Hall effect, single-cylinder synching problem

[Pekellop]

Hello,

I have a problem with the speed detection from a Hall effect sensor, more specifically with respect to cyclic variation in speed. I have a single-cylinder engine with a cam Hall effect sensor, picking up a 24-1 toothed wheel. The controller is a Microsquirt.

The engine is mounted on a dyno, with a flexible coupling and one-way bearings, to avoid problems related to inertia (the inertia of the dyno being quite a bit more than that of the 3.5 hp engine...) and stress in the crankshaft. With a previous setup, without one-way bearings, speed detection and timing was working properly.

Now, the Microsquirt is losing sync, at cranking and running. I removed the spark plug to eliminate engine compression and motored the engine from the dyno, and everything is working fine! This leads me to believe that variations from cycle to cycle, or within a cycle, due to compression, are causing the controller to lose sync.

I have tried increasing the next pulse tolerance (for cranking, up to 90% and running up to 70%). This seems to resolve the synching problem, but the rpm values are much higher than in reality. Is there a way to solve my problem, through some sort of parameter in the software? I would hate to have to increase the engine’s inertia with a bigger flywheel.

Thank you

[Matt Cramer]

Could you post a copy of the MSQ file? I'd like to have a look at the settings to see why the RPM is too high.

[msiddalingaiah]

Have you completely ruled out interference from the ignition system? When you were motoring with the plug removed, was the ignition on? Maybe you could remove the plug, connect it to the coil and ground the plug case. If you see spark, then it's almost the same as running the engine.

Just a thought...

[Vincent]

Here's the MSQ file.

infinity_2010_06_20.msq

Effectively, the ignition was OFF during the test without plug. So it could be an ignition interference problem. I will change the setup as proposed and retest, then post the result. I will also try to change sensor gap and orientation.

Thanks for your help

[Matt Cramer]

Hello,

I have a problem with the speed detection from a Hall effect sensor, more specifically with respect to cyclic variation in speed. I have a single-cylinder engine with a cam Hall effect sensor, picking up a 24-1 toothed wheel.

Then why is your MSQ file configured for a 30-2 wheel?

[Vincent]

Sorry guys, I attached the wrong file.

megasquirt201009080848.msq

As I said, with this MSQ we have good RPM reading when spark plug is removed, so I think the 24-1 settings are OK... I tried with ignition ON and OFF and it makes no difference on RPM fluctuations. Any other idea of settings that can be ajusted (or hardware modifications) that can help cranking sync problems?

[Matt Cramer]

Try no dual spark, 24 skip teeth. You may also need to turn off alpha-beta-gamma filtering, or adjust the values.

[Vincent]

No dual spark and 12 skip teeth solved the problem. Thanks for your help!