1.4.0 idle issues

[elaw]

I'm running a self-customized version of the 1.4.0 code (but the *only* change I made in the idle area was to batt V correction) and am having a number of vexing issues. I sort of had things under control during the summer, but for whatever reason as the weather's gotten colder the issues have become worse.

I think there are several interacting problems going on but for now I'll stick with the most confusing one.

The attached log file is a snippet taken after a warm restart... so there should be no warmup-related issues.

If you look after the 1530-second point and again at the 1550-second point, PID idle engages, and the PWM idle duty starts sharply *increasing* even though the RPM is above target and the target is ramping downward. I think the only reason RPM ends up going down instead of up is that the idle valve is fully open (or near so) and thus having little effect, and the idle timing correction is what brings the RPM down. At some point the PWM idle duty reverses its upward trend and later the idle begins oscillating and due to RPMdot it eventually kicks out of PID idle but that's obviously a tuning issue. But I can't see how the PWM idle duty going the wrong direction could be a tuning issue?

.msq attached, along with an .ini file that can be used to view it.

Thanks!

[elaw]

Here's my modified idle file if it helps. Just look for comments with "ELaw" to see what I changed... basically I changed the PWM idle V compensation from being an adder to a multiplier, and the PWM idle value that's sent to outpc is the uncompensated value, not the compensated one.