Does anybody have any theories to the attached picture.
This is a home made electronic ignition system using a Powerspark Halls sensor and a homemade rotor with embedded magnet.
The magnet position in the rotor is indicated by the black dot. The pick up in the red sensor is roughly where the black line is.
I have used a variable speed hand drill to spin the rotor.
At very low revs (hand twiddling speed) the spark seems to be in the right place.
As the speed increases the spark becomes more and more retarded. Approx. 160 degrees in the attached pic.
Slightly higher again and multiple sparks appear, in a regular pattern.
As the speed increases to what is probably a slow tick over speed it seems to settle down to a single spark in the correct position.
Any thoughts anyone. I'm worried my kick over speed could be where the spark is in the wrong place.
Thanks
Pete
Interference from the electric drill?
I'm guessing you're using a strobe light on the rotor to see at what point the spark is firing vs the rotor position.
Assuming also you are using a standard Lucas style 4ohm 12v coil?
The coil dwell time may be altering relative to RPM. Depending on how many degrees it's retarding with increase in speed, this would be a good ignition on a bantam! But for a 4 stroke, it needs to advance. Or better put retard less as revs increase.
Try opening your spark plug gap to a few mm, the spark should easily jump the big gap at atmospheric pressure.
You also want to see what revs your firing at on a 4 stroke your sparking every other revolution. So 4000rpm is sufficient, unless you're running a wasted spark system for a twin with 2 magnets and one sensor? (Or you can use 1 magnet and 2 sensors which can limit the amount of times you need each coil to fire.
I fitted a Thorspark ignition kit to my M33. The instructions advise that it isn't possible to time the ignition statically. Perhaps your setup is acting the same.
Thank you V500, Allan G and Rog1 for your replies.
V500, the motor of the power drill is roughly 12" away from the magneto. A simple matter to lengthen the twiddling stick, so I'll give that a go.
Allen G, I'm not using a strobe yet. The action has been filmed using an iphone in SLO-MO mode. That's 240 frames per second, played back at 30 fps (1/8th speed). This produces an amazing result but a huge file size. The bike is a B33, so single cylinder. The magneto runs at half engine speed, so one spark each time the magnet passes the sensor. The spark in the picture would be as the piston approaches the top of the exhaust stroke! and it delays further from there.
The coil is a standard Lucas type 6v 3ohms. The Halls sensor is supposed to work with 6V.
Rog1, I'm hoping that by spinning the rotor with a hand drill I'm simulating dynamic tuning, but without having the engine running. My initial aim was to get a firing position near enough to enable the engine run, then progress to proper dynamic timing with a strobe light.
Pete
Does the hall sensor "switch" the coil via a module ?
You're switching about 2 amps with a 6v system and 3 ohms primary.
Remember, the collapsing primary magnetic field will itself induce back emf that will go back into the circuit
Rhyatt, thank you for your response. There is no module. I shall forward the rest of your comment to my uncle, he's the clever one who made the system for me.
Pete
The early vw golf and Vauxhall cavaliers used a hall effect sensor which drove a small module , albeit a 12volt system.
Part of their circuitry was to deal with the collapsing back emf which - on later cars , sent spurious spikes back towards the ecus.
Of course , they all used neg ve earth , whereby the coil had a live and the earth path was switch by the triggering sensor and module, same as with points earth switching.
The diameter of the magnetic, also may have an effect , as there is no finite cutting point of its magnetic field and its relation to the sensor.
The vw had a cutting disc to create the cutting point.
Optical triggers has a light and a photo electric sensor and a solid disc between the 2 , with a tiny hole in the disc at the trigger point
Hope it helps
rhyatt, thank you for your further comment.
A bit of background info. We essentially copied a system that a popular bike restorer youtuber chap posted. His system works and seems to be identical to ours. I don't believe an extra bit of kit is required, but you may be on to something with the magnet. As before, I'll forward this on to my Uncle.
Pete
Does the Hall effect sensor need a magnet? Many are just a ferrous toothed wheel. I could be you are getting a trigger of the front of the ally rotor as well as the magnet. I appreciate yours is ally, but you may also be getting eddy currents.
limeyrob, thank you for your comment.
I didn't know about the ferrous toothed wheel. It seems possible that the sensor is picking up multiple signals from the spinning magnet. Maybe the magnet is too powerful or too close. Currently the system is off the bike and being bench tested.
Pete
I think there is a possibility that what you are seeing is to do with the frame rate of the camera (remember wagon wheels rotating the wrong way in old films?).
If so, you will need a strobe light to see what is actually happening.
Quote from: Aquarius on 22 Nov, 2025, 11:37 I think there is a possibility that what you are seeing is to do with the frame rate of the camera (remember wagon wheels rotating the wrong way in old films?).
If so, you will need a strobe light to see what is actually happening.
I worked with video systems of many years and I think that is probably what's happening.
Quote from: Rog1 on 21 Nov, 2025, 13:52 I fitted a Thorspark ignition kit to my M33. The instructions advise that it isn't possible to time the ignition statically. Perhaps your setup is acting the same.
It is possible to set the timing statically with a Thorspark, but not accurately.
I fitted a Lucas RITA to an A65, it had an inductive (rather than Hall effect) pick up and it had to be timed using a strobe as it continued to advance to 3000 rpm. But it came with a tool (a 3/16 strip of metal) that you could use to get a good enough static timing to get the engine running safely for the strobe timing. It worked pretty well and was only a few deg out. There clearly is a repeatable relationship between static set up and dynamic timing, the problem is knowing what that is on a bespoke set up.
https://www.ebay.co.uk/itm/191112866807?_ul=GB&mkevt=1&mkcid=1&mkrid=710-53481-19255-0&campid=5339019731&toolid=10001&customid=eb%3Ag%3Avms%3Aeb%3Ap%3A191112866807%3B{GCLID}&_trkparms=ispr%3D1&amdata=enc%3A1JI4DdDIRRFSN2QActaccnQ43&ad_type=pla&loc_interest_ms=&loc_physical_ms=1007061&gad_source=1&gad_campaignid=20747179918&gbraid=0AAAAADmMgiiDHp2qyA3znNTwdmQrS6X-3&gclid=CjwKCAiA24XJBhBXEiwAXElO3xfA4UOrOA5rJkYjpH3H-7qnffxLWc6uh4r8qdXPhVUe16AOXiTOlRoCZSsQAvD_BwE
these are popular hall conversions for Mg cars
whereby you have 4 magnets(I think) moulded into the black plastic rotor to give 4 triggers(same as a 4 cylinder ignition car that it replaces).
so the principle is sound but you just need for a 1 cylinder bike
Thank you for the recent responses.
I'm familiar with the wagon wheel (stagecoach not a round chocolate confection) or stroboscopic effect/aliasing. I may not have made it clear, but I was running these tests at very slow speed, under 3 rotor revolutions per second. I could see with the naked eye and hear that odd things were happening. The camera was used to record and retain a record. At 240 frames per second there were no stroboscopic illusions going on.
Rhyatt, we used a sensor very similar to the one you have linked. There's no shortage on Ebay. One magnet with the rotor we made, as it's a 4 stroke single and the magneto runs at half engine speed.
It seems I may have fitted the wrong coil. I'm told it should be a non ballast coil, not a ballast coil. And I thought this was going to be simple!
Pete
It seems I may have fitted the wrong coil. I'm told it should be a non ballast coil, not a ballast coil. And I thought this was going to be simple!
what voltage are you using on the bike ? 6 or 12 and pos or neg earth
you may find this useful - it says about coils with 1.4 ohms max(12 volt system)
https://www.mgocspares.co.uk/DocumentHandler/72022/g421-accuspark-ignition-fitti-pdf?srsltid=AfmBOormEQtjilVBUB0Y5mNWkh82S9cT0Hmd0Nda_VkQ773xleA2qD7o
rhyatt. 6v / positive earth. I will read through what you have sent. Thank you.
Currently looking at 6v non ballast coils. No shortage on the interweb, but they all seem to be the larger (55mm diameter) size. Going to make squeezing one in discretely a tricky job.
Pete
When you say the spark is occurring at the wrong place, I take it you mean it is jumping from the plug body and not from the spark gap..?
It's at the wrong place, not the wrong time..?
40mm diameter 1.5 ohm coils are readily available - I have a couple which have been used for only a few minutes (failed experiment!). You could have one for say £12 posted (UK).
Quote from: scifi on 22 Nov, 2025, 22:14 When you say the spark is occurring at the wrong place, I take it you mean it is jumping from the plug body and not from the spark gap..?
It's at the wrong place, not the wrong time..?
scifi. No, the spark is in the correct place, just not at the right time.
Quote from: iansoady on 23 Nov, 2025, 10:05 40mm diameter 1.5 ohm coils are readily available - I have a couple which have been used for only a few minutes (failed experiment!). You could have one for say £12 posted (UK).
iansoady. Thank you for the offer. Is it specifically a 6v non ballast coil?
Pete
No, the spark is in the correct place, just not at the right time.
Ah, the Andrew Preview problem.. ;D
So to clarify,
The HT spark is occurring across the spark plug electrodes , initially at the correct degrees btdc , but then wanders and changes it's point of timing all on its own
rhyatt.
That's about the gist of it.
To recap -
Firstly, the rotor with the magnet is fitted into the end of the magneto shaft. It is not screwed in, but is free to rotate. The rotor is being spun with the aid of a wooden rod. Initially by hand, then with a slow speed hand drill. At no point during the observed test was 180 rpm (rotor speed) exceeded.
At very slow speed the spark occurred as the magnet passed at its closest point to the middle of the sensor. As the rotation speed was increased the spark became further and further delayed. In the picture on the first message the spark occurred approx 160 degrees after the magnet passed the mid point of the sensor. A further small increase in speed and a second spark just after the first commenced. A little bit quicker again and a third spark happened. So now three sparks per revolution of the rotor?
Culprits.
Faulty sensor - Unlikely.
Faulty or incorrect coil - possibly.
Magnet in rotor too strong or wrong polarity - maybe, bench testing this now.
Pete
Quote from: PeteB31 on 23 Nov, 2025, 18:01 Quote from: iansoady on 23 Nov, 2025, 10:05 40mm diameter 1.5 ohm coils are readily available - I have a couple which have been used for only a few minutes (failed experiment!). You could have one for say £12 posted (UK).
iansoady. Thank you for the offer. Is it specifically a 6v non ballast coil?
Pete
Yes. It has a screw in HT lead connection. It's one of these: https://www.classicbikeshop.co.uk/6v-screw-top-ignition-coil.html
ok
most car systems use INDUCTIVE on their crank sensors and front wheel ABS sensors and these drive modules that are capable of dealing with the Amps that an HT coil needs.
a crank sensor has its own inbuilt magnet that just needs a metal trigger whizzing past it to make it work(it generates an AC voltage as it whizzes past the crank sensor and this triggers the module that triggers the HT coil)
if you find your magnet it causing the issue , maybe consider a small inductive sensor
your initial picture , looked like this untit ----
worth speaking to ACCUSPARK and all you need maybe is a single trigger built into the black circular slip over cam
I'm thinking two things:
Magnet too strong
Ally rotor - its not a full circle and its conducting. So it has edges passing the sensor.
If eddy currents are the problem then they would increase sharply with speed.
The retarding is a serious issue, the Lucas RITA I'm more familiar with automatically advanced with speed as the inductance increased and triggered the spark earlier so it was inherently reliable.
Aside: I must ask "why". I much prefer magnetos over coil ign, there's a reason most light aircraft flying today are still using magnetos and I notice that when they raced A65s in flat track many were swapped from coil to mag - a significant mod. Mags fell out of favour because they cost too much, not because they did not work.
Thanks all for your comments.
Iansoady - That's exactly the coil I bought for this exercise.
rhyat - You may be on to something. We didn't buy the sensor direct from the supplier, so it seems a bit of a cheek to quiz them.
limeyrob - My feeling is that it's something to do with the magnet installation in the rotor. Testing currently happening, hopefully that will spill the beans.
The bike runs fine with the magneto. The only issue being that it was difficult / impossible to get a low steady tick over. Anything other than fully advanced and the engine would stall. I'm fairly confident that the timing was correct, so deduced that the problem may have been a weak spark when the magneto was not in its best flux position (fully advanced). A contributing factor is the very low quality of replacement points available.
The up side of all of this messing about is that nothing is spoilt and I can go back to the original magneto / points system.
Pete
Is it manual A/R or auto? I much prefer manual but it has the disadvantage that it moves the spark away from the best flux point. With an auto the mag is always at its best point.
I've recently been round this loop on my RR (K2F). Hard to start and weak spark at idle. I have a box of old mag bits so I stripped it down, swapped a few bits, cleaned everything, put new brushes in and now its fine. There was no obvious fault but it made a big difference.
Ah ok.
Dare I ask if you were using a resistive spark plug cap ?
As you should not have anything in the way of the spark leaving the slip ring.
No resistance at all , just pure copper ht lead , plain cap , non resistor plug
limeyrob.
The bike has a manual advance / retard lever on the bars. I'm loath to send the magneto away for an overhaul as I can't categorically prove it's performing below bar. The electronic ignition seemed a low cost option and certainly worth a shot. Total spend so far is under £50.
rhyatt.
Champion spark plug. 10k OHM.
Copper core HT lead.
Pete
I assume the 10kohm is the resistive plug cap.
Do away with it.
Mags have never advocated a resistive plug cap.
Non resistive screw on plug caps available green spark plug company.
While that's coming in the post, just wedge the end of the naked copper lead in the fins and give it a kick , 1/4" gap should give a good blue crack of a spark
Yes, sending the mag away may not be needed, but a clean up on the bench may work. Your electronic ign set up looks nice, its just a pity that its doing something odd. Pardon my ignorance, but how are you keeping the manual A/R?
rhyatt.
We will try different HT lead and cap fixing options during the bench testing. The original lead was a heavy gauge copper lead with a spade end screwed directly to the top of the spark plug.
limeyrob.
The advance / retard works in exactly the same way as the original points. The whole assembly is retained by the same sprung ring as the original cam washer. There is a notch in the lower plate that is activated by the advance / retard cable nipple. So the whole assembly moves through 20 degrees or so as operated by the handlebar lever.
Pete
The reference to non usage of a resistive plug cap applies only to ht systems when used with a magneto only.
Any other coil based ignition can utilise a 10k plug cap
Thanks for explaining the A/R, that's neat, I couldn't work that out. Will you leave the mag armature in or run a dummy shaft? The mag could be generating a lot of electricity with no where to go.
If the points plate is missing there's no HT "load" generated.
Quote from: limeyrob on 25 Nov, 2025, 09:03 Thanks for explaining the A/R, that's neat, I couldn't work that out. Will you leave the mag armature in or run a dummy shaft? The mag could be generating a lot of electricity with no where to go.
The Thorspark electronic conversion retains the manual advance\retard in exactly the same way.
Quote from: limeyrob on 24 Nov, 2025, 12:06 Is it manual A/R or auto? I much prefer manual but it has the disadvantage that it moves the spark away from the best flux point. With an auto the mag is always at its best point.
I've recently been round this loop on my RR (K2F). Hard to start and weak spark at idle. I have a box of old mag bits so I stripped it down, swapped a few bits, cleaned everything, put new brushes in and now its fine. There was no obvious fault but it made a big difference.
That's interesting.My B31 has a rather 'Lumpy' tickover on full advance,but pull the lever back an inch or so,and it chugs away nice and evenly like a cement mixer.
Checked the ignition timing,but never thought about the magneto flux.
That's quite normal in my experience - full retard should give a better idle as on full advance the spark is occurring too far before TDC.
Retard should give a slower idle, but dunno about "better".