Err.. it is a job in progress! Nothing is finished yet!

Yeah, but that will mean new mounting points inside the crankcase cover. Let us see. Tomo, I will get some time to work on this.

good luck eng saab,i know u will finish it,.u r a genious born with soldering iron in one hand and soldering met in other hand..goodluck.finish it soon.

I did not get the optimal components. For eg - the transformer. I had to make do with a generic line transformer. Also had to use an LED as a Timing Light! The timing jumps around a little - less than 1mm on the rotor through the timing window. I guess this is too much. Will need to tackle this.

Tried to do some tweaks today. It is basically working fine at idle. Now, I need to make a sample map and see what happens at higher rpms.

Also, trigger coil seems to be situated at around 28 degree from TDC. Can anyone confirm?

Will update later.

@drvmtm - Thanks! But you seem to be exaggerating quite a bit!

Ripped out some nice parts from a brand new computer PSU. Now, converted to DC ignition system from AC one. The AC ignition was taking too much effort to run properly - the engine used to stumble at 2k rpm and refused to go above due to weak sparks. Of course DC ignition is better than AC ignition from all sides, but I wanted to keep it simple. Anyways, DC it is. The primary side voltage is 400+ V. The secondary side would be somewhere around 40.000 V assuming a 100:1 turns ratio.

Now, the spark is so hot that RFI interference causes the micro controller to shut down and restart! Damn it! I solve one problem and stumble into another. Anyways, I feel I am finally in control of the situation. Of course I may stumble into yet another unseen roadblock. I will try out the DC ignition tomorrow.

This is a 40kV, 80 mJ spark from the DC-CDI. Now, need to fit in the bike and play around with the timing. Asking once again, any idea, at what degree BTDC is the trigger coil located on a Pulsar?

Shit looks like lightning to be.
Good going Abhi, all the very best for the completion of your project.
Sorry can't help you anyway as these things are way above my knowledge.
You sure had to be in some R&D of automobile giants.

^^ Thanks for the appreciation dude. This time it almost looked like I bit off more than I can chew. Still looks like that sometimes!

I have found out the trigger angle, a +pulse at 28deg BTDC and another -ve pulse at 10deg BTDC. The stock CDI uses this second signal to fire at startup. At other times, the first signal is used. This means maximum advance is limited to 28deg which the CDI uses at 3500 rpm and is constant from there on. Another roadblock! I will need to get the trigger signal at atleast 40deg BTDC to make any improvements over stock. i do have some ideas. Let's see what happens.

Second problem, the DC-DC converter part which generates 300V+ from 12V has some problem charging up quickly enough. I get some misfires at 5000 rpm plus. Probably, flyback transformer inductance kickback problem.

Other than these two, I have solved many other problems - too many to talk about here. Actually these days, I solve a prob a day, I kid you not! I will be unable to work on this for about 15 days due to some other works that needs to be done (not related to this proj). I would like to solve these problems before I upload the source code and schematics. So, its a wait till then.



The circuit as of now!



Testing it out.



Finding out the various degrees.



Bonus pic! This is the custom wounded stator coil which I had done earlier. 19AWG wires. Powers a 100/90 bulb effortlessly.

nice to see that progress,
ignition signal at 40 degree btdc(max advance) means design is like retard type,
misfire, flyback transformer inductance, try using a optocoupler between controller and drive circuit.
are you still using trigger coil for signal? try hall ic for pickup(much better to mount,easy to tune)...
ot. you said 100/90 bulb, any side effects,heat prob etc (was planning a similar mod)

Hmm.. will keep in mind the optocoupler idea. AFAIK, it is only needed when using different power sources, like 12V from battery as well as say, 220V line AC in the same microcontroller as i/p or o/p.

Yeah, I am still using the trigger coil. Since, I have no problems there, will keep it as it is. That will mean the system will still be cost effective as well. BTW, do you have any particular IC in mind? Would like to have a look in the datasheet.

No side effects after 3 month use till date. I used thicker wires from coil to battery, a custom made RR unit, relay for h/l and ambassador reflector. Everything works fine till date.

isolating the microcontroller output can save it if anything goes wrong at discharge circuit.
parts availability always remained a problem to me, still i was able to get some ssa443a to experiment with.

Thanks for the suggestion. And in any case, I have installed the microcontroller on sockets. And I do have some BJTs between the micro and the rest of the circuit.

And, you seem to have worked on something similar. Please do share your experience.

i was working on a cdi, it was not digital but totally analog. used multiple sensors(hall ic's) with selector circuit so only a single sensor operates at a time.
design was stone age, but it worked(main advantage was i could select the operating sensor, even when the bike was running).

Actually, the suggestion to use an optocoupler to separate the actually ignition circuit from the microcontroller does seem like a good idea. For one it will help keep the microcontroller alive when one of the transistor or other components on the HV side goes bust.

Are you still having trouble keeping the microcontroller running stable? Since, in that case I would suggest build two distinct power rails, one for the microcontroller with a heavy amount of filtering and one for the boost supply. Again, having the HV side optically isolated will complete the shielding of the microcontroller from weird interference.

The micro runs rock stable. The inductance kick from the flyback transformer takes up some time when charging up the CDI capacitor. So, it does not charge fast enough at high rpm. Hence the misses. So, the inductance kick from the transformer primary needs to be controlled. I am thinking of using a simple diode in reverse. Anyways, I would work on this only after about 10 days. And I use simple 10uF, 0.1uF and 1000uF filters in parallel for clean power supply. I guess its good enough!

Ok, how did you mount the 2 sensors? The SS443a right? Mounting tips, in case I may want to do the same too!

any progress ? .