Motorcycle Capacitive Discharge Ignition System

[Matttyga]

Hi Guys,

I'm brand new to this forum, so first of all hello to all. I'm also a novice electronic hobbyist, so please go gentle on me

Moving swiftly on, I'm very interesting in making a DC CDI for one of my motorcycles. Specifically a 125cc single cylinder two stroke.

I've been trawling through the site for a couple of days seeing what's been done and reading various threads, and although that's been very helpful I felt the need to make a post to see if I could get a few thoughts from people far more experienced than myself.

Now, I'm not asking for the simple answer as I won't be learning anything that way, so I'll try to explain what I've done so far and where I'm at, and the problems that I'm having so far.

I found a few schematics online for DC CDI's and the one I chose for a starter is from the Sportdevices website: https://www.sportdevices.com/ignition/inverter.htm

I've built the circuit as closely as I can using the parts that I have available to me and I did actually have the circuit hooked up to an ignition coil and sparking a couple of nights ago, so I'm guessing that I'm not too far from the truth.

In the write up, they are using pin 3 of the UC3843 chip to turn the oscillator on and off by providing a 0V and >1V signal. However, they don't actually show that in their schematic. Just to be different, I used an Arduino UNO hooked up to pin 1 through a 2N2222 transistor to pull the pin to ground to stop the oscillator. I then fire an ON-OFF pulse to the SCR, followed by a short delay to allow the capacitor to discharge fully and hence allow the SCR to switch off, and then start the oscillator again. FYI, I am inputting 12V from my power supply, which is possibly not the best option, and measuring 300V at the output of the secondary.

I did actually jump for joy when the spark plug started sparking!!

I played with the timing a little, but unfortunately after about 10 minutes of playing it stopped. I haven't gone as far as checking all the components yet, but the mosfet seems fine, as does the SCR.

Comparing the sportdevices circuit to mine, I have a couple of difference.

MR856 Diode - I'm using a FR207

I put a 10 ohm resistor in series with the mosfet gate, along with a 10K pull down resistor, as I figured that it should be pulled to ground to make sure that it turns fully off.

I also changed the 47pF cap that they show on pin 4 to a 470pF. If my calcs are correct, the 47pF along with the 47K resistor on pin 8 makes for a frequency above what the UC3843 is actually capable of.

I put a 100nF ceramic cap from pin 8 to ground, as that was recommended in the datasheet.

The transformer I wound my self (first time I've done it), so there may be an issue there. It's an EE type ferrite core, approx 6mm2 with an airgap on the middle leg of maybe 0.5mm (I didn't measure it). The outer two legs are touching. I wound 25 turns of 0.3mm wire (which I doubled up, so two strands), then 4 turns of 0.6mm wire (for the primary) followed with the remaining 25 turn of the secondary wire. I didn't "choose" the wire diameters as such, as this wire was what was used on the transformer that I stripped, so I just re-used it. Probably the wrong thing to do, but hey, I'm a novice and I don't have any enameled wire at my disposal.

Oh, and as a final thing, I've connected it all up on a breadboard, which as far as I know is exactly how NOT to do it.

So anyway, after a rather long first post, I'd certainly be interested in hearing any comments. I'm sure to hear some criticism but that's fine.

I've read the UC3843 datasheet several times, and while I grasp some of it, certainly not all of it.

Happy to take a picture of my circuit or whatever if anyone thinks that may help. Also just as happy to scrap the whole thing and try something else if you think that the sportdevices circuit is not the way to go. As I said, I'm certainly no expert.

Better leave it at that for now.

Cheers,

Matt

 

[Diver300]

As your transformer is home-made, it may have more leakage inductance than expected.

https://en.wikipedia.org/wiki/Leakage_inductance

Leakage inductance will result in voltage spikes on the transformer primary, and it might be a good idea to have some suppression, maybe a small capacitor and a series resistor, between the MOSFET drain and ground.

I can't actually work out why the transformer is made as an isolated transformer. The end of the secondary with the dot could be connected to the drain of the MOSFET (as well as the primary).

 

[Matttyga]

Thanks for your reply.
I'm not really that familiar with transformers, and this is my first project that is using one. And as I wound this one myself then I'm pretty sure that it's far from perfect.
As to why use an isolation transformer, I can't answer that as I was just following their schematic. I did actually contact them a few days ago but as this circuit was designed 14 years ago the current crew have little knowledge of it and understandably don't really want to get involved.
Not sure that I follow your closing statement about the wiring of it.
I'll probably have a little free time to play again tomorrow but typically, work keeps getting in the way!
A friend of mine suggested a higher turns ration (more secondary turn) but he didn't eleborate on why.
How about my wire choice for the transformer? Any come ts in that? As I said, I just used the old wire that was on the transformer I pulled apart.
Cheers
Matt.

 

[Diver300]

I can't really comment much on the construction details. I don't see why you doubled up the secondary. The current in the secondary will be much less than in the primary, so you don't need wire that is at thick.

With only 4 turns on the primary, leakage inductance will definitely have an effect.

I think that it would be very difficult to fault find a circuit like that without and oscilloscope.

On the circuit, the adjuster can vary the output voltage from about 120 V upwards, but there is no limit. If the potentiometer is turned up to maximum, you could get enough voltage to damage something. It would be better to put a 5 k resistor between the bottom of the potentiometer and ground, instead of directly connecting the potentiometer to ground. That would limit the voltage to around 250 V.

I would test a circuit like that by first finding the transformer part can produce the voltage you need. It would also be a good idea to find if it is producing the current that you need. It says that it charges a 1 μF capacitor to 250 V in 1 ms, so that is about 0.25 A. You should be able to run a 60 W lightbulb from the output, the top of the 470k resistor.

 

[Matttyga]

Thanks Diver300. That certainly gives me a couple of things to think about.

I doubled up on the wire as that's how it came off the transformer I took apart. No other reason. I'll need to get some wire and have a play I suppose. So are you suggesting more turns on the primary, and then also the secondary to keep (or perhaps increase) the turn ratio?

Agreed on what you say about the voltage divider for the output. As it is I've set it at 300V. Don't need more, so I will add the extra resistor so limit the voltage to no more than I need and safeguard the circuit.

I have an oscilloscope, so I'll do some probing and see what I can find. Thanks for your numbers. I'll see what I get and report back.

Off on a tangent, I have an Aliexpress DC-DC converter that takes 12V input and adjustable output up to 450V. I did have a play with this thinking that I might be able to use it. The chip on board is an 8 pin DIP and from looking at the circuit, it's definitely from the UC384x family, even though it just has Chinese writing printed on the top. Just for a laugh I removed the output capacitor, attached my multimeter (with the module set at 200V) and grounded pin 1 of the chip. Instantly the voltage dropped to zero (as I was hoping), but then the voltage rose again to input voltage. Wasn't expecting that, so without a bit of deeper investigation I don't know what's going on there, but it sort of put a damper on my plan. Oh well.

Anyway. More later.

Cheers.

Matt

 

[Diver300]

Many boost converters are not isolated. For this application there is no need to have isolation, which is why I suggested that you could change the transformer connection. Ordinary car ignition coils, as in the classic contact breaker ones run from a 12 V battery, are non-isolated boost converters.

Non-isolated boost converters will usually provide the supply voltage (12 V in this case) when they are not actually operating. The input voltage can flow through the transformer without the switching element, the MOSFET or the contact breaker, turning on or off. I think that is what is happening on your DC_DC converter.

It is possible that you could use the DC-DC converter in your CDI ignition. You would need to remove the output capacitor, but it would the replace everything to the left of the vertical line in the schematic diagram in the link that you posted.

 

[Matttyga]

Thanks. Your closing statement on the DC-DC converter is encouraging as that was certainly a plan. I already removed the cap and was measuring things when I saw the voltage drop to zero and then back to input voltage, so it's interesting what you say about the voltage leaking through.

Its my day off tomorrow, so I'll have to convince the wife that I have important work at the factory and come back for another bash.

Cheers

Matt