Closed-Loop Idle: Car Dies Coming to a Stop

[blasphemy101]

This post was originally much longer, but FireFox shit the bed...

I know I've seen this problem come up before, but I can't find the topic through search.

Bottom line, I've tried setting up CL Idle using these settings:



The idle is fine and solid (crank-to-run taper is a little long; I set it to 20). My problem occurs when returning to idle from 1k - 1.8k RPM (whether while driving or free-revving); the RPM drops well below the idle speed, often stalling the engine.

If the car is driven/revved anywhere over 2k, it returns to idle fine...

Any thoughts?

[nando]

settings look OK. I'd consider using the shift settings for keeping the valve closed, but nothing stands out as being wrong.

what does your target map look like? can you post a screenshot of a datalog where it stalls, and one where it catches?

[whodwho]

A couple quick observations

Your PID gains are not optimized well and not for quick response
You want to zero out all the gains
Check that you are in CL
Bring in the I gain just enough to hit your target RPM
Increase the P gain to increase response to load changes(you want this as high as you can get away without oscillations)
Leave the D gain at zero (you shouldn't need to use this unless you are getting uncontrollable oscillations)

I would try a higher Min Duty for PID to a bit under what is needed to maintain the target idle

I would also try a little higher RPM with valve Open

What is the crank duty for warmed up set at compared to what is needed to maintain an idle?

Are you using idle advance?

[blasphemy101]

settings look OK. I'd consider using the shift settings for keeping the valve closed, but nothing stands out as being wrong.

what does your target map look like? can you post a screenshot of a datalog where it stalls, and one where it catches?

I think they should look okay. I think they came from you somewhere down the line (I copied them from another thread I found).

A couple quick observations

Your PID gains are not optimized well and not for quick response
You want to zero out all the gains
Check that you are in CL
Bring in the I gain just enough to hit your target RPM
Increase the P gain to increase response to load changes(you want this as high as you can get away without oscillations)
Leave the D gain at zero (you shouldn't need to use this unless you are getting uncontrollable oscillations)

I would try a higher Min Duty for PID to a bit under what is needed to maintain the target idle

I would also try a little higher RPM with valve Open

What is the crank duty for warmed up set at compared to what is needed to maintain an idle?

Are you using idle advance?

That's how I originally tried tuning the PID settings, but I ended up trying these settings. Do you think they could be slowing the loop's response enough to let the car stall? I can try tuning them once I get some of these other settings resolved.

I'll have to check on the crank duty when I get home, but I believe idle advance is set at 18deg. Why would the crank duty % have an effect here?

[nando]

Crank duty can come into play when the RPM level drops below your cranking RPM threshold (typically 300-350).

BTW the cranking duty on BMWs is pretty much always set at 50% DC.

[whodwho]

It seems to use the crank setting for the last known position to start and I have a feeling by watching and having changed it around that it may use it more often, worth checking/trying.

The min duty for PID is another one that I would look into

[nando]

Does MS2/e still use "last known good duty" or did they switch over to the CLT/RPM duty table like MS3?

[blasphemy101]

It seems to use the crank setting for the last known position to start and I have a feeling by watching and having changed it around that it may use it more often, worth checking/trying.

The min duty for PID is another one that I would look into

Gotcha, I'll check it when I get home. I'll look into the min duty for PID. What kind of setting should I be looking at?

I know you said just under the duty % required to maintain the target idle. If I'm understanding you right, then it should be a about 15-20% (ie 25-30% total). Does that sound like a more typical value?

Crank duty can come into play when the RPM level drops below your cranking RPM threshold (typically 300-350).

BTW the cranking duty on BMWs is pretty much always set at 50% DC.

I believe my cranking RPM threshold is 350 RPM. Again, I'll check it later tonight.

Does MS2/e still use "last known good duty" or did they switch over to the CLT/RPM duty table like MS3?

I believe it uses the last known good duty logic from reading the documentation.

... I'd consider using the shift settings for keeping the valve closed...

I'm a little fuzzy on the purpose of these settings. They're for closing the idle valve during shifts..? If so, why do you want that?


By the way, thanks so much for all the help guys! I swear I've tried searching, but a lot of this info doesn't seem to be readily available. I'm really trying to learn this stuff to get out of the "I know how to set my fuel/ignition maps" stage... I can tell MS is a powerful tool, and it's frustrating to know I can hardly see the tip of the iceberg.

[whodwho]

Does MS2/e still use "last known good duty" or did they switch over to the CLT/RPM duty table like MS3?

it appears it still plays in MS3 as well, from - http://msextra.com/doc/ms3/Idle_Control.html

Dashpot Adder (percent or steps) - This setting sets the amount that the valve is opened above the previous known-good valve position calculated by the PID algorithm when PID control was last engaged. The dashpot adder is added to the last known-good position each time throttle is lifted.

[Bearmw]

This thread is gold.

[blasphemy101]

Just wanted to post here and point out I haven't forgotten about this...

It looks like my TPS is bad. It won't read throttle at light throttle inputs (reads 0-2%), so I think the car is staying in CL at light throttle. I assume this would throw off the PID loop.

I'm going to pick up a new TPS and report back.

[whodwho]

Are you using a M50 style TPS? Did you switch pin 1&2 at the TPS connector?

[blasphemy101]

Yes M50 TPS...Do what now? I've been using MAPdot for everything, so I've never even used the TPS, just installed it for the future if I wanted stuff like CL...

[whodwho]

When adding a variable TPS you need to switch pins 1&2 at the TPS connector to get the proper signal, it will give you an unstable signal if not

[blasphemy101]

When adding a variable TPS you need to switch pins 1&2 at the TPS connector to get the proper signal, it will give you an unstable signal if not

I haven't noticed any instability until now... It seems to detect higher throttle inputs, but like I said, I've never paid much attention to it.