On my wiki page
http://www.vems.hu/wiki/index.php?page=MembersPage%2FFPhil%2FInjectionI got the following answer
"- I have not understood what the option "Asymmetric PID conf" does.
.. if enabled, bigger (twice) values to feed more air (push RPM up) than to remove air."
Last week I started to use the I term of the idle control, but I have not understood how the I action works when the increase and decrease limits are set opened. I may report on my page.
As said elsewhere better to open only the decrease limit and to choose the IAC DC reference for an idle rpm well above the rpm idle target. In this case the integrator term starts to add up (for decreasing IACDC) when the rpm first crosses from under to above the rpm idle target.
If the rpm happens to become lower than the target then the integral term decreases (for increasing IACDC). I did not notice in this case (rpm<target) the integration is done twice faster when the "Asymmetric PID control" option is enabled.
Beware each limitation (on 1 byte) is done without the integration constant which is multiplied afterwards, which means that the actual limit for the action on the IACDC value depends on this integration factor.
Sometimes I feel it could be easier to have the fw code to read.