For some reason with anything other than 0 in the "Integral" my IAC valve setting slowly drifts downward with subsequent worsening of the idle. It works great for the first 10minutes or so after start-up before this happens. Restarting resets the IAC position and I'm good for another 10 or so.

With only P and D values set it works fine, no drift at all. It's actually perfectly acceptable with the I set to zero but was curious if anybody else has experienced this or know why?

 

You have to really play with those settings to get it to work exactly right with integral. My settings are actually 2 proportion 7 integral 2 differential. That produces the best idle I've experienced. Other users have their variations too some which don't involve integral and some that do. So i really think it kind of depends on your other settings.

http://www.boundaryengineering.com/instructions.zip

Check out these files for some settings suggestions.

 

I'm running also with 272. It seems to be the best setting for my setup so far.

 

I think I figured out what was going on. I was going into closed loop idle outside of the control band because my neutral timeout was set very low and on occasion I make very slow shifts when I'm putting around town. The idle flag was on anytime I held the clutch down too long when shifting which is probably driving the iac valve closed to get the idle in the control band. I guess the "Integral" control is what does this?

"The ECU will only actively control idle speed under the following conditions:

1. The throttle is closed;
2. The clutch/neutral input is active, or the "Throttle cracker" is enabled and the vehicle is stationary;
3. The actual RPM is less than the target RPM plus the "Control Band" (note that there are two control bands; one for normal operation and the other for when the air conditioner output is on), or the other conditions have been met for the period of the "Neutral Timeout".
Once these above conditions are met, the closed loop idle speed will be controlled in a closed loop mode by the ECU."

 

I know exactly what was going on then. The integral factor adds correction over time. So you were adding some great big numbers to the integral error at 4000 RPM and 0 percent duty cycle. This is called integral windup in PID theory. So when you were inbetween shifts the integral value just kept getting bigger and bigger because regardless of what the PID controller did the RPM wasn't getting better, and there was one HUGE error between the 4000 RPM you were shifting at and the 1000 RPM it wanted to idle at. Finally you got the value so big it just overwhelmed the system and cut the IAC valve.

 

Thanks for clearing that up. Idles great now.