Can some one explain for me why we are getting kickbacks during starts?

Here is how I see it.

If you have your spark at TDC two things can happen.

* It will ignite (the mixture is sort of correct), the engine will continue to rotate in the same direction as the starter is trying to rotate it and the engine will hopfully start.

* It will not ignite at all (mixture is totaly wrong).

Am I missing something? For me the kickbacks are a mystery. If I take my oem P16 ECU. Install a 996 or a 748 EPROM the bike will start without kickbacks. Even with a EPROM designed for a 748cc bike in a 996 with bigger cams.

Why does MyECU give us kickbacks.

Can some one explain it so I understand. It can not be because is spontinius ignition as the cylinder pressure is not very high during cranking and if the engine has not been running it is also very cold.

Jocke.....

If you logged some data as was suggested, we might be able to tell you.
As it is you post about problems, disappear, then come back and post about the same problem.

As it is I have no idea what you did or where you're at.

There are many possible reasons for this. Please post your map and data logged and hopefully we'll find out which one.

In case you fear Hyperterminal for whatever reason, try using my Logworks plugin for logging. You just tell it what serial port to use and it will ask for a log file name. The file written will be exactly what you would get from Hyperterminal. You don't have to start Logworks though I guess it need to be installed.

I think it's an interesting discussion. From what I understand, in theory it just has to be bad timed firing. But why/when is it firing? Cliff's post is good and it illustrates a very plausible scenario.

In practice though, I know the "usual" kickbacks I had until last spring did not root from ECU resetting. I usually have the Optimiser mounted and I've only seen two ECU brown outs while cranking. One of them ended up in disaster, but that's another story.

Just like greenmonster, my experience is that too much fuel will likely result in kickbacks. I have no idea why - maybe these are not actually back fires per se, the engine seem to be forced to a halt for some other reason.

My kickbacks just disappeared for no obvious reason. I had done a number of things that may be part of the solution: better ground strap from battery, DIY starter overhaul, backported the voltage boost components, replaced regulator (it did not charge at all below 3000 rpm), set air bleed and butterfly opening to WHB spec and adjusted idle speed by leaning mixture in the map instead. All of a sudden it just started to behave like one wants it to: it starts quicker than you can get your thumb off the button. I would love to know why but, of course, I'm happy anyway...

I may be wrong but my gut feeling is that the single most important change was when I stopped trying to set idle using the idle screw and/or bleeding screws - but leaned the idle cell instead. Incidently, I think that is what Cliff has recommended all the time.

greenmonster wrote on 11/16/09 at 00:39:11:


Not correct. I have be converting my 851 to P16 by installing a 916 / P16 harness and made it to fit so I never touched the started stuff.

Good thougt but did not work  this time.

I am not too concerned about the kickbacks. I am sure that it can be cured. I am more interested in why it happens. I think that Cliffs reply could be something to look at.

Jocke.......

It's a pity you can't actually log the first crank revolutions. I don't even know at what RPM the starter manage to crank the engine before a successful ignition. My logs usually jump from 0 rpm to about 900 rpm with a time warp and a diagnostic message in between. I'm not saying it's important enough that it should be "fixed" though. Unless that would be fairly easy?

greenmonster wrote on 11/16/09 at 00:39:11:

I have an older version of MyECU hardware (Mk II) which did not feature the voltage boost. Cliff helped me with the details for incorporating the few components that make it up.

Cliff, apart from resetting (or upsetting) the CPU, could a borderline voltage have more subtle effects, like not fully driving the FET gates? Quote from your schematics: "The FET requires about 10V to turn on fully so an extra stage of buffering with the BC549 is used to raise the microcontrollers 5V to the 12V of the bikes electrics." So what happens if the gate only gets, like, 9.5 volts? Is it a little worse or much worse? Then again, this shouldn't be a problem on a Mk III.

greenmonster wrote on 11/18/09 at 01:27:09:

Yeah. We can see that the OEM P8 fires at TDC while cranking. It's weird that only one of his logs show a voltage drop during cranking.

The 4 degrees can be explained, I think, by the way the input pulse is tracked. TDC is when the waveform crosses 0. I use a small positive offset. I'll have a look at this soon to see how it can be improved.

Instability ? Not sure about that one. I'll recheck that on my Cali. Do you have any waveform captures that show this? I could produce special firmware that can log some realtime data so that we can see what is going on.