I just took a scope picture of one full cycle of ignition timing, and the result showed that the engine, cranked by the starter, (sparkplug removed) rotates at 5.868Hz (cycles per second) corresponding to 352 RPM.

Looking closer at the previous closeup of the timing peaks reveals that there is more like 16 mS between the peaks.

Calculating like before. I get the first peak to arrive 34deg before the other. If that last one is positioned 3 deg. before TDC as Mike has shown in his ignition curve thread, then it means that a stock system can time from 37deg. before TDC.
But it would not be difficult to mount small neodynium magnets exactly where they are desired to be, to place the trigger pulses at appropriate timing points.

So far so good, I now have the circuit built on a makeshift breadboard, not pretty, but it works.
The test rig looks like this: It contains the level shifter, with the associated batteries, to the left. At the botton the +-31 volt supply is connected.
Top right is the terminal strip where the Ignitor module is connected, and the small square dot left of it is the switch which can supply 12V to the black wire.
Left of the rig is the 12V supply to the Module.
All in all 5 power supplies to perform this task.
(And at this point, I am still nervous that I shall fry the module when I turn the supplies on)
Top right in the picture is the high voltage probe, which will measure the voltage drop on a resistor that is acting as stand in for the ignition coil primary coil. At this point I am not interested in generating sparks with the associated high voltage ballihoo. I just need a voltage step indicating if the current through the ignitor is interrupted. This means that the voltage across th resistor will be 12V when the coil is charged, and it drops to 0V when the ignitor disconnects the coil.

The tests did not bring any usefull conclusions.
Even after reconfiguring the test bench to supply +-60V into the trigger input, did I manage to make the module respond to the trigger signal in an even remotely coherent manner.
The only exception being at 3000RPM, where it locked in and produced reliable trigger: and where there is enough time to establish a current through the coil for the successive firing.

It is true that my trigger signals are pure square waves, whereas the signal from the timing coil of the bike are more like sinusoidal. This may be the culprit, although it is beyond mee how a comparator can miss a fast pulse.

Then there were the most frequent response which would occour over the range of revs, like this totally erratic response.

When the switch is opened, and current stops flowing into it,  it is a property of a coil, that it will try to maintain this current flow. To this end, it induces a voltage large enough to do so, and with this voltage it drives a current into charging the capacitor, spending the magnetic field in it during the process. When the coil is demagnetized, it is the capacitors turn to act, and it discharges through the coil, depleting the energy it had stored, and building the magnetic field back in the coil. Only a bit of the energy is lost due to the resistance in the coil.
Like a seesaw, this goes on and on untill eventually all energy is lost as heat.
This is a well known phenomenon, it is called the LC ressonant ringdown, and it was waveforms like this I was hoping to see.
I will have to do some measurements with the module in the bike to try to find out what I am missing.

If you feel like diving a bit deeper into the function of the good old points breaker system, here is the absolute best article I have found that covers the subject:


https://ttypes.org/conventional-ignition-systems/

To be continued

Axman,

Picture IMG_20250616_183557_NR, shows the output of the pickup coil while cranking the engine without a spark plug installed. This image was made before I received a new used module.

The recent entries show how I tried to trigger this new module with waveforms generated by a signal generator, ampified to the same voltage level as that produced from the pickup coil.
As you will see, the module does not respond well to the signals produced, I don't quite understand why, but suspect that the fast risetime of this signal is the reason.
I am presently working on a way to faithfuly replicate the sinusoidal waveform from the pickup coil here in the electrinics lab, so that I can investigate the ignitors way of producing the spark.

Obviously, the ignitor is able to drive the bike with a sufficient spark, the bike starts and runs fine as it is, but the spark looks awfully weak to my eyes. Then again, the last time I fiddled with engines was in the 19 seventies, when point breaker systems were the rule of the day.

Cheers, Finn Hammer

Axman, all

When taking measurements on a running bike, with sensitive equipment, some signals show up which are not really part of the signal path, but find their way into it by induction and by other means.The generator signal being one such thing, an artifact, not intended to be there, but nontheless looking real on the scope trace.

For a while now, I will pause the trying to simulate the behaviour of the module on the bench, because, as we surely all can agree, there is nothing that beats the real thing.
I therefore moved the scope out to the bike to take some realtime shots from the trigger coil line and the ignition coil primary, to see how they line up relative to each other, over a range of revolutions.

The yellow trace is the pickup coil with 20V between divisions, the cyan trace is the voltage across the ignition coil primary, also with 20V between divisions.

I took readings at 1200RPM, 2000RPM, 3000RPM,4000RPM and finally 5000RPM. The latter being the fastest I feel comfortable revving the bike without the road as a load.

These readings revealed a couple of interesting things one of them I had previously overlooked: The voltage out of the pickup coil, when loaded by the internal resistance of the module, starts at 12.8V @ 1200RPM, then gradually rises to something like 31.2V "5000RPM. This makes a lot of sense and will greatly facilitate replicating the waveform for later bench tests.

The other thing that came out of the test today, is the clear correlation between the +- voltage peaks and the ignition timing.
I have copied closeups of the timing sequences related to the various engine speeds, and as you can see, the ignition advances from alignment with the negative pulse @1200RPM untill it aligns with the positive pulse @4000RPM. Revving the engine higher than that does not produce further advancing of the ignition.

How the module manages to produce the advancing timing is anyones guess at this time.


For obvious reasons, I was unable to verify the possible cut out @>10000RPM, and I did not feel like cutting the black wire to investigate what effect it has.

Axman88 wrote on Yesterday at 17:48:55:


You are more than welcome, I appreciate the conversation. I am not an expert myself so we can figure this out together, and your questions help me to focus on things I did not think about.
The RIGOL  DS1054Z of mine has many advanced features but it is really an entry level scope, and extremely good value for the money.
What Hantek scope do you have?

Axman88 wrote on Yesterday at 17:48:55:

I was thinking along the same line here...

Axman88 wrote on Yesterday at 17:48:55:

Exactly, I did not pay much attention to the polarity, since they are references towards each other, and not to ground.
Axman88 wrote on Yesterday at 17:48:55:

Yes, and that is a very low value indicating that there is perhaps not a lot of energy put into the sparkcoil. I am still trying to figure out what to make of what the cyan trace tells me, I will talk about it in a later post.

Axman88 wrote on Yesterday at 17:48:55:


Timescale for:
1200, 2000,3000 is 2mS (so you are close)
4000 is 1mS
5000 is 500µs

Axman88 wrote on Yesterday at 17:48:55:


Yes, same time base for all traces on the same screen.

Axman88 wrote on Yesterday at 17:48:55:


That is because the reference of the signals is the positive terminal, which is continously connected to positive. The other terminal is intermittendly connected to negative to initiate current flow through the coil (and that happens as you say ~2 divisions before the trigger) then abruptly disconnected to create the voltage spike which in turn induces the high voltage event in the secondary coil to make the spark. when the spark dies out, about one division later, both terminals are at 14V connected by the internal resistance of the primary coil.   And then the sequence starts over....


Axman88 wrote on Yesterday at 17:48:55:


Yes.
Neg? is when the coil begins to be charged
B is the trigger point, so there the spark starts
P is the initial amplitude of the secondary during the spark
S and E: I am not so sure. I don't think the spark lasts that long.  I will show you here in the end an expanded view of what is going on with the primary coil.


Axman88 wrote on Yesterday at 17:48:55:


I have not heard of a scope that has this function, but perhaps it exists on the really expensive high end types. Letme know what Hantek you have, it probably can store signals too.

I still find usefull features in my equipment. For example. I could export the trigger coil waveform from the scope and save it to a USB stick as a CSV file, then load it into my signal generator and play it from there. Pretty cool.

Anyway, here is the 3000closeup, which shows the going on's of the primary ignition coil. I think it shows that the spark event is over in around 200µS or so. It ressonates at 23kHz and since the inductance of the primary is around 11mH, then it has around 4nF to ressonate with. That is not a lot.
However, 4.3 ohms primary resistance allows around 3 amps to pass, resulting in an energy stored of 50mJ. Injecting this energy into a 4nF capacitor should result in a voltage of  5000V  across it's terminals, so something is not right. Lots to ponder here.....

(an online LC ressonance calculator is your friend: https://www.omnicalculator.com/physics/resonant-frequency-lc
Did you read and understand the fabulous article about classical ignition systems? https://ttypes.org/conventional-ignition-systems/

This is what I think:
Proper ignition systems like point breaker and the later ones where the breaker point is replaced with a thyristor: They are great because they establish a proper LC ringdown, with a good size capacitor, which transfers proper power to the spark in a well controlled manner.

Capacitor Discharge Ignition: Same, only the position of the switching element is different. Still a well defined capacitor is discharged into an inductance in series, resulting in a ringdown.

Then we have this Savage Ignitor system..... It just seems flaky. Perhaps it is due to the age of my specimen. I miss to see a beautifull sinusoidal ringdown.

-And yes, I know, thousands of Savages and S40's run for decades with this system without a problem.