Kokusan Blues

[blitz11]

I had a Honda CX500 which used a CDI. After 25 years, the capacitors went bad, and the ignition failed. Honda wanted $400 for the CDI box, which was probably more than the bike was worth.

I found on the web someone who had taken a CX500 cdi apart, and figured out the layout. For a hobby for a while, i built tube guitar amplifiers, so i took out my circuit board equipment, laid out, and built a CDI. I used a potentiometer to bias the thyristor so i could set the advance break point.

I built 2 for $25; one to use, and one as a spare.

I sold the bike, but the first ignition is still working with the new owner. Granted, it is a street bike, but as a point-to-point circuit, it has yet to fail. Using a big thyristor was key; it can dissipate heat, which keeps everything happy.

My guess is that because one can choose one's parts here, you could build a reliable ignition for not too much $$$. You can buy an enclosure and potting material, and if you're worried about EMI, you can do two layers of potting with the layers separated by foil.

But, I never had RFI/EMI problems with the CDI i built, and it was just a single board in a project box with NO potting. I don't see that as much of an issue.

This is all great information. Excellent work on Sport devices. That's REALLY great.

Thanks.

blitz

[GMP]

That was almost certainly a basic analog unit, plus, the 12V source on a street bike is nice and clean (the battery is a great capacitor). Not much to screw up in there. Now with digital powered by a magneto coil, its another story. Power/signals must be clean within certain thresholds. Poke around on a bike like the GG with an oscilloscope and you will be horrified. I would bet that the OEM CDI has a well thought out supply circuit, and the inside of the enclosure has metalized coating for sheilding.

I'm not saying to anyone don't do it, I think its great to experiment with this stuff and learn something. I just want you to be aware that there are more things to be considered for a good result.

[blitz11]

Good points.

1. Actually, i like the guy's idea about running it "total loss" off of a battery. That in itself eliminates power supply problems. I've not put a scope on my GG.

2. The secret to good audio (guitar, audio amps) lies strictly with the power supply. if you have a junk power supply, you have a junk amp. probably the same thing here....my guess is running total loss is the optimal.

3. I don't think that the CDI as drawn would be susceptible to or generate that much EMI/RFI. The power supply might cause issues with the PIC, but the external battery would eliminate that.

4. If you don't want to lay out $350 for a CDI, and you're handy, it might be worth doing. I might build one this winter and see where it takes me. Or I'll just hope my ignition lasts. PICs are easy to program, so building your own advance curve is all up to you.

This might make for a good "tech article."

Thanks for the follow up.

blitz

[GMP]

There are ways around the supply issue without going to a battery, you just have to understand what you are dealing with as far as noise.

Back in the mid '80s I built my own enduro computer prototype. It was fully programable like an ICO and had an LCD dot matrix display. It was based on an Intel 87C49 micocontroller and Dallas smart clock, and coded in Intel Assembler. It basically worked on the bench with simulated inputs. The KDX200 bike supply was too dirty so I went with a battery but the consumption was quite high with the ICs back then. I got laid off from that company and lost after hours use of the development system/emulator so I was never able to finnish it. That stuff was big $$ back then.

[jiauka]

Originally Posted by GMP

Poke around on a bike like the GG with an oscilloscope and you will be horrified.

100% true, I was about to connect my Garmin GPS using a simple diode bridge an a large cap, and I checked the output before connecting my garmin, and I was really horrified and I ended up with a bridge+cap and a DC/DC converter to get a clean signal. A similar aproach can be done on a DIY CDI. I do have experience with EMI/EFI microcontroller problems, a good separate digital ground and a shielding solves 99% of the issues.

If I have a spare weekend without any broken leg or arm I will try it.

have fun,

j.

[GMP]

I ended up with a bridge+cap and a DC/DC converter to get a clean signal. A similar aproach can be done on a DIY CDI.

Thats exactly the way I would go, and I suspect whats in the Kokusan. I would also isolate all external signals.

A lot of my experience comes from designing medical lab instrumentation for BD, primarily blood cell counters. These instruments have very high gain analog systems as well as microcontrollers, plus pumps, valves, and other noisy components. Also spent a few years at a TUV office here in the USA. And you?

[iancp5]

This is quite an interesting thread.
I wonder if it accounts for the some GG's don't idle properly issue? Also the fact that whatever I did with my 300 it was too abrupt in power coming in from idle. All the time people telling me the 300's a pussycat! I gave up trying to get a reliable idle or a smooth enough power entry for our wet and slippery enduros and bought a 200.

The 300 always felt to me as soon as I came off idle, wallop a load of torque appeared and would spin up in wet conditions. Turning on wet grass was a nightmare.

[GMP]

The CDIs do have different curves. They vary by year and displacement, some years the 250s and 300s having different curves. The 300 can be more of a beast with a 250 CDI. My '03 250 was a bit of a dud off the bottom until I swapped in an '00 CDI. Maybe your bike started life as a 250 and was converted to a 300 at the factory or importer, its been done here.

[jiauka]

Originally Posted by GMP

A lot of my experience comes from designing medical lab instrumentation for BD, primarily blood cell counters. These instruments have very high gain analog systems as well as microcontrollers, plus pumps, valves, and other noisy components. Also spent a few years at a TUV office here in the USA. And you?

I use to design medical monitoring equipment, now I'm doing digital TV systems that need to be ON 24/7, I guess you already know the related software+hardware problems on a system that needs to be allways ON, in some way a CDI has the same issues + the radiation problem and the harsh environment, but with modern SMD technology this is not a big problem.

have fun,

j.

[GMP]

Cool. If you experiment with the ignition idea keep me posted. I don't have a lot of electronic fabrication resources where I work now. We do custom process control stuff for our own use so its primarily buying subsystems and doing the software.

Good Luck.