Idle control on specific conditions

[AVP]

Hello all,

im struggling a bit with the following problem:
Although idle seems to work nicely on most occasions, even on high electical load it does not dip that much, i have this problem:

as im cruising, if i step on the clutch and go to neutral while the car is still moving, and on top of that if i press on the brake slightly, the rpm needle as it drops towards 1000rpm, it dips, as if it will stall, then jumps up and settles to my idle speed.

i have tried various idle configurations to avoid this from happening, all unsuccesfull

I presume that this happens because as the car goes to idle config once the gear is out and TPS=0, there is not enough air allowed through the idle valve to allow for the idle settle properly without dipping.

I tried using min PW allowed up to 30-35% and above that it makes my idle really high so i dont like it.
This somehow helped a bit, but not much.

I tried playing around with PID and deadbands, again without any real change.

I tried altering the PW in hardware settings, again without real luck.

can you help me with this one? which area of the idle settings should i focus on?


what does the IAC delay do as well?

thanks
vasilis

[gunni]

what happens with the lambda while it goes down?

you need to be analytical and see exactly what the reason is for the behaviour,

not enough fuel/air , to much? to late?

[AVP]

ego works until 400rpm. the idle bin there is 1.01

when this happens, scenario is somewhat like this: i accelerate in gear and then immediatelly go to neutral and TPS=0

lamda on that point would be on the equivalent bin to accomodate and once in neutral, would go to the 46kpa bin which is =1.01

also the same happens if im on overrun and go to neutral in which case im lamda = lean and then goes to 1.01

i tried today PID settings with D=0 or 1/2 and it seems to be better. i do struggle though to have the idle lower than 1000-1050rpm though.

im wondering if i have to readjust the TPS a bit for that since this way i get no dips in my specific circomstances.

ill keep testing it for now and get back to you.

i think it was lack of air that causes this or PID control

[AVP]

i have tried for a couple of days the idle with PID: 28/12/2 and it seems much better.

however idle still has a specific issue with VEMS.

the higher the IAT, the higher the idle.

now this is something that needs additional adjustment with a table perhaps that has IAC and MAT as  variables. Its the only way to keep it from climbing up as this happens.

[mattias]

the higher the IAT, the higher the idle.

Lower your value for "refpos addition for MAT, slope".

[AVP]

its 3.1 now
should i get it even lower?

[AVP]

i reduced it to 1.2 4 step for now. ill recheck

[mattias]

You should also consider that you are probably not letting the PID control to have enough freedom to move away from the reference curve. Check the integral values in the datalog, to see if they rail against the min/max values you have set.

[AVP]

my integral values were set to 20 min/max and i reduced the max bit to allow the idle to settle lower if needed. i have always used small numbers here, maybe under this gonfiguration i should use higher.

PID, i tested lots of ways, starting with P only then working on but as soon as i introduce D into the equation, the problem reappears as i described it above. Also if i use high values on P and I, but relative to eachother, i find that idle may jump around violently on startup, so using values of 28/12/2 seems to work fine at the moment.

should i make the integral decrease limit even higher than 20?

[mattias]

I would use a larger integral limit. I generally use 0 for increase,  and 50-80 for min, that's because I define my reference curve to be 200-300 rpm above the target rpm. The PID control then beds down the engine at the target rpm after first settling at a higher rpm. This behaviour has somewhat been removed and it is clear we need some re-factoring of how it works, it is discussed here:
http://www.vems.hu/wiki/index.php?page=MembersPage%2FDamirMuha%2FIACintegral

Remember to take into account the effects of having ignition based idle control active.. it can hide  problems with the PID tuning of the idle air valve.

[AVP]

yes im always concerned regarding the ignition based idle. so far whether i use ignition and iac getting after 2 or up to 9deg after the ignition to control idle makes no difference, so i have left it to 4deg after.

im going to put the integral limits in now.

[AVP]

right

on warm engine and MAT of 33-38 as the car was stationary i had 1030rpm which is ok, BUT as soon as the integral dec limit went above 20 (tried even 50) the idle speed went up, close to 1150rpm. best deal was with integral dec limit of 10.  Maybe it has to do with the fact that my PID has low values as well: 28/12/2

the mat slope is on the second choice now, 0.8%. this way the idle is very good under high MAT conditions. again, changing the deg of advance that the IAC intervenes from 2/4/8 deg didnt have much effect on idle speed. It may have on changes of idle speed and the way it is controlled though.

[mattias]

What is your target rpm with a warm engine ?

[AVP]

940rpm

If i have ignition timing OFF, and PID 0/0/0, idle jumps to 1200-1300rpm. It is PID that slowly brings it down to the target rpm.

[mattias]

Please share a log file, lambda fully warmed up, no ignition based control and whatever PID settings you want and 50/50 for the integral dec/inc limit.
In the log, let it idle stationary, and then pump the throttle to bring it up to 3-4k and let it settle down to idle.