DC DC inverter which would replace HV coil - Couple of questions

[Integer]

Hello!

I have couple of questions regarding DC-DC converter, but first of all I would like to present you with my problem.

So thing is that i have small 2 stroke motorcycle which has  CDI ignition. But problem is that it's High voltage coil which charges capacitor is failed and new part is very expensive, so I've come up with idea to build dc dc converter which would generate that voltage(about 250-300V).

I found this schematic on internet for the complete DC CDI, but thing is that I want to use only inverter from that circuit because I want to keep my original CDI(except spark, original CDI controls power valve in exhaust)
http://www.transmic.net/DC-CDI-v20/DC-CDI-Schematic.pdf

So this is part of the schematic that I would use:

My question are:
1. I was thinking of adding some voltage protection because maximum voltage taht my cdi can handle is 400V( i dismantled one old cdi and found that capactior is 1uF and rated on 400V)
I think that 300V would be enough so I added one varistor in schematic which should serve purpose, but I don't know is  that enough, and what else should i add for maximum reliability and protection?
2.On this shematic for inverter which is basically same: http://www.sportdevices.com/ignition/inverter.htm Is staded that "The last 1N4007 diode is necessary to load the capacitor with the current provided by the transformer" I was planing to replace 1N4007 with some shottky diode which are faster. On my shematic this diode is connected beetween CDI ignition output and ground.
3. Biggest problem is that HV coil generate voltage only when magneto passes near it and with this design voltage would be generated all the time and it would needed to be somehow wasted, otherwise there will be huge voltage build up. For that purpose i put varistor, but I don't know is that enough.
4.On original transmic shematics that problem is solved with pin 1 of the UC3845 which turns on generator when activating scr, so all voltage that is induced goes to the spark.

If you have any advice feel free to post because I'm not so competent in analog circuits and if I could make this, it would save me a lot's of money,and also because this is common problem on motorcycles it would help other people too.

Best regards and thanks in advance.

Josip

[Paul Price]

The original CDI circuit has a voltage sense circuit to regulate the high voltage that charges the 400V capacitor. You don't want to use a use a voltage-protection device, the original design already takes care of this. The 1N4007 diode does not need to be replaced, it is perfect for this design, besides that, a Schottky diode can't be found that handles this kind of high voltage anyway.

Also the pin 1 control of the IC is used to shut it down while the CDI capacitor is discharging itself into the ignition coil, and so the UC3845 is not trying to charge a virtual short-circuit while the SCR grounds the CDI capacitor and a spark is being made.

If the original power MOSFET that was connected to the burnt out part and otherwise the rest of the CDI circuit is ok, you would do well to repair the original circuit. This CDI is much more sophisticated than your replacement circuit since it has automatic spark advance with rpm, using the PIC MCU chip to do this, and this does a lot to make your bike run well at high rpm.

What you need to do is make or find some inductor that could replace the old burnt out coil. You can find some cores to make an inductor using raw materials you can get cheap, like the toroidial cores you might find  in a discarded desktop PC power supply. If you feel up to experimenting, maybe make a multi-tapped inductor, by winding and making  taps at each 25 turns of a total of 100 turns of AWG 20 wire upon  a 2-in diameter toroidial core. You should insulate the core body and each 25 turns of wire winding with transformer tape. You will have to see how it works on the bike with each tap, starting with the 100 turn end. Here's where cut and try might fix the problem.

The original circuit erroneously labels R2, a 100K ohm located on the drain of the MOSFET, shows this resistor as 100-ohms. Also it shows connecting the drain of the coil-driver MOSFET through mating connectors (P2-S2) to ground, this would short things out.

The original UC3845 circuit limits current into the coil, so there is little chance of damaging the CDI circuit with a DYI coil of your own.

Since the HV is regulated by the UC3845 driver IC, the power consumed by your CDI circuit is quite small when the rpms are very low or when you are stopped. When the voltage across the CDI capacitor reaches about 400V the controller shuts down and this makes the current used quit low.

I don't see anything in the CDI schematic that controls the power valve on the exhaust.

[Marco]

Why can't you get some generic CDI and just feed it the ignition signal? I assume your original CDI doesn't need the 300V to control the exhaust valve ... so it can just keep doing that any way.

[Integer]

Paul,

first of all thanks on comprehensive answer.

Also the pin 1 control of the IC is used to shut it down while the CDI capacitor is discharging itself into the ignition coil, and so the UC3845 is not trying to charge a virtual short-circuit while the SCR grounds the CDI capacitor and a spark is being made.

Can I make some voltage divider and opto isolator which would connect to the original CDI ignition coil output an when capacitor would discharge opto coupler would conduct and turn on T2 transistor?

If the original power MOSFET that was connected to the burnt out part and otherwise the rest of the CDI circuit is ok, you would do well to repair the original circuit. This CDI is much more sophisticated than your replacement circuit since it has automatic spark advance with rpm, using the PIC MCU chip to do this, and this does a lot to make your bike run well at high rpm.

My orginal CDI has also automatic spark advance. That, and power valve servo are the reasons why I would like to keep it.

What you need to do is make or find some inductor that could replace the old burnt out coil. You can find some cores to make an inductor using raw materials you can get cheap, like the toroidial cores you might find  in a discarded desktop PC power supply. If you feel up to experimenting, maybe make a multi-tapped inductor, by winding and making  taps at each 25 turns of a total of 100 turns of AWG 20 wire upon  a 2-in diameter toroidial core. You should insulate the core body and each 25 turns of wire winding with transformer tape. You will have to see how it works on the bike with each tap, starting with the 100 turn end. Here's where cut and try might fix the problem.

That is a lot of hassle and I even don't know would that solution work. Because that coil is in engine bay where are high temperatures and also it is bonded in resin along with the whole stator.

The original circuit erroneously labels R2, a 100K ohm located on the drain of the MOSFET, shows this resistor as 100-ohms. Also it shows connecting the drain of the coil-driver MOSFET through mating connectors (P2-S2) to ground, this would short things out.

The original UC3845 circuit limits current into the coil, so there is little chance of damaging the CDI circuit with a DYI coil of your own.

P1, P2, S1, S2 are the pads for transformer.

I don't see anything in the CDI schematic that controls the power valve on the exhaust.

Yes, because original CDI would control power valve, so only purpose of my schematic is to build inverter which would charge coil of capacitor and to replace original alternator coil.

Why can't you get some generic CDI and just feed it the ignition signal? I assume your original CDI doesn't need the 300V to control the exhaust valve ... so it can just keep doing that any way.

Marco,

that is good idea, but I would first like to try this inverter solution. Because it's hard to find CDI which has same ignition curve as my motorcycle and I don't know can I split pick up signal into 2 and feed one to original cdi to control valve.Original CDI needs pick up signal for RPM calculation and based on RPM it controls the valve.

Is pick up signal strong enough to be split?

Regards

Josip

[Paul Price]

If you try to trigger t2 with an optocoupler you would bypass the microcontroller chip and not have any spark advance.

I don\t understand what a powervalve for the engine looks like or what is it's function and I don't see the wiring from the connector connecting to the powervalve.

The same pickup signal wire from the crankcase  that computes rpm also is used by the PIC chip to determine when to deliver the spark

[Marco]

that is good idea, but I would first like to try this inverter solution. Because it's hard to find CDI which has same ignition curve as my motorcycle and I don't know can I split pick up signal into 2 and feed one to original cdi to control valve.Original CDI needs pick up signal for RPM calculation and based on RPM it controls the valve.

That's not entirely what I meant, the generic CDIs seem to me to just trigger on an ignition signal ... there is no curve. If you just tap the trigger signal for the SCR from your original CDI and use it to trigger the generic CDI it will still use the original curve.

If we were to take the PIC schematic as an example, the PIC would stay functional and take in the pickup signal ... but everything other than the PIC would be replaced by the generic CDI.

[Integer]

If you try to trigger t2 with an optocoupler you would bypass the microcontroller chip and not have any spark advance.

I don\t understand what a powervalve for the engine looks like or what is it's function and I don't see the wiring from the connector connecting to the powervalve.

The same pickup signal wire from the crankcase  that computes rpm also is used by the PIC chip to determine when to deliver the spark

I'm planing to trigger opto coupler with the ignition coil output from the original CDI. So when capacitor charges it will turn on opto coupler which would then connect UC3845 Pin 1 to GND.

I attached new eagle schematics, to make clear what I want and how I imagined this should work.

I hope that clarifies things a little, because I think that we have misunderstanding probably because of my bad English, and little bit complicated subject.

that is good idea, but I would first like to try this inverter solution. Because it's hard to find CDI which has same ignition curve as my motorcycle and I don't know can I split pick up signal into 2 and feed one to original cdi to control valve.Original CDI needs pick up signal for RPM calculation and based on RPM it controls the valve.

That's not entirely what I meant, the generic CDIs seem to me to just trigger on an ignition signal ... there is no curve. If you just tap the trigger signal for the SCR from your original CDI and use it to trigger the generic CDI it will still use the original curve.

If we were to take the PIC schematic as an example, the PIC would stay functional and take in the pickup signal ... but everything other than the PIC would be replaced by the generic CDI.

Aham, now I get your point. But how can I trigger scr of the inverter, because my original CDI is bonded in resin and I don't have access to its SCR. I only have access to ignition output which goes from CDI. Can I put there one opto coupler and with that signal trigger the scr of inverter? Something similar I draw in attached schematics for controlling the UC3845.

Thank both of you on help.

Regards!

Josip

[MarkClair]

Yes, how to trigger the scr is something which need to be discussed deeply as he has access only to the ignition output from the CDI. Sorry I don’t have a vast knowledge in this. I just bought an inverter yesterday from Staticon http://www.staticon.ca/product-detail/stativolt-inverters .They are SCR based, include transformer isolation, failsafe magnetic control and inherent surge rejection. I’m currently into setting it up.