil in that thing "it cant be that complicated"
theres defenantally no fancy timing circuts in there lol
The bias idea sounds like it could work, but a fair bit of current could be needed. It'll be interesting to hear how you go.