Ignition with a 3-tooth trigger wheel?

[SnailRacer]

Hi all-

We're running a Yamaha YZF600r, which in stock configuration has two coils and runs in a wasted spark ignition mode. The stock trigger wheel is pecular, though. It has only three teeth: two large 60 degree teeth and a small 8 degree tooth. The large teeth are each spaced 120 degrees apart (think symmetry), while the small tooth preceeds one of the larger teeth by about 15 degrees. The end of the large tooth that follows the small tooth aligns with the pickup at TDC.

In theory, this 3-tooth system is enough to calculate RPM and tell which coil to fire (remember, two coils + wasted spark). However all of the other engine management systems I've seen or heard of use ALOT more teeth, as I'm sure you're all well aware. Seems to me that with a 3-tooth trigger you'd loose too much precision to implement any (accurate) spark timing algorithm.

But Yamaha obviously got it to work. We'd like to make it work too, as the stock trigger wheel is also part of the charging and balancing systems, and there doesn't seem to be any convinent place to mount a custom trigger wheel. Our ideal solution is running a Megasquirt with modified code and an extra ignition trigger output for the second coil.

So anyone know how to structure a successful ignition timing algorithm around a three tooth trigger? Or am I making a bigger deal of this than it needs to be?

Thanks!!
Jeremy

[SnailRacer]

Hi all-

We're running a Yamaha YZF600r, which in stock configuration has two coils and runs in a wasted spark ignition mode. The stock trigger wheel is pecular, though. It has only three teeth: two large 60 degree teeth and a small 8 degree tooth. The large teeth are each spaced 120 degrees apart (think symmetry), while the small tooth preceeds one of the larger teeth by about 15 degrees. The end of the large tooth that follows the small tooth aligns with the pickup at TDC.

In theory, this 3-tooth system is enough to calculate RPM and tell which coil to fire (remember, two coils + wasted spark). However all of the other engine management systems I've seen or heard of use ALOT more teeth, as I'm sure you're all well aware. Seems to me that with a 3-tooth trigger you'd loose too much precision to implement any (accurate) spark timing algorithm.

But Yamaha obviously got it to work. We'd like to make it work too, as the stock trigger wheel is also part of the charging and balancing systems, and there doesn't seem to be any convinent place to mount a custom trigger wheel. Our ideal solution is running a Megasquirt with modified code and an extra ignition trigger output for the second coil.

So anyone know how to structure a successful ignition timing algorithm around a three tooth trigger? Or am I making a bigger deal of this than it needs to be?

Thanks!!
Jeremy

[Steffen Hausler]

We had a similar problem with our WR450 single cylinder. Yamaha put a +1 wheel surrounding the magnetic core for the AC magneto. The problem we had is that our fuel injection system only takes a 12 -1 signal. How we solved it was to machine a ring that will match this 12 -1 signal and we put it over the existing magnetic core, seeing we could not get rid of it or find any other place to put the ring.

And as to using the existing signal, unless you have some EEs on your team with really good skills in signal processing, the out look is about the same on what we decided we had to do.
Now some EFI systems can take alternative signals, not necessarily the normal 60-2, 32-2, 12-1, etc.

Hopefully this will help you. Feel free to ask me more information if you want.

Steffen Hausler
Cal Poly San Luis Obispo

[Jersey Tom]

3 teeth? Interesting.

As I understand, in a steady state scenario (ie engine held at some RPM) having a small number of teeth isn't a problem. Its rapid acceleration of the engine where having more teeth greatly increases accuracy.

What EFI system are you running exactly? Stock, and just modifying your advance and fuel maps? Or are you running a standalone system?

If the latter, go with whatever is best supporteed by their system out of the box. For example, the stock Honda F4i system uses a 12 tooth wheel, with a kind of odd duty cycle. Far from 50%. When we first migrated to our new standalone system, which takes a 12-1, we just ground off one tooth and worked with that. Gave really crappy response. I then machined up a new 50% cycle 12-1 wheel, put it on that spline, and it was a very noticeable benefit.

Not sure on your engine how your wheel is on there... on the F4i its just a spline interface with basically a bolt holding it on. If thats all it is, its real easy to machine up a couple different wheels, throw em on, and see if they make a difference.

3 tooth though, weird.

[B Lewis @ PE Engine Management]

If you actually go through the math just using the laws of motion, it will become evident to what extent the number of teeth effects ignition timing accuracy. Having gone through this several times while designing our systems, I will try to explain the major points:

1) The largest amount of error in the crank position calculation occurs at low RPM. This is good because most engines don't reach peak torque until later in the operating range.

2) Even if you assume a ridiculous amount of acceleration at the crank, almost all of the error can be accounted for with a relatively low number of teeth. For example, if we assume an angular acceleration of the crank to be 10,000 RPM/sec (which is insanely quick), at 5000 RPM the max error will be less than 0.25 deg with a tooth spacing of 60 deg and not accounting for any crank acceleration in the calculation.

3) How the crank position is calculated in the ECU is as important as the number of teeth. For example, a three-tooth wheel (like the one mentioned earlier) can be extremely accurate if implemented correctly. With three teeth (one offset), the computer can determine TDC of cyl 1/4 and cyl 2/3. Assume the maximum timing needs to be 50 DBTDC for the engine. If you set one of the teeth on the wheel to indicate 50 DBTDC, you only have to account for acceleration of the crank over 50 degrees. Using this method, you can easily get accuracy of greater than 0.5 degrees without even estimating the crank acceleration.

Well, it got a little long, but I hope this helps.

[Jersey Tom]

Well I feel educated now. Sweet.

[SnailRacer]

Thanks everyone!

Steffen - I like your team's ring idea, we'll probably give that a try -- it opens up the opportunity to use whatever engine management we want.

Brian - Your insight is quite enlightening! Makes this project seem a little more feasible...