Number of teeth / cylinders conflict

[racingmini_mtl]

Generating virtual teeth is likely to give you less accuracy than simply using the lower number of real teeth. Either will work fine at constant speed but as soon as you have fast transients, your generated teeth are out of sync. Unless you take into account the acceleration and deceleration of the crank wheel which is another level of complexity which can also lead to some accuracy issues.

Jean

[jimmy76]

Thanks for that advice!
So I should better use 12-1?
Wouldn't it be the same from accuracy point of view?
The 2 generated teeth would be calculated of the previous tooth 'length'. So there is only a 2.7 degree latency in case of acceleration deceleration.
I'd expect it is still better than using 12-1.

[racingmini_mtl]

It won't be better because you're using false data for the extra teeth when the wheel is not at a steady state. And the MS2/Extra code does use all the teeth to determine speed and acceleration so false extra teeth will screw up with this while fewer real teeth won't.

Moreover, timing uses the same teeth and the virtual teeth are not located to anything real but to a time that may or may not be correct. Real teeth always correspond to a real engine position.

Jean

[rickb794]

A compromise between more teeth for accuracy and less teeth for full rpm span is a 36 tooth wheel add a cam sensor or a single tooth in a distributor to provide phase detection and you can rock all the features MS can provide.

[jimmy76]

I've tried to make a worst case calculations about how 'wrong' that virtual tooth could be.
As a hypothesis I've calculated with an engine which could speed up from 1000 to 7000 under 2s. That is 50 turn/ s2 acceleration.
At 1000rpm 1 tooth to pass by takes about 500us. (I have 132 teeth)
So if I'm correct than from speed point of view the real rpm would be 1001,5rpm on the next tooth.
(As I'm calculating the position of the virtual tooth based on the time length of the previous one)
From phase point of view it seems to be 0,0015 thooth.
I can even think about measuring not only one but 2 teeth and compensate with some acceleration / deceleration.
Please don't get me wrong. I'm just curious what I can do with an additional micro.
Thanks for all the advices!

[racingmini_mtl]

... the virtual teeth are not located to anything real but to a time that may or may not be correct. Real teeth always correspond to a real engine position.

Regardless of what computations you do, the above statement remains true. Anything you do in your code to time virtual teeth is very unlikely to improve or work with what is done in the MS2 code to get a precise timing. But you can do it if it amuses you. I'm done here.

Jean

[jimmy76]

If there is no advantage compare to a 'real' 12-1 than I won't waste my time.
Thanks

[DaveEFI]

B&G code does allow the use of PWM control of low-Z injectors. I ran with that for years. Unless your number of teeth issue disables that function - I used EDIS. But it would be silly to start a new project based on the old and restrictive B&G code, unless you are certain it will do all you want.

A reliable tach signal is essential to MS. Surely it can't be that difficult to fit a 36-1 trigger wheel to your crank pulley? Willing to bet others have done just this.

[LAV1000]

@Jimmy76,

Do you know this one:
http://www.ti.com/lit/ds/symlink/cd4017b.pdf

[jimmy76]

Thank you guys!

If I use the old B&G code I have 2 possibilities:
- let MS read both my VR sensors (this requires the microsquirt mod) but than I lose the 2 injector PWM outputs
- or I use the 2 PWMs but in that case I can only read 1 crank signal

I'd like to use this 2nd solution. But in that case I not only need a divider but I also have to combine my 2 VRs into a single missing tooth output. That's what my PIC does right now.
With the B&G it seems that I can use even a 66-1 (which is only a division by 2)
If I would like to use the extra code the only solution could be the 12-1 (divide by 11).

Yes. I know an additional 36-1 wheel with a single hall sensor would be the easiest. But my project right now is to make a plug&play MS2.
I do not plan to raise the boost, no need for a start program, I just want to get the 'modern engine' feeling with a better computer.

[jsmcortina]

With the B&G it seems that I can use even a 66-1 (which is only a division by 2)
If I would like to use the extra code the only solution could be the 12-1 (divide by 11).

66-1 should work on both firmwares. EDIT: See next posts.

James

[racingmini_mtl]

With the B&G it seems that I can use even a 66-1 (which is only a division by 2)
If I would like to use the extra code the only solution could be the 12-1 (divide by 11).

66-1 should work on both firmwares.

James

Are you saying that 5.45 degrees per tooth is not an issue with MS2/Extra?

Jean

[jsmcortina]

Are you saying that 5.45 degrees per tooth is not an issue with MS2/Extra?
Jean

Oops, you are correct.

James