Making spark, cave man with a smart phone

[61BenC92]

great video on the 3 different twin designs. it talked about increased crankcase pressure on the 360 degree; good crank case vent is needed which might help with power and keeping the gaskets from leaking oil.


Sent from my iPhone using Tapatalk

 

[kbongos]

I got some of the new toys to scope out ignition. Just trying them out. The Hantek 6022 software sucks, but fortunately the openhantek6022 makes it bearable. The clamp on current probe seems to work good, the capacitive spark wire probe as well. But I have little to compare it to as this is all new to me. The toys:

A trace on my CM400T running with the 'discount' AC CDI unit:

A trace running with the original CM400T CDI unit:

I did a few bike runs with the discount CDI, 5-mile run up to 75 MPH, and one slower 7-mile 2-way trip. This is with switch to manually change from least to full advance. Mostly able to ride it on full advance, but have to rev 2-3K at stop lights if left on full advance to avoid stall. I have some better connectors now and want to make a better harness to more easily switch from one to the other for further tests.

I'll also be try scope captures on cars to see how it compares. Also working on better bench testing of smart COP units using a micro board(raspberry pi pico) running micropython to do the timing in a convenient way. Want to get solution working off just F mark idle non-advanced trigger. That would be used at start and idle as spark target, and advance would be calculated for higher revs. It would be nice to work with either existing pulse trigger circuit or add on hall sensors. The AC capacitive system has advantage of being able to trigger immediately with no care about dwell, whereas the non-capacitive requires careful timing of primary current(dwell) started 1-2ms before fire point.

 

[Maraakate]

Interesting on the dirtiness of the waveform compared to a genuine unit. Where do you attach the probe for grabbing the trace? And is the other end of the probe grounded? I have some 10x probes and an old Tektronix 20hz (I believe it's 20 need to recheck) that I use for alignment of antique radios and vintage hifi equipment. Would be fun to put it to some use on seeing the curve outputs compared to the books waveforms. I can make a video of it then as well for the CM400A for fun.

Great work as always, been a lot of fun seeing your work.

P.S. I'm gonna guess, but could be wrong. You do ham radio work as well? Most people getting into the level of electronics hobbyism such as yourself tend to also be into ham radio. I have an Extra class license myself. If you don't have a license, you should get one! Lots of electronics fun, tons of different modes and could be cool to have a digital mode conversation (or even SSB phone, though I'm mostly a CW/Morse guy myself) with a fellow vintage biker.

 

[kbongos]

Hey @Maraakate,
I'm not sure why the OEM trace set looks cleaner at this point. These tools are all new to me.
I will be re-doing some comparisons. I am making a better wiring harness for discount CDI.
This initial one I had the discount CDI mounted away from the coil by the fuse box, new harness
it will be placed directly under the OEM CDI making it a more fair comparison, and more easy to
switch from one to the other.

I am not a ham but I have total respect for that. When I was about 10 years old I got a 200 in one
Radio shack experiment kit and making a crystal radio was a favorite. Forest Mims did some great
educational writing for Radio Shack.

For the OEM CDI I managed to clamp the amp probe around the hot(yellow) wire going from CDI to coil. For discount
CDI it was a very long run of wire. So yellow is trace from high voltage spark plug cable with
this hantek capacitive clamp on spark plug wire. I also got a 20:1 attenuator from Hantek,
but it didn't seem needed at first try with spark clamp so just went direct with that. Not sure if it has
much attenuation built into it, it does have a small box that might have some electronics.
I think it is fairly simple and just a raw wire put next to spark cable could work as well, I have seen
a few you-tubes that suggest this, and people making there own simple pickup probes. It's a little
concerning that it could hurt the scope, but whatever, it's not that expensive.



The blue trace is the current clamp. It has two settings, 1mv per 10ma, and 1mv per 100ma. Something like
that. So at lower setting, 1 volt difference should be about 10A. We do see about +-.7V swing, so it would
be around +-7A. I have learned from others that 4A to 8A is typical of what goes thru primary.

So I made a minature gas tank for testing.

It's chrome! Pretty cool huh? So I'm trying out my new
mini gas tank, and low and behold gas spews out the overflow bowl spout. Hmm, I've been riding this
and have not noticed a problem, but I suppose it could overflow and I could be none the wiser as it sprays out
into the wind. Now after a while one gets clued into the fact that these bowl gas flow stops can be
a little flaky. So I made this super cool overflow detection sensor.

Now the gas can has a top. I've got this old school can opener that gives a clean cut of a can top
so the top is still usable. Coolest thing ever.



Idea is that at end of driving I can look at it and see if any overflow occurred.

Another random tip - To get pictures off phone wirelessly from Android phone to PC I use Android free
app Amaze and it's FTP server feature. Then on PC I use another free app called WinSCP.

 

[ancientdad]

Another random tip - To get pictures off phone wirelessly from Android phone to PC I use Android free
app Amaze and it's FTP server feature. Then on PC I use another free app called WinSCP.

I'm still in the dark ages. I use a USB cable with adapter to USB-C for my Galaxy A42 (poor man's Galaxy S series) to move pics to my computer... it feels like a lifetime ago when I built a Heathkit 7 transistor radio form one of their kits. It was a lot of fun but looking back it seems so antiquated now, mechanical tuner and built-in ferrite core antenna. Ran on 6 C cells and was only an AM radio. Radios used to be so simple that when I was about 12 I took apart a medium sized radio, figured out where the amplifier section was and cut the board to remove that part from the tuner, mounted it in a cigar box with the 3" speaker from the radio and the input wires (mono) soldered to a jack to connect one of those suction cup listening pickups you could stick on the outside of the old desktop phones to "listen in" on my sister's phone calls. I've made no further progress since then...

The stuff you guys are doing is so cool.

 

[kbongos]

I'm still in the dark ages. I use a USB cable with adapter to USB-C for my Galaxy A42 (poor man's Galaxy S series) to move pics to my computer... it feels like a lifetime ago when I built a Heathkit 7 transistor radio form one of their kits. It was a lot of fun but looking back it seems so antiquated now, mechanical tuner and built-in ferrite core antenna. Ran on 6 C cells and was only an AM radio. Radios used to be so simple that when I was about 12 I took apart a medium sized radio, figured out where the amplifier section was and cut the board to remove that part from the tuner, mounted it in a cigar box with the 3" speaker from the radio and the input wires (mono) soldered to a jack to connect one of those suction cup listening pickups you could stick on the outside of the old desktop phones to "listen in" on my sister's phone calls. I've made no further progress since then...

The stuff you guys are doing is so cool.

Yes, I have a nice Samsung Note20 phone, but the USB connector works like crap. I'm happy if I can get the USB C charger going, I don't touch it until charged. Actually picked up some cheap wireless charger pads as backup just in case USB C fails completely someday. I figure the less I use it the better. So getting the Android Amaze FTP server going with my WiFi along with WinSCP client on PC is a little intimidating to initially setup, once you get it going it is quick to use and I wonder how I lived without it. WinSCP has a two panel file manager, like the old PC App from Peter Norton called Midnight Commander back in MSDOS days. And after reading about Motorcycle Twins I see references now to an old Norton Commando motorcycle, so I guess that's where Peter got the name for his MSDOS file manager program. Linux has got a copy cat program Midnight Commander program as console file manger, and you can even find a Windows version of it, if you are into that sort of thing.

 

[ancientdad]

Yes, I have a nice Samsung Note20 phone, but the USB connector works like crap. I'm happy if I can get the USB C charger going, I don't touch it until charged. Actually picked up some cheap wireless charger pads as backup just in case USB C fails completely someday. I figure the less I use it the better.

My wife and I both have the lower end Galaxy phones, hers is the A13, and both have the USB-C port but we've had no issues with them. My only complaint was the phones came with a short C to C cord which is useless, no chargers either. Not sure what they were thinking of since our chargers are typical full-sized USB. Good news is the internal, non-replaceable batteries (which I hate but my original replaceable S5 battery lasted almost 8 years anyway) are 5000 ma as opposed to the 3500 ma in the S5, so battery life is much better now. Ended up buying a couple of cables that have C adapters on the end so they'll work with the previous micro USB too. We need good battery life for using nav while in the mountains each October.

 

[kbongos]

A few more traces, two with discount CDI, two with OEM CDI. Hantek 6022 captures 10K samples. The A/D readings are only 8-bit(0-255) resolution. You can use this scope software to zoom in and out. You can record screen snapshots(what you see here). You can export datasets as CSV or JSON.

With just a single trace it captures 20K samples. The scope locks up regularly while taking these, where I need to unplug USB hardware and restart program. I would guess this is either due to me running laptop battery(less reliable USB power?) or due to the high voltage readings causing interference, as it does not seem to have this problem on the bench with lower voltage recordings.

Did another 2x4 mile lower speed trip on bike with discount CDI and toggle switch, ran good. It's now mounted up in reasonable spot and I can switch from discount to OEM CDI fairly easily. Advance toggle switch has longer wire so easier to reach.

The first channel measuring HV spark is not real regular, I would guess it is just hard to measure with the extreme high spiked voltage. After the high spike you will notice some offset. In a number of other posts by people recording these, this can indicate(or they say it does) a continued spark duration. I have no idea if that is what we are looking at here. Most of the ones I see from other people(on youtubes) are not CDI but with cars and so the odd bounce we see in the blue trace measuring primary current I believe is unique to this capacitor arrangement.

 

[kbongos]

Just want to share some pics and new info as I go. First I want to leave a link to Thierry in France who has done some hobby work making ignition gadgets for motorcycle. His site is at: https://transmic.fr/ His latest work includes TCI and CDI controllers using ESP microcontroller that provides WiFi connectivity. He mainly has done Yamaha motorcycles. Anyone ever heard of those?(That's a Honda joke).


The following screenshots start with the Smart COP unit. I use a Pi Pico controller running micropython to arrange primary charge time(dwell) of 1ms, 2ms, 4ms and 8ms. At the end I have few traces of the Dumb COP unit where I arrange a 2.2uf 400v cap and just brush the coil-cap end to ground as briefly as I can. I just got some of these 2.2uf 400v caps, I have some newer OnSemi IGBT switches coming in about a week that are specifically made for ignition control. But for now just some crude testing brushing the contact, kindof like a points contact would do but more random. There is also a adhoc resistive ballast added to the dumb coil circuit. This consists of about 8 feet of a spool of thinner wire that is around 1-2ohm. I had used this previously in trying to test car battery 'wellness'. You can get larger watt small ohm resistors, but for some testing(like loading a battery), you need to dissipate some serious heat generation. I also use a big starter solenoid as a test load. The incandescent light bulbs of various wattage can also serve as good test loads.


The ballast is added mostly to avoid too much current with these crude tests. I get quite a lot of spark where I brush the primary on briefly. Interestingly it does not generate a spark(at the plug anyways) if I leave out the capacitor. These circuits work in complex ways where the cap and inductor work together and this includes a flyback or kick. So even a point system using 12v on primary generates a 200-400v primary pulse at this spark time. It does take longer to get current flowing in this 12v switch arrangement. The CDI is a faster arrangement due to running primary side source at higher voltage. And generally these 2 cylinder bikes run at higher RPM to where this charge time can be a factor.


You can see somewhere around 4ms the primary current limits itself(somewhere around 8-16A). It appears as this sharp cutoff, I believe it just hits a natural limit of the coil to where it just acts as a resistor. I have also learned from watching a lot of diagnostic utubes that 2 plugs firing in series as 'wasted spark' arrangement does not split the 'spark' energy in half between the two plugs, and this is because they have a much easier path to spark over in the uncompressed lean cylinder.


Smart COP traces at 1ms primary charge:

Smart COP traces at 2ms primary charge:

Smart COP traces at 4ms primary charge:

Smart COP traces at 8ms primary charge:

Smart COP traces at 2ms primary charge:

Smart COP traces at 4ms primary charge:

Smart COP traces at 8ms primary charge:

Dumb COP traces sparking manually:

 

[kbongos]

Here we capture a few MAP sensor traces hooked up to the carb sync inlet access ports. I was thinking this might offer a less prehistoric method to sync carbs. I have never synced my carbs using these ports, I don't have the tools and just did the best I could with the bench sync. But the thought occurred that these common MAP sensors could be useful for sync or other diagnostics. This is just an initial test, but does show decent readings. I made a adapter out of a plastic barb fitting by just screwing it in along with a short tube to the sensor. Notice how the pulse is cylinder specific, synced to every other spark fire(wasted spark). The sensor is just a inexpensive one I picked up on ebay to see how well it could work.

 

[61BenC92]

In post 59, where you talk about the spark traveling the easier path to the lean, uncompressed cylinder, does this mean that one is at TDC on exhaust stroke yet it gets the stronger spark? 180deg crank engine.


Sent from my iPhone using Tapatalk

 

[kbongos]

In post 59, where you talk about the spark traveling the easier path to the lean, uncompressed cylinder, does this mean that one is at TDC on exhaust stroke yet it gets the stronger spark? 180deg crank engine.
Sent from my iPhone using Tapatalk

@61BenC92, Something like that. This was something of a correction to a previous post #44 where I said I believe the 'waste spark' splits the spark energy between the two plugs. One spark is not needed(wasted) on the opposite exhaust stroke. I have since learned that spark for the compressed with gas state takes more energy than when it is in non-compressed state, so the assumption that the energy is split evenly was wrong. This was part of pondering the benefits of going COP where you would have independent coils for each plug. There is also a spark duration, and you can see it in the traces as a period of time after the initial HV jagged line. The CDI traces earlier above you can see it is like a .4 msec event, while my COP style tests it shows some trailing event much longer 1-2msec that I am told is the spark sparking.

 

[kbongos]

Here is something to wrap your head around. They have wasted spark single cylinder systems. So one cylinder sparks at both power stroke and opposite waste exhaust stroke. Why would you do this? So you can trigger it from the main crankshaft(spins twice for 4-strokes) and not the camshaft(spins once per 4-strokes). So my CM400T does it's spark timing from the main crankshaft, unlike most of the older twins with points that get trigger from camshaft. So if I wanted to NOT do waste-spark I would need to know which cylinder is on power stroke, and the camshaft is the most obvious place to get that signal.

https://www.jetav8r.com/Vision/Ignition/CDI.html - some interesting reading on ignition.

 

[61BenC92]

ok, I follow you now. Longer spark or multiple spark discharge would be a great advantage for a more complete burn which should make more power, run smoother less fuel smell from exhaust. To pick up from cam, maybe you could adapt a pick- up ring and cam position sensor / hall effect from some modern engine. I am interested in all this because I am considering adapting a modern dirt bike fuel injection/ elect. ignition system to my 125cc Benly to increase reliability.


Sent from my iPhone using Tapatalk

 

[kbongos]

ok, I follow you now. Longer spark or multiple spark discharge would be a great advantage for a more complete burn which should make more power, run smoother less fuel smell from exhaust. To pick up from cam, maybe you could adapt a pick- up ring and cam position sensor / hall effect from some modern engine. I am interested in all this because I am considering adapting a modern dirt bike fuel injection/ elect. ignition system to my 125cc Benly to increase reliability.

You would think that applying some new technology to old systems could be useful. The EFI sounds like a big leap, and it is, but it could be worth it as a retro upgrade. I have been learning about that. They sell kits to do this. It does take some skills. And it's probably easier just to buy something newer with it built in. I read they made a Honda CX500 with EFI turbo and computer control in Y82. Here is a youtube of someone retro fitting a EFI on a nija: https://www.youtube.com/watch?v=NYsP2JYXbGs&t=1001s

So I was over working on a my buddy's Y81 XL250 again, upgrading to LED lights. It's a 6v system, but since the battery is toast it is more like a 1 to 3V system and a pulsating 6v system. Added a new 12v blinker that did work. Still have a minor problem with brake light. Gave me a chance to get some scope traces of his CDI ignition, here are some traces:

 

[61BenC92]

Finally had time to watch that efi video. Cool how he made it an all in one. I was planing to use a submerged pump, no return style. I had not considered the coolant temp./ air cooled head temp. sensor difference or the dual O2 sensor need.


Sent from my iPhone using Tapatalk

 

[kbongos]

So I got my IGBT switches from ebay. Rigged up to the pico board and arranged to fire the Escape 2 wire cop coil with 1ms,2ms,5ms.
'20pcs NGD8201AG NGD 8201AG Integrated Circuit IC' https://www.ebay.com/itm/284061806406
Here are the caps I picked up: '50pcs 225J 2.2uF CBB 400V 225 Metallized Film Capacitor P=20MM' https://www.ebay.com/itm/293687769937
I couldn't find a direct matching schematic for IGBT, ones close showed they should trigger with a nice low 2-3v gate level, but these required
something higher like 6-9v. Maybe they are reject floor sweepings, who knows.
Anyhow, it did seem to work OK firing similar to smart 3-wire COP(those worked fine with 5v trigger).
Basically taking about 1-2ms to get a some good amperage(5-10A) going on primary. And it peaks out current somewhere around 3-4ms.
And they didn't self destruct or smoke, which is a concern. It looks like they do add specific protection diodes in this
specialized IGBT switches to deal with that, which is nice if it contained everything needed.

I'm thinking something like this could be useful for point replacement, along with micro to do dwell and advance and hall sensors for
trigger. And it might be useful if the electronics could be explained and handled and leave the mechanics to the mechanics to arrange
the hall sensors to replace the points.

So for the hall sensor triggers I think they commonly use a raw hall sensor($0.50 or so) with a small magnet attached to rotor.

Another arrangement is where they make probes with a magnet built in to probe, and a hall sensors as a probe that picks up on a rotor
tooth and use a rotor with a number of teeth along with a missing tooth gap to indicate the absolute position. This you find in newer
arrangements typically including EFI. People do experiment with DIY solutions like https://wiki.speeduino.com/en/home, or https://www.diyautotune.com(megasquirt).

My goal on my CM400T would be to get 2 smart COP's along with micro working on bike doing triggering and advance, with just the 1 F-idle trigger.
It might be nice to learn how to apply the newer stuff(IGBT's) to a points based older bike as well.

I may try this with my OEM CM400T coil just to see what I get. My understanding is the coils maybe significantly different.

Here are a few traces of this IGBT, the cap and the 2 wire dumb Escape COP:

This is all just on bench with plug in the air, not in bike or anything.

 

[Maraakate]

Yes, I've seen people use the speeduinos on some other 80s bikes. You have to go to the junkyard and look for throttle bodies that will fit the bike and/or fab up adapters for this. Then, you have to get a fuel pump, fab up mounting for that as well and the relay for it. Then you'll have to fab up something for the sensors. Practically requires access to CNC.

Not impossible. I have seen some good speeduino threads before on such conversions, but it is a massive undertaking. If you're not a fabricator and don't have a CNC it will be incredibly difficult for you. That's what has stopped me from attempting it. However, I think this will be the way forward in 20-30 years when you can't get any reliable parts for the carbs any more. Specifically the float needles. Aftermarket ones are all junk. I could see such a thing becoming a kit on more desirable and popular bikes such as the CB750 and early Goldwings.

 

[kbongos]

I got my hands full just learning about the ignition. But it is interesting to see that they make kits to add EFI if one were that ambitious. Personally I would probably just buy a newer bike with EFI. But it is useful to learn about the details of this stuff. So for example I fixed my Escape that had the fuel pump die, cost me about $20 bucks. Where for most people this would be $1000 shop bill. Pleasantly surprised you can get to the pump in the gas tank thru a trap door under the back seat instead of having to pull the gas tank off.

 

[Maraakate]

I got my hands full just learning about the ignition. But it is interesting to see that they make kits to add EFI if one were that ambitious. Personally I would probably just buy a newer bike with EFI. But it is useful to learn about the details of this stuff. So for example I fixed my Escape that had the fuel pump die, cost me about $20 bucks. Where for most people this would be $1000 shop bill. Pleasantly surprised you can get to the pump in the gas tank thru a trap door under the back seat instead of having to pull the gas tank off.

In some cars, yes. Others you have to drop the whole tank... like when I had my 91 LeBaron. Really OT, but I had to go through 3 aftermarket pumps until one actually supplied enough pressure for that car. First one was a Bosch, so I thought it couldn't be that. Found someone at the shop who had DRB III (anyone remember that?!) to scan things to narrow it down to not being something else. Second one also was junk. Third time was the charm. I had someone warn me that the aftermarket pumps for older models are pretty terrible QA. I figured it was an exaggeration, but apparently not.

 

[Maraakate]

Do the AC CDI on these bikes put out more volts? Or asked another way, is the coil unique to these as compared to a 12v point systems coil? So is the voltage/current generated by these much different than say a 12v point system? I believe the AC CDI gets some higher voltages(higher than 12v, like 100's of volts AC) from the stator. LDR mentions changing coils in his use of a modern retrofit ignitech(?) that uses the 12v DC battery system. I'm just wondering if the coil has to be different if not run from the AC stator because perhaps it does not put out as much when coupled to 12v system. I tried to measure some, but it's not straight forward, the cap is embedded in the CDI, the output of the cap is grounded thru the coil primary(so looks like ground mostly, other than a spike at spark time). Spikes are hard to measure. When I do a search for coil replacement, the search results look like the coils are probably only used on these motorcycles with AC CDI, and perhaps not used on the older bikes with points. Looks like there are plenty of fairly cheap aftermarket coils I could buy for the bike. Just wondering if they are appropriate if say one wanted to experiment with a retrofit to newer 12v DC system say.


Maybe I'll try bypassing all of this and try some COP wires scavenged from the junk yard. That could interesting.

I'm assuming at this point because of the 400V capacitor that gets discharged its probably around 100-400V going from the yellow/green 2-pin from CDI to coil, but I will verify shortly.

 

[Maraakate]

Didn't read anything on AC or DC. Oh well, guess it requires a scope. I'll have to get an HV probe for it one of these days.

 

[kbongos]

Didn't read anything on AC or DC. Oh well, guess it requires a scope. I'll have to get an HV probe for it one of these days.

Hey @Maraakate, Good job on the latest CDI reverse engineering! I've been trying to reverse engineer this electronic blinker, but it's a pretty good challenge. This is just to figure out how it works. I ran across this web site that I think the circuit is roughly equivalent to: http://techlib.com/electronics/flasher.html

I don't think the coils used with CDI are very different than the ones used with points. I think they both develop a short pulse of 300-400 volts on primary at spark time, we just can't see it in my scope measurements. It was a little confusing to me because the points or electronic switch type are just working with 12V, but I believe when it is fire time then it let's it snap and it's generating this high spike of 300v or so on primary at that time. I plan on pulling my cm400t stock coil and see how it compares to the COP dumb coil on the bench test, trigger it with the IGBT(and cap) same as I do the dumb COP coil. I believe this is roughly equivalent to a points system, just replace the points with this IGBT switch.

 

[Maraakate]

Hey @Maraakate, Good job on the latest CDI reverse engineering! I've been trying to reverse engineer this electronic blinker, but it's a pretty good challenge. This is just to figure out how it works. I ran across this web site that I think the circuit is roughly equivalent to: http://techlib.com/electronics/flasher.html

I don't think the coils used with CDI are very different than the ones used with points. I think they both develop a short pulse of 300-400 volts on primary at spark time, we just can't see it in my scope measurements. It was a little confusing to me because the points or electronic switch type are just working with 12V, but I believe when it is fire time then it let's it snap and it's generating this high spike of 300v or so on primary at that time. I plan on pulling my cm400t stock coil and see how it compares to the COP dumb coil on the bench test, trigger it with the IGBT(and cap) same as I do the dumb COP coil. I believe this is roughly equivalent to a points system, just replace the points with this IGBT switch.

No idea. I just assumed it sends the ~400V OUT from CDI to coil through the yellow and green since it's wasted spark and it's confirmed through reverse engineering that green is truly ground from CDI to Coil. I figured putting a probe on yellow the other on green with an analog meter would show ~400V, but it didn't so maybe I am probing it incorrectly or it needs an HV probe hooked up to a scope to see anything interesting. I was hoping that could be used as a test for people to see if their coils were getting voltage to them.

What I might do later is hook it up directly to an early prototype CDI that I mounted on a 2x4. Easy access to attach clamping/finger probes and see what happens. If I bork it and fry something, no big deal since I have no real use for it anymore.

I was reading on Mike Nixon's site that the COPs put out less voltage than standard coils, typically, because it's all computer controlled with fuel injection so they don't need as much energy to make ignition happen as opposed to carb. No idea if this is true, but Mike's got a lot of good real world advice so I think it's safe to assume he's right. And I'm assuming you researched to find something that meets the specs so it would work.

Something I'd like to do in the future is build (or find cheaply) a 400V-600V switching supply and make some mounts for a jig that you could put a coil in it and test for spark. Not just jumping the gap on the plug, but much further. Similar to the old coil testing units. Resistance readings are helpful to determine if it's failed or on it's way out. But, things fail once real voltage goes through them and not the pitiful 0.5V or whatever it is from an ohm meter. Could be even more interesting if the PSU was adjustable so you could see where it breaks down.

For fun, you should try and repeat the same tests I did with an analog volt meter on the blue and white at crank with kill switch off, and during run at idle and throttle. I was able to do that test so you can easily verify the stator is working when actually doing things and not just the resistance readings. I updated the Hondamatic diagnostic test https://www.vintagehondatwins.com/f...s-for-CB-CM-400-450-Automatic-Trans-1978-1983 let me know what readings you get. I was even able to read the smaller voltage increase for the fixed pickup coil as well. It works for the advance coil, but the reading is so minuscule I didn't bother to post the readings on that, yet.

 

[Maraakate]

We like fun, right? Got bored at lunch and pulled the scope out. It's rated up to 400V so I can hook the probes up to it. I didn't do two channels, just one. Simply put the probe on the yellow at the 2-pin coil connector, the ground to well... ground! Here is the wave form, notice around 22 seconds when I put it into gear and the time of the wave changes going from 7.5 to 15 degrees showing this works (and that your change relay and light blue connector is working):

Here is the wave form for Low and High RPM. Probe to White and Ground, then Probe to Blue and Ground. Notice it caps off at a certain RPM range for the low RPM coil and high RPM has much more to go. Something you can verify now with analog meter, but still neat to see:

Finally, here is the Advance Coil. Probe to Pink and Ground:

Didn't want to get fancy with the probes on the Orange/Brown for pick up coil. Will try to do that some other time. Had to have my phone sit on the crawler seat, so it's a funny angle, but you get the idea.

I also have very little experience in using a scope. Unless a service manual dictates to me the settings on the scope for a waveform alignment (i.e. DC Zero Voltage Offset on a Vectrex, etc.) I just adjust the settings until I can get a waveform I can reasonably see. You can see some of the triggering isn't set just right so you see those double lines at lower speeds and when I get it some RPM then it stabilizes out. But, for what we're doing here I think this is good enough to get an idea how it works and what it should more or less look like.

 

[ancientdad]

Interesting stuff. When I worked for the local reman center for GM, Ford and Chrysler radios and ECMs back in the early '90s, I got to watch the bench techs work on evening shift when I came in with stuff to get into the queue for repair. It's fun to watch. I probably should have stayed in electronics school in the early '80s but working full time and going to school 16 hours a week too put a big strain on my health and study time allotment while living by myself back then, so I dropped out after 4th quarter because my grades were falling and I'd still have that loan to repay. But this stuff has always fascinated me.

 

[Maraakate]

Interesting stuff. When I worked for the local reman center for GM, Ford and Chrysler radios and ECMs back in the early '90s, I got to watch the bench techs work on evening shift when I came in with stuff to get into the queue for repair. It's fun to watch. I probably should have stayed in electronics school in the early '80s but working full time and going to school 16 hours a week too put a big strain on my health and study time allotment while living by myself back then, so I dropped out after 4th quarter because my grades were falling and I'd still have that loan to repay. But this stuff has always fascinated me.

I never went to school for any of it, beyond taking Electronics I as an elective in High School. If there's clear directions to follow I can do that, and can more or less diagnose faults eventually and obviously I've figured out how to reverse engineer somewhat basic circuits (true dual layer boards would probably make my head spin, unless it was a small circuit) but reading the schematic and following how it works and why is somewhat foreign to me. Depends on what it is. Luckily I've been able to have friends who can explain it to me when I am curious. But, it hurts me when following the path of a circuit to see where it fails. Old radios it's a bit easier. You can follow where the signals go and see if it's amplified or lost, but other stuff I have no idea sometimes and takes me much longer than a person who does it as a profession to narrow down the issue.

We're lucky the CDIs in these things are discrete components. If it was a single custom IC that had a "blob" over it to hide what it was then it's a lot of work to get it x-rayed and decapped. Then you have to know how to make sense of what that bare miniature circuit is, if it has firmware (which it will) then you have to convert that back to a binary form in order to reprogram it to something else. Lots of work. Usually the only stuff that gets that done would be for something very popular, like the YMF262 IC that was used in millions of MIDI controllers and in turn arcade/video game systems and even computer sound cards in the 90s:

https://docs.google.com/document/d/18IGx18NQY_Q1PJVZ-bHywao9bhsDoAqoIn1rIm42nwo/edit

 

[kbongos]

Did some scope traces on bike running at idle. First few here are discount CDI primary voltage in ch-2:blue,
ch-1:yellow trace is fire signal I'll call k15. Both probes at 10x, so 5v is 50v. It also clips at 5v range and
primary voltage is clipped, so larger than 50v. I have a 20x attenuator but of course could not find it at this time.

Next few traces are OEM CDI, same primary voltage ch-2:blue, and K15 ch-1:yellow.

Next is advance-pulse signal, call it K43, with same K15.

Next is this advance stator signal I will call bias.

Next one is one of yellow alternator wires.

This next is white wire AC supply and K15

This next one blue wire AC supply and K15

Next I took OEM coil off bike took it to bench and arranged to test with IGMT switch, measuring primary current(ch-1),
versus secondary clamp-on volts. Coil to 12v, other side IGBT switch pulls to ground for various dwell times,
then let's it open at fire time.
First few here are with 2.2uf 400v capacitor added as snubber to IGBT switch.

The next ones here are with no cap/snubber.

I did take the old analog tek scope out was able to mostly verify that hantek discount scope was not lying to
me. Unfortunate the primary Volt measures here are still clipping so left to speculate what value it reaches.
I did arrange my DIY 25x(or something like that) attenuator for the white and blue wire AC supply
voltage test.


OEM coil worked similar to COP coil I'd say.

 

[Maraakate]

The "blue channel" for blue and white looks practically the same as mine. All cool stuff.

 

[Maraakate]

Well I was all wrong, it just needs a 12V battery for these tests and a capacitor:

https://www.youtube.com/shorts/i-Z_NxmbafI

Building one is simple. Yellow from coil to positive, Green from coil to 2-4uF 400V film capacitor to negative. Ground part of spark plug to negative. Put a splitter on the Green from coil with a jumper. Touch that jumper on the battery terminal repeatedly it will spark. Now to build something that would trigger it more frequently than by hand...

 

[kbongos]

Hey Maraakate! Good stuff. I captured a few traces like that on the dumb COPs above, as you experienced it's a rather noisy affair,
where you get plenty of spark from the wire you brush to simulate the points. And that's with the cap which is pretty much there
just to act as a snubber to dull out the points like spark.


I made some initial spark dwell circuits, one with a 555, another with a cap and schmitt trigger inverter gates, but I wasn't real happy
with either of these. So then I got some of these ON Semi IGBT transistors that are targeted at ignition, and the Pi pico RP2040
little ARM boards that I use for these timing tests. It's rigged up to do about 20sec of 1ms, 2ms then 5ms dwell.


I figure now I can experiment more aggressively, given we have backup now, your new replacement CDI, and I have this discount one rigged up
to function(with manual switch for advance). So I'm not so afraid of blowing up my OEM CDI. And I mostly proved to myself
that these dumb COP coils are not too different than the OEM CDI coil. So I rigged up 2 of these Escape COP coils to the OEM CDI
and ran it briefly. Tried the coil primaries 2 ways, in series and in parallel. Both started and idled for a bit.
Wiring is very sketchy so didn't run it long. Maybe I will find time to make it less sketchy enough to try on some rides.


I'd also like a bench test setup where I have an adjustable spark gap to see how long(distance) of a spark a system can produce.
Honda mentions it in FSM as a test method, and it seems reasonable to me.
It would also be cool to verify 'spark duration' somehow. With a long dwell time you can get up and over 2ms of what
should be a spark duration(it's much shorter on a AC CDI setup). My 960 FPS high speed phone camera might be able to capture
2 sparks, but that is still crude. And of course someone has already I am sure done this and it's just a matter of finding
it. Here is some study done back before I was born at 40K FPS. https://www.youtube.com/watch?v=HGWKdfSF6zg


It would appear that the CDI spark is shorter with higher voltage than the points or IGMT. Notice how when the cap is used
with CDI or IGMT the current and voltage bounces, some oscillation. Whereas no oscillation with IGMT and no cap.


Here is a trace of Hantek current probe on it's high amp 1mv/100ma setting for the starter current. This is pushing limits
of this current probe(say it is good for up to 65A). This is roughly showing we get an initial surge of around 100A and
then it tappers off to 50A or so. You can see some peaks where it is drawing more due to piston compression. I have the kill
switch on so it does not start, so starter is cranking at around 400 RPM is what I figure.

And here is a trace of current at high setting of weller 100W solder iron with probe around solder tip.
And again on medium 1mv/10ma you can see it clips out the reading.

 

[kbongos]

Well my buddy's Harley has been out of commission for over 2 months due to ignition problems. It's funny how people are so certain they know what the problem is but then they really don't. They throw parts at it. Buy a new battery, or coil, or ignition module. Turned out just to be some flaky wiring I think. I took some measures with my scope, tinkered with some connections and its starting. Next day it is not. So I come back, hook up scope, clean some connectors and then it's starting again.

So I rigged him up LED's that flash on the primary of the coil and on the pickup sensor wire between the CAM and the Ignition control box. Here is a video of it:
https://www.youtube.com/shorts/hAm7XlYoXyI
The next day I came back and cut off 3 pin connector between CAM and ignition module I thought could be a problem and just soldered them together. Added a third green LED that lights slightly to show 5v supply from module to sensor in CAM cone. So these LED's will hopefully provide useful information next time it won't start.

Here's a pic of my baby scope measuring the sensor reading from the nose cone.

I did capture a few primary amps and secondary measure with scope.


Here's my buddy's buddy's chopper that he hasn't ridden in years. It's Texas chopper, with a S & S engine, which is like an American made copy of a Harley engine. It's all about big bore in cubic inches. I think it looks pretty cool and makes for good garage furniture.

 

[ancientdad]

I think it looks pretty cool and makes for good garage furniture.

LOL... to me, looking good is one thing but this bike is on the outer fringes of form over function. I couldn't ride that bike more than a few blocks without being in physical misery, but to each their own. Cool idea to set up indicators for potential connectivity issues, hopefully you'll be able to track down his real problem (which I'm sure is rooted in custom wiring)

 

[kbongos]

Cool idea to set up indicators for potential connectivity issues, hopefully you'll be able to track down his real problem (which I'm sure is rooted in custom wiring)

The Harley continued to have no spark issues. I believe the add on LED's were helpful. The LED and resistor across primary of coil is a pretty easy addition. Tough to say how robust this will be given that there is some spikes of high voltage, up to 300v, that could burn out the LED. The idea partly came from his buddy who said he had used a trouble light across primary to troubleshoot and would get a flash indication that way. I originally snuck a single LED and resistor on the 3-wire connector(with some thin wire wrap wire) between nose cone and ignition module. But I thought maybe that connector could be the problem so came back and cut it off and soldered it, and added back in the LED and added yet another green LED to show the 5v supply. He had no-spark issue again, and I examined the nose cone electronics, put back together and running again. I poked and found the nose cone electronics potted with softer material and plucked out a lot of it to examine it better. My assessment is that is where the problem is. I want to either replace it, and/or clean it off better and check for bad solder joints.

Here are some pics. The nose cone, where it has sensor, I believe it is probably a hall effect sensor and it reads 2 slots in the tuna can like rotor. The sensor has a slot in it where it sees these tuna can slots and translates that to a low signal. Here is that 3-wire connector, with wire wrap snuck in to get scope reading:



Here is LED snuck into to flash on that connector:



And after cutting that connector out and soldering, I moved and added green LED up closer to ignition module where there was already some wiring connections I wanted to check.



A pic of the LED and 5k resistor added to primary coil



Some pics of the CAM position sensor in the nose cone



The sensor with filling mostly removed for inspection



I believe this could be replaced by these that go for around $50:



Buddy's buddy had these old parts



So apparently the mechanics remained mostly common, and you could switch back to points, but that would need mechanical advance with springs and weights and rotor cam for points. So this electronic system does electronic advance, there is ignition module under seat the 3 wires go up to. This includes some other wires for maybe kill, but also a vacuum pressure sensor to do something with advance. Other electronic modules seem to put everything on controller in nose cone instead of the separate box. Here's the box:



But some of those modern replacements they want like $300 or more for.
This is a wasted spark arrangement - meaning a single coil(2 connectors primary). These harley bikes can optionally use a dual coil pack(3 primary connections), they call this dual-fire, as apposed to single fire(or maybe other way around . The more expensive controllers can be configured to do either.

Some of the controllers have a LED built in that turns on for fire time and can be used for initial static timing, as that cam plate can be adjusted, just like points system.
Notice the tuna can slots are not 180 degrees apart, but somewhat odd based on the twin cylinder slightly odd angle.

This only has a single carb, which looked rather easy to take off and mess with compared to the dual carb systems typical with Honda



There is a lever that is pulled. I called it a choke, but was corrected, it is referred to as a enrichment control or something.

If you have read thru this entire post you are now officially certified to work on Harleys. God bless.

P.S. He is up and running. I told him to take some tools, at least some screw drivers along. I hope he doesn't get stranded.

 

[ancientdad]

Well, I'm familiar with that carb having cleaned one up for the garage-find Harley I had about a year and a half ago, simple enough to get working (in my case after sitting covered up for 15 years in the PO's garage). Fortunately I didn't have to bother with any electrical, it fired right up and ran great. Your buddy's situation is so very different being custom, I wish him good luck with it.

 

[kbongos]

Can anyone give me opinions or experience with transistorizing your points. Good? Bad? Ugly? Like Hondaman offered or DIY HEI. Some of the benefits touted are longer points life, easy 'upgrade', can be removed easily to go back to original. Down side I would say is it is a rather small upgrade, you still have contacts, mechanical advance. But upgrading those is quite a bigger overhaul away from original. I don't have anything with points, but maybe someday I will as these old bikes are pretty cool(thus VHT), mostly points on the old bikes. The technology keeps evolving, like these newer IGBT transistors are better than what they were 10 years ago. It might be nice to come up with some shared ideas. I could help with prototypes, circuit design. But first things first, it would be nice to gauge how useful this is. So the auto industry used these with this HEI module, I guess they thought it was useful. I do like the idea of keeping it simple where you could easily go back to points when you get stuck on the side of the road due to your burned out transistor


Jay, thanks for perking my interest in the spark area again, glad to see you experimenting with this stuff. Thought I'd ask over here just to avoid to much noise in your recent post, I tend to get rambling.

 

[ancientdad]

With two of you working on the idea I'm sure there will be more progress.

 

[LongDistanceRider]

The problem with using the GM HEI modules is the bulk and the heat generated requires a large heat sink and space is limited. I had the Acacia system on the 350 for awhile and the only place for the module unit was in front of the battery box, just barely fit after some grinding on it.
That system used the stock advance with magnetic switches instead of points.

 

[kbongos]

The advances in IGBT(transistor) has reduced the heat generation problem to a large degree, so hopefully that problem goes away. The other nice thing is there are IGBT's like the one I mention trying that are specifically designed for ignition and include the critical protection diodes needed to avoid frying them.

The problem with using the GM HEI modules is the bulk and the heat generated requires a large heat sink and space is limited. I had the Acacia system on the 350 for awhile and the only place for the module unit was in front of the battery box, just barely fit after some grinding on it.
That system used the stock advance with magnetic switches instead of points.

 

[kbongos]

My previous post on just using points, adding transistor to offload points load I no longer think is a good idea. I will explain more below. I would also agree with people posting to get rid of the mechanical advance. I'm with you now on that. A microprocessor solution is needed. I'm going to work toward that. My target is to get something inexpensive that can work with points as a input, or with other input triggers, like my 81 CM400T with the pulse coils, or whatever. Would be nice if it worked with existing coils or with newer coils(CNP say).

At least for initial prototype I want the micro code to be as simple as possible. I will focus on this initial start/idle F signal spot, then work in code to calculate an advance at higher RPM.

I have over the last few days come to a few conclusions. One is that only adding transistor/ignitor to take load off points is a bad idea. Using points in general is a bad idea. The use of points on a CAM is crazy caveman stuff. It gives you a variable dwell time based on RPM, where this dwell time is 20 times the length of time at slow RPM compared to high RPM. So at high RPM it is maybe 3ms, while at slow RPM it is 60ms. I suppose the old coils were made big enough or in such a way to tolerate this abuse. But if you notice, newer coils are much smaller and lighter and modern electronics can deliver a carefully crafted 2 to 4ms dwell time. The fact that you can stop your points based engine motorcycle(but leave your power on) and the points can get stuck in the on state and burn out your coil or drain your battery is testament to how cave man crude this old system is.

I also tend to agree that advance should be handled electronically and not by some old analog weights and springs. Even the CDI approach at that time I dislike - using odd impossible to understand analog circuitry.

The existing commercial solutions are not good - reuse the existing advance mechanical method? You call that a feature? It's a bad limitation. The commercial solutions available to do the advance work have other problems. There is no focus on 40-50 year old bikes. They are over priced. They are mostly proprietary.

What I envision is a box with a micro and circuitry to work with any type of output coil arrangement, and any type of input cam/crank position sense, that includes points, pulse signals, digital signals derived from opto or magnetic means. A focus on only these old motorcycles. It should be affordable, like less than $50. It will preferably be an open design so you have many people that can build off it, change it, support it.

 

[kbongos]

Got blink program compiling! Wahoo! Blinking lights!



Existing CDI setup, red is start/idle pulse coil signal, black is advanced pulse coil signal,
yellow is CDI HV charge coil. red and black go to toggle switch that selects between advanced and normal.



This is what the bike pulse coil signals look like.



Initial breadboard to convert pulse coil signal. Settled on optoisolator as easiest solution. Also can easily arrange to trigger on negative(first) or
positive spike(second).

Some sketchy initial examinations of the patient.



Initial breadboard mess


Breadboard of current setup



Try it out on bike



Circuit so far


Even more sketchy live trial. Have not let the smoke out yet!




Info on the 453 scope: https://w140.com/tekwiki/wiki/453

It all started with the erector set, check it out: https://www.eliwhitney.org/museum/-gilbert-project/-collections/erector-sets

So I got it working(briefly)! This is going to revolutionize 1970's ignitions everywhere! Well, maybe not, but we can try.

So my CDI setup has the advantage that it fires at the exact moment I signal start fire to CDI. This is due to high voltage capacitive discharge getting charged to HV beforehand.

Now if you compare that to points or modern 12v delivered to coil, this requires a controlled DWELL time. So you have to know ahead of time when to start dwelling(building current in coil) before the moment you wish to fire it off(end of dwell).

So with my CDI setup, my dwell time is not part of the equation, it fires at start of this signal. If I want to apply this to non-CDI, 12v system, then it will fire only after the dwell time, and so trying to perform this timing is more dificult. But we would like a solution that could be applied to either setup.

So a starting place to operate is at start and idle at the boring non-advanced F place. You can just transfer signal directly to the output to fire when this is sensed. This has the advantage that it should start quickly. If you are kick starting, this could be important. Even with start motor, I like it to start immediately. And then keep things at that boring non-advanced rock steady mark for a while.

At higher RPM you need to advance the timing or your engine will only obtain limited power. So for that, another rotor pickup signal could be used, or a mechanical advance like used on points.

The method used here is to calculate it. How accurate this will be the timing light can show. Does it jitter? Mine does a little at the moment, but it's running the engine, wahoo!

Now let's say you want to apply this to a 12v system. Can we ignore the dwell time(say 3ms for example) at start/idle? I don't think so, but you could arrange the rotor signal to compensate(move the plate).

So the basis here is to try and keep things simple with a single rotor signal. I will point out that using leading or trailing edge of signal is important. That using the trailing edge could be problematic to filter, while using the leading edge is easier. And that would apply if you wanted to use mechanical points triggering on trailing edge. Perhaps adjusting it to work from leading edge could be considered. Of course a non-mechanical switch(like a HAL sensor) could be desirable, but this entails adding this mechanically to replace the existing points.

There are digital solutions available for doing all this, they will cost you some money and time. Or you may just like the simplicity of points and
mechanical advance. Nothing wrong with that. But we can consider getting adventurous and revolutionizing 1970's ignition to the digital age.

In the beginning there was the cave man, then he created spark, then came the Honda Twin. Long live the Vintage Honda Twin.

 

[ancientdad]

In the beginning there was the cave man, then he created spark, then came the Honda Twin. Long live the Vintage Honda Twin.

This, the cave man that I am, can absolutely understand and agree with.

 

[kbongos]

I'm still stuck on my micro controller where it jitters around the calculated advance. Now I could use the max advance pulse signal to work around this, but I would prefer the simplicity of a single input signal.

As a side note and service to the community(and my curiosity) I am buying some cheap DC CDI's and will see how well those work with my CM400 setup, where I have toggle switch to use non-advanced at start and then switch to advanced pulse signal for running. One of them has a coil. Other is not potted, which will allow me to look at the circuit.
Will report back with findings. Here's what I got:

 

[Maraakate]

I'm still stuck on my micro controller where it jitters around the calculated advance. Now I could use the max advance pulse signal to work around this, but I would prefer the simplicity of a single input signal.

As a side note and service to the community(and my curiosity) I am buying some cheap DC CDI's and will see how well those work with my CM400 setup, where I have toggle switch to use non-advanced at start and then switch to advanced pulse signal for running. One of them has a coil. Other is not potted, which will allow me to look at the circuit.
Will report back with findings. Here's what I got:
View attachment 23423View attachment 23424

Bud, if you want the actual PCB with component list from the real CDI PM me and we can work out a VHT member and competent hacker deal. No need to waste time with these inferior toss away products. I can tell by looking at the board on the right it's the absolute cheapest components they can source.

Just FYI to anyone else, I'm discontinuing the kits except on a per-request basis. Causing too many problems. If I am confident that you're competent to solder and isolate and debug properly (like kbongos!) then I will sell you one.

 

[kbongos]

So I did get these cheap DC CDI ignition units in. I got them working on the bench. Made a nice trigger circuit circuit with some schmitt CD4093 logic gates I had. It's really easy to make a square wave just adding a resistor and cap with one of these, plus they work on a wide supply range(up to 20v). And you can add some diodes to make it something other than a square wave.

To my dismay they did pot the yellow one(a transparent potting material at least), but whatever, the components appear all unmarked anyhow. Other one came with a small coil and cable assembly. I compared it to a dumb COP coil used in for my ford escape. I got both of them to make spark, and they both appear about the same resistance for the primary. Measuring resistance of these primaries is a challenge, I ended up finding a high wattage 2 ohm resistor, putting them in series to a 12v car battery, and then measure the voltage between them. That's one way of doing it. Typical multi-meters cannot measure that low.

Now it was interesting in LDR's thread, he uses a DC-CDI Ignitech module along with a car coil from a 3800 engine. I've had some cars with this and have burnt out a few of these coils.
I think these coils are driven not by CDI(300v cap) but TCI(12v dwell). And I understand that there are some differences in CDI versus TCI coils, where CDI ones being higher voltage are lower primary resistances, less inductive. I can understand that the coils may be slightly different, working better with some characteristics tailored to the mode of operation, but my recent experience is that there must be a lot of overlap and they are really not so different. So the coil that came with this cheap DC-CDI measured roughly the same as the one aimed at the ford escape(TCI).

And I do believe that the 3800 engine coil packs use TCI(12v) and not CDI(300v). Now maybe the 'pontiac bonneville SSEi' had a special controller with CDI, but I doubt it, I'm thinking it was TCI.
Looks like this was a 1990's car. My Buick Lesabre with the 3800 was I think a 1999. I'm guessing these used the same controller, which was I think TCI(not CDI), but who knows. These 3800's were used for a few decades. I didn't have much luck googling on what the controllers used. ICM(ignition control module), well that doesn't tell you much. I did find some pics of it, it sat underneath the 3 coil packs, and found one blog and pics that made me think it was not CDI(no big caps). Now I have heard reference to CDI being used with cars, it's just tough to find info on with what. Seems like more of a specialty item with cars to me and that TCI is more common.

I did get these DC-CDI's to fire up my CM400T and idle, first with the honda original coil, then with the 2 escape COP coils. It might be worthwhile to make a better wiring harness for this and road test it just to make sure it could be a viable solution, but I am somewhat more interested in getting a TCI solution working. Like with the smart CNP units that Jay is interested in. These get interesting because of the dwell time needed to work into the timing, and hopefully the micro-controller can be arranged to do that.


A CDI solution is easier because you don't have to work in a long dwell before fire time.

 

[LongDistanceRider]

The coils I use are for the General Motors DIS ignition system which appears to be a TCI type

 

[kbongos]

I arranged my bike to use 2 Ford Escape 2003 COP coils, along with the discount AC-CDI. It runs well. I'm going to run it like this to prove it out. The output of the CDI is
wired to the 2 coils in parallel. The original with all the wiring is still there in case I get stranded. As I have mentioned I have a switch that selects non-advance to start, then I switch it to advanced for the duration of the run. Aim is to just verify that these inexpensive ignition components can work. The switch for advance is a bit of a hassle, but that's OK, it's simple.



The microcontroller project is still in the works. My current thinking is that I cannot use a calculated advance position from just a single rotor position, just not accurate enough after a near full rotation. My bike has 2 original pulse(coil) pickups, one at max-advance, and one at start/idle(15 degrees). So an easier target would be to use both of those signals, if you want some advance between them, a calculated position could be used. I may do this next, I could apply it to what I have with the CDI, taking the place of the toggle switch.

Here is a test circuit to test ignition using some CDxxxx logic gates.



Some possibilities after this will be to try and arrange a TCI with the existing pulse sensors. The 3ms dwell needed to work in prior to fire makes this tricky. At higher RPM 3ms is a lot of time. So my advanced will not be so advanced if I want to trigger off that existing sensor. I believe my CM400 rotor sensors are fixed and cannot be altered easily. So next step would be to add and use alternative trigger sensors. These are ultimately needed if you wanted to upgrade a points system, as points are crude.

Adding a disk with trigger points is common. You could use 1 or more sensors. A single sensor can measure two positions, leading and trailing edge. Modern systems use missing tooth gear to measure many positions along the rotor. The pulse sensor is a coil and a magnet, they also call these reluctor sensors. The simple ones(like my CM400 has) will be 2 wire and just the coil directly. Alternatives to this include Hall sensors(these will be 3 wire, as some power required). The pros to reluctor are the simplicity of a coil, not much to go wrong, high temps not a issue. Cons to reluctor are they require a slimmer more controlled spacing, a little more complex to process. The Hall sensors can experience problems with high temps. Optical interruptor sensors also are a possible choice.

I'll probably use the Hall at first because they are small and inexpensive, like a transistor. Room to arrange these is confined. I'd probably want to test how high temperatures can damage these.Maybe these reluctor coil sensors are needed. The Harley I fixed had a flakey HAL sensor. Maybe missing tooth gears is the way to go.

A hall(or reluctor) can pickup on a passing magnet, say fixed to a disk. Alternatively the magnet can be built into the sensor and then detect a passing iron disk lobe(like a missing tooth gear/disk). Similar to a electric guitar pickup detecting the metal string. Common automotive use includes rotor sensors and the ABS brake sensor.

 

[Jays100]

My understanding of how the Chevy smart coil system works is the first is a trigger from the missing tooth wheel and all other cylinders are calculated from there.

I would think (and being totally ignorant of the CM series system) that the initial trigger could be at full advance then a lookup table of delay v rpm be utilized to initiate the spark at the proper time. this is valid up to full advance, 3500rpm?, where it triggers direct.

 

[LongDistanceRider]

My understanding of how the Chevy smart coil system works is the first is a trigger from the missing tooth wheel and all other cylinders are calculated from there.

I would think (and being totally ignorant of the CM series system) that the initial trigger could be at full advance then a lookup table of delay v rpm be utilized to initiate the spark at the proper time. this is valid up to full advance, 3500rpm?, where it triggers direct.

SOHC CB/CM 400/450 engines get full advance of 43*, +/-2*, at 4500 to 5350 rpm per the FSM. Nice tight exacting spec's.:lol:

 

[Jays100]

Good thing the engine isn't fussy, that's a pretty big spread for timing.