12V - 350V 200mA converter for motorcycle CDI

[abuhafss]

Based on the discussion in the thread

https://www.electronicspoint.com/th...-ma-load-using-555.269429/page-5#post-1621004

I would like to have tips to modify that configuration to get an output of 350V 200mA to power a CDI. Although it has been mentioned in one of the post that this configuration would not be suitable for high voltages, I just want to do some experiments for the sake of learning.

 

[KrisBlueNZ]

So you have an automotive input voltage (around 12V)?

Are you sure 200 mA is enough? What's the capacitance you're charging, and what's the maximum spark per second rate?

You would need a boost/flyback controller IC driving an external MOSFET. The converter might need to be optimised to handle the widely varying output voltage.

Have you searched the web for existing designs?

 

[abuhafss]

So you have an automotive input voltage (around 12V)?

Are you sure 200 mA is enough? What's the capacitance you're charging, and what's the maximum spark per second rate?

You would need a boost/flyback controller IC driving an external MOSFET. The converter might need to be optimised to handle the widely varying output voltage.

Have you searched the web for existing designs?

Yes, the input voltage would be 13-14V.
The capacitor would be 1μF, 500V.
The maximum RPM would be 12,000 i.e. 200 sparks per sec.

Here are few designs which I have found on the web:

Another circuit which I obtained after reverse engineering a Yamaha CDI. The HV converter section is quite similar to the first one. The value of Q1 could not be read after TIP.


Here another UC3845 based



Here is another with a choke instead of a transformer

 

[Harald Kapp]

I modified the title as per the recommendation of Kris in the linked thread (#1).

The circuits you show are 2 separate classes:
1 and 2 show a full CDI that does not need a separate 12V -> 350V booster, this is part of each circuit.
3 and 4 are 12V -> 300V booster circuits, but without the ignition circuit.

What is your question with regard to these circuits?

 

[abuhafss]

What is your question with regard to these circuits?

Perhaps you could not follow my request in the first post. I want to make DC to DC converter for CDI based on 555.

Kris asked about the circuits which I had searched on the web so, I have posted them above. Forget about the ignition circuits in the first 3 schematics and just consider the Converter portions.

My primary question
1. How to modify the circuit in the linked thread to have 300-350V for CDI?

Secondary questions
2. What is the difference between the DC-DC converters shown in first two schematics above?
3. Where the pin #3 of UC3845 should be connected in the 3rd schematic?
4. What changes are required in the last schematic to use it as CDI supply?

 

[KrisBlueNZ]

This will be the third time I tell you that the 555-based circuit is not a suitable starting point for a high-power boost converter. I have removed "555" from the thread title.

Thanks for posting those diagrams.

Forget your primary question. Forget the 555 circuit.

You're right, the boost converters in the first and second circuits are very similar. In the second circuit, D6 is a zener diode that sets the output voltage.

The big difference is in the driving of the SCR. In the first circuit, all of the circuitry in the top half does signal processing and drives the SCR. In the second circuit, the processing is simpler.

But you say you're only interested in the voltage converter section. In that case, either of those blocking oscillator circuits can be used, though the part numbers of some important components are missing - the main switching transistor, as you mentioned, and the core and wire details for the transformer.

In the third circuit, pin 3 of the UC3845 is supposed to be connected to the MOSFET's source, and there is supposed to be a low-value resistor between the source and the 0V rail. I don't know why the designer didn't do this, and what kind of "control" signal is supposed to be fed into pin 3. You need to find the text associated with that diagram and post the URL.

The fourth circuit may be a good topology (a standard current mode boost converter), but the MAX1771 data sheet suggests a maximum output power of 24W, not the 70W you say you need.

CDI is not a standard application for a boost converter, because the output capacitor is being repeatedly discharged. There is also a very harsh mechanical and electrical environment. In that kind of special case I think it's best to go with what has worked before, so your second design, reverse-engineered from an existing product, may be the best to work from.

 

[abuhafss]

Although I mentioned about my concern for the 555 circuit, anyway leave it.

First two circuits are from two different models of Yamaha bike.

D6 in the 2nd circuit is a 250V TVS diode.
The main switching transistor in the 2nd circuit is TIP***, could it be a darlington?
Can you please help me understand how the oscillation is made and what is the function of the feedback secondary in both these circuits. The feedback network in the first circuit has only one resistor+diode and the second has 2 resistors+cap+diode......how both these techniques performs same function?

Here is URL for the 3rd schematic http://www.sportdevices.com/ignition/inverter.htm
And here is URL for the ignition schematic http://www.sportdevices.com/ignition/ignition.htm

 

[KrisBlueNZ]

Yes, the main switching transistor could be a Darlington. Lift the base and measure the base-emitter voltage to find out.

I think that configuration is a blocking oscillator - see https://en.wikipedia.org/wiki/Blocking_oscillator but there may be some clever interaction going on between the output capacitor and the oscillator circuit. As I said in post #2:

... The converter might need to be optimised to handle the widely varying output voltage.

The voltage on the energy storage capacitor will clamp the voltage on the feedback winding, and this will affect the oscillator in some way. Presumably it will be designed so that the oscillator frequency and/or duty cycle changes during the charge-up period. The same way that a photo flash circuit changes its frequency as the capacitor charges up. If you've ever heard a photo flash circuit recharging, you'll know that the frequency changes during the charge-up period. Of course the charge-up period for CDI is a lot shorter!

In the third schematic, the "control signal" on pin 3 is explained in the text on the page you linked to:

Current sense input is normally used to control the maximum current in the mosfet. It depends on transformer inductance and oscillator frequency, but not on load (with this topology).
Here, this input is used to control the temporary switching-off the power supply while the PIC activates the SCR to discharge capacitor. When line is 0 volt, the oscillator is ON, when line is higher than 1 volt, oscillator stops. This is done for two reasons: 1. SCR needs the current to be zero to be able to self-disconnecting. If power supply stills active, risk of permanent SCR activation exists. 2. To increase power economy, all power supplied during discharging is wasted.

 

[abuhafss]

In the third schematic, the "control signal" on pin 3 is explained in the text on the page you linked to:

The 3rd schematic is to power a PIC based ignition (schematic of the same I already linked).
However, I want to power a simple version as shown in the 2nd schematic and even in the 3rd on the right after the gray line. In that case, there is no PIC to offer any control. Here is my idea:


The 2N3904 and 1.5k is used to pause the oscillation when the pulse is received from the pick-up coil. But where to connect pin#3?

 

[KrisBlueNZ]

If you have a look at the UC384x data sheet you will see that it can be disabled either by grounding pin 1 or by forcing a voltage onto pin 3. In schematic #3 it is done using pin 3; in your circuit it is done using pin 1. It's actually most reliable to use pin 3.

You have shown the pickup coil feeding the SCR and the 2N3904 transistor directly, but that's not workable in practice. You are going to need some signal conditioning circuitry. This circuitry needs to provide a good strong clean positive pulse to the SCR's gate to trigger it. That pulse can also be fed to pin 3 through a suitable voltage divider, so that during the pulse, pin 3 is forced to around +1V and the UC384x is disabled. If you drive pin 3 like this, you do not need the transistor connected to pin 1.

Normally a compensation network is connected around pins 1 and 2 of the UC384x but circuit #3 doesn't have this. I'm not sure whether this will be a problem or not.

 

[abuhafss]

The circuit between the trigger input and the SCR gate is the conditioning circuitry.

So, I should remove 2N3904+resistor and connect the pin#3 to the gate of SCR. Normally the pulse is only ±5V, so do we need voltage divider to connect to pin #3?

 

[KrisBlueNZ]

No, that's not the kind of conditioning circuitry you need. You can't connect a simple inductive pickup directly to that point in the circuit. Post a link to the pickup you plan to use.

The conditioning circuitry needs to make a clear decision on the correct time to fire the SCR, and it needs to generate a pulse of the appropriate duration to ensure that all or nearly all of the energy stored in the capacitor is dumped into the coil. The leading edge of that pulse will trigger the SCR, which will unlatch when the voltage across it is too low to sustain conduction, and the flat top of that pulse will disable the converter.

The amplitude of the pulse, after conditioning, will presumably be about +12V so you should use an 11:1 voltage divider to pin 3. Yes, remove the 2N3904 and resistor.

If you want to try driving the SCR from the pickup without any conditioning circuitry, which I don't recommend, you can try the following:
D1: 1N914
D2: 1N914
D3: 1N914
R1: 10k
R2: 10k
Pickup positive to D1 anode
D1 cathode to R1 end 1
R1 end 2 to UC384x pin 3
R2 end 1 to UC384x pin 3
R2 end 2 to 0V rail
D2 anode to UC384x pin 3
D2 cathode to D3 anode
D3 cathode to 0V rail.

This will produce a signal at UC384x pin 3 that doesn't go negative (because of D1), is clipped to about +1.4V maximum (because of D2 and D3), is pulled to 0V (because of R2) and is at +1.0V or higher when the voltage from the pickup is at least about +2.7V (because of voltage drop across D1 and voltage divider effect of R1,R2).

 

[abuhafss]

No, that's not the kind of conditioning circuitry you need. You can't connect a simple inductive pickup directly to that point in the circuit. Post a link to the pickup you plan to use.

The conditioning circuitry needs to make a clear decision on the correct time to fire the SCR, and it needs to generate a pulse of the appropriate duration to ensure that all or nearly all of the energy stored in the capacitor is dumped into the coil. The leading edge of that pulse will trigger the SCR, which will unlatch when the voltage across it is too low to sustain conduction, and the flat top of that pulse will disable the converter.

The amplitude of the pulse, after conditioning, will presumably be about +12V so you should use an 11:1 voltage divider to pin 3. Yes, remove the 2N3904 and resistor.

Kindly see this German site, http://www.motelek.net/zundanlagen/cdi/cdi_lektion2.html showing 5 schematics of AC-CDI (powered by HV from bike's generator) all of them has similar conditioning circuitry. Even the one which I reverse engineered (Yamaha OEM) has identical circuit.

 

[debe]

This is what I used to power a CDI ignition on an engine, just using a small transformer. Uses minimum components & works.

 

[KrisBlueNZ]

Kindly see this German site, http://www.motelek.net/zundanlagen/cdi/cdi_lektion2.html showing 5 schematics of AC-CDI (powered by HV from bike's generator) all of them has similar conditioning circuitry. Even the one which I reverse engineered (Yamaha OEM) has identical circuit.

OK, fine. Then do what I suggested in the bottom part of post #12. As I said in that post, I don't recommend it.

Or use debe's circuit. It's called a Royer oscillator - see https://en.wikipedia.org/wiki/Royer_oscillator for details of the specific core characteristics needed.

 

[abuhafss]

Have limited space so, a bigger transformer could not be accommodated. By the way, how to pause the oscillation in Royer Oscillator? Grounding the center tap using a transistor?

 

[debe]

The transformer in my post is only 50x45x25MM fairly small. Its an iron core 240v to 20v CT, I made it to run a CDI unit I built for a generator that had a faulty CDI. Why would you want to pause the oscillator?? you just disconnect the power to it.

 

[debe]

You keep it simple unless you want adjustable electronic timing.

 

[abuhafss]

The transformer in my post is only 50x45x25MM fairly small. Its an iron core 240v to 20v CT, I made it to run a CDI unit I built for a generator that had a faulty CDI. Why would you want to pause the oscillator?? you just disconnect the power to it.

That transformer is fairly big as compared to EE ferrite core transformer roughly about 20mmx15mm.

The oscillator must be paused to stop the HV supply in relation to the triggering of the SCR, to avoid short path when SCR dumps the charge of the capacitor.

 

[abuhafss]

If you want to try driving the SCR from the pickup without any conditioning circuitry, which I don't recommend, you can try the following:
D1: 1N914
D2: 1N914
D3: 1N914
R1: 10k
R2: 10k
Pickup positive to D1 anode
D1 cathode to R1 end 1
R1 end 2 to UC384x pin 3
R2 end 1 to UC384x pin 3
R2 end 2 to 0V rail
D2 anode to UC384x pin 3View attachment 15011
D2 cathode to D3 anode
D3 cathode to 0V rail.

This will produce a signal at UC384x pin 3 that doesn't go negative (because of D1), is clipped to about +1.4V maximum (because of D2 and D3), is pulled to 0V (because of R2) and is at +1.0V or higher when the voltage from the pickup is at least about +2.7V (because of voltage drop across D1 and voltage divider effect of R1,R2).

Is this what you are suggesting, or do I have to remove the old conditioning circuit?