Idle drop and stall after change from open to closed loop idle
#1
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
I'm looking for some help please sorting this little gremlin out. Every now and then, on average once per drive this stall will happen. 2001 nb8b ms3 basic.
It does it when driving and I have no throttle such as rolling to a stop and clutch gets depressed, the car with just stall, once started again its fine for usually the rest of the drive. It will also do it occasionally if I am idling and i blip the throttle, it will drop to 500rpm and stall. Hot or cold it doesn't seem to matter.
I have data logged the event, nothing jumps out to me, will try and post the file if anyone wants to see it.
It only started after I went from open loop idle to closed loop. Car starts well, runs well just this gremlin, it feels mechanical in nature but only started after switching to closed loop...
Thanks for any insight or pointing out the obvious.
It does it when driving and I have no throttle such as rolling to a stop and clutch gets depressed, the car with just stall, once started again its fine for usually the rest of the drive. It will also do it occasionally if I am idling and i blip the throttle, it will drop to 500rpm and stall. Hot or cold it doesn't seem to matter.
I have data logged the event, nothing jumps out to me, will try and post the file if anyone wants to see it.
It only started after I went from open loop idle to closed loop. Car starts well, runs well just this gremlin, it feels mechanical in nature but only started after switching to closed loop...
Thanks for any insight or pointing out the obvious.
#3
Elite Member
iTrader: (1)
Join Date: Jun 2006
Location: Warrington/Birmingham
Posts: 2,642
Total Cats: 42
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
Could be a gremlin.
I've posted many times on MSExtra where the dashpot isn't applied. The settings for when dashpot is/isn't added aren't configurable, and the developers have stated that it's not applied in certain circumstances, e.g. small throttle blips. Unfortunately IME these can be replicated when driving and cause a really low idle until PID recovers, or if severe enough a straight stall.
you can fudge it out by using the table rather than last known, but it's not perfect and that method is flawed when external conditions change.
I've posted many times on MSExtra where the dashpot isn't applied. The settings for when dashpot is/isn't added aren't configurable, and the developers have stated that it's not applied in certain circumstances, e.g. small throttle blips. Unfortunately IME these can be replicated when driving and cause a really low idle until PID recovers, or if severe enough a straight stall.
you can fudge it out by using the table rather than last known, but it's not perfect and that method is flawed when external conditions change.
#4
Boost Czar
![](/mt2011/images_td/ranks/supporter1.gif)
![](/mt2011/images_td/ranks/admin.png)
iTrader: (62)
Join Date: May 2005
Location: Chantilly, VA
Posts: 79,688
Total Cats: 4,113
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
Before crying wolf: Actually review logs, see what's happening or not happening when you stall and correct the problem by tuning.
dashpot is maybe what 2-4% idle bump? if you're stalling because you don't have dashpot adding, then you have other issues.
you have to make sure you hit the max decel value so CL idle will kick in even when coasting in N. It's imperative you tune the value correctly to prevent stalling in this very scenario (high gear/no gear and low rpm), it should be set JUST lower than what you typically idle at, and then you want to tune it correctly in your a/c activation settings.
It could also even be overrun settings not adding fuel back in time for the motor to recover if you're coasting in gear.
I can think of a few more things, but without looking at logs...
dashpot is maybe what 2-4% idle bump? if you're stalling because you don't have dashpot adding, then you have other issues.
you have to make sure you hit the max decel value so CL idle will kick in even when coasting in N. It's imperative you tune the value correctly to prevent stalling in this very scenario (high gear/no gear and low rpm), it should be set JUST lower than what you typically idle at, and then you want to tune it correctly in your a/c activation settings.
It could also even be overrun settings not adding fuel back in time for the motor to recover if you're coasting in gear.
I can think of a few more things, but without looking at logs...
#5
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
Before crying wolf: Actually review logs, see what's happening or not happening when you stall and correct the problem by tuning.
dashpot is maybe what 2-4% idle bump? if you're stalling because you don't have dashpot adding, then you have other issues.
you have to make sure you hit the max decel value so CL idle will kick in even when coasting in N. It's imperative you tune the value correctly to prevent stalling in this very scenario (high gear/no gear and low rpm), it should be set JUST lower than what you typically idle at, and then you want to tune it correctly in your a/c activation settings.
It could also even be overrun settings not adding fuel back in time for the motor to recover if you're coasting in gear.
I can think of a few more things, but without looking at logs...
dashpot is maybe what 2-4% idle bump? if you're stalling because you don't have dashpot adding, then you have other issues.
you have to make sure you hit the max decel value so CL idle will kick in even when coasting in N. It's imperative you tune the value correctly to prevent stalling in this very scenario (high gear/no gear and low rpm), it should be set JUST lower than what you typically idle at, and then you want to tune it correctly in your a/c activation settings.
It could also even be overrun settings not adding fuel back in time for the motor to recover if you're coasting in gear.
I can think of a few more things, but without looking at logs...
#9
Elite Member
iTrader: (1)
Join Date: Jun 2006
Location: Warrington/Birmingham
Posts: 2,642
Total Cats: 42
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
dashpot is maybe what 2-4% idle bump? if you're stalling because you don't have dashpot adding, then you have other issues.
you have to make sure you hit the max decel value so CL idle will kick in even when coasting in N. It's imperative you tune the value correctly to prevent stalling in this very scenario (high gear/no gear and low rpm), it should be set JUST lower than what you typically idle at, and then you want to tune it correctly in your a/c activation settings..
you have to make sure you hit the max decel value so CL idle will kick in even when coasting in N. It's imperative you tune the value correctly to prevent stalling in this very scenario (high gear/no gear and low rpm), it should be set JUST lower than what you typically idle at, and then you want to tune it correctly in your a/c activation settings..
If dashpot isn't added and other factors take over i.e. environmental/load changes since the last known position I can get a very low idle, or a stall.
Max del load etc is all irrelevant really as this is before PID has engaged.
IanR, Attaching the actual datalog rather than the screenshot of some of the values would be much more useful.
The aim, and the way OEM does it, (AIUI), is to hit a higher than normal idle then engage PID relatively quickly to drive it down. That's certainly how my OEM MK1 works, and the missus Suzuki, and my parent's FIAT500 & Landrover. (Other cars may work differently).
#10
Retired Mech Design Engr
![](/mt2011/images_td/ranks/supporter1.gif)
iTrader: (3)
Join Date: Jan 2013
Location: Seneca, SC
Posts: 5,011
Total Cats: 859
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
1. As a precaution for when CL is missed,go to idle valve test mode and, with valve closed, at normal operating temp, adj idle screw so you have 750 RPM. Then when CL does miss, you are less likely to stall.
2. It looks to me that your idle valve max duty cycle is too high. Did you actually test for this?
3. This is important: you cannot put slider to zero. Pretty sure that kills CL parameters. The Advanced allows you to adjust the ratios of PI&D, but the slider still controls the magnitude.
2. It looks to me that your idle valve max duty cycle is too high. Did you actually test for this?
3. This is important: you cannot put slider to zero. Pretty sure that kills CL parameters. The Advanced allows you to adjust the ratios of PI&D, but the slider still controls the magnitude.
#11
Boost Czar
![](/mt2011/images_td/ranks/supporter1.gif)
![](/mt2011/images_td/ranks/admin.png)
iTrader: (62)
Join Date: May 2005
Location: Chantilly, VA
Posts: 79,688
Total Cats: 4,113
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
not to mention the PID values themselves look bogus.
it's a key in preventing CL to kick in when you dont want it. It's possible to have the idle valve engage while you are coasting in low rpm and in high gear, the CL code might kick in, try to lower you RPMS while youre still moving, close the valve completely, and when you hit the clutch to stop the valve is closed because CL was running and the car stalls.
I absolutely hated using last known value. I know others still use it, but I was all about the initial idle duty table, that worked brilliantly for me--the last known value always gave me headaches and I was always bugging Ken to change it when that was your only option.
Max del load etc is all irrelevant really as this is before PID has engaged.
I absolutely hated using last known value. I know others still use it, but I was all about the initial idle duty table, that worked brilliantly for me--the last known value always gave me headaches and I was always bugging Ken to change it when that was your only option.
#12
Elite Member
iTrader: (1)
Join Date: Jun 2006
Location: Warrington/Birmingham
Posts: 2,642
Total Cats: 42
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
How do you deal with additional loads that don't have an adder i.e. lights when using the initial duty table.
I found that with the initial duty table it was spot on on the drive to work, (lights off), but on the way home with the lights on the idle went irritating low until PID could take over.
I found that with the initial duty table it was spot on on the drive to work, (lights off), but on the way home with the lights on the idle went irritating low until PID could take over.
#13
Boost Czar
![](/mt2011/images_td/ranks/supporter1.gif)
![](/mt2011/images_td/ranks/admin.png)
iTrader: (62)
Join Date: May 2005
Location: Chantilly, VA
Posts: 79,688
Total Cats: 4,113
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
I never had that issue, so i dunno. My headlights would reduce the voltage so I tuned the battery voltage comp table. But otherwise, lights or fans would have no real effect on the idle valve duty...I'd always expect to be idling around 32-33% idle duty.
It would be a cool feature to allow extra inputs to bump idle duty by X%.
It would be a cool feature to allow extra inputs to bump idle duty by X%.
#15
Retired Mech Design Engr
![](/mt2011/images_td/ranks/supporter1.gif)
iTrader: (3)
Join Date: Jan 2013
Location: Seneca, SC
Posts: 5,011
Total Cats: 859
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
I tried moving the slider to zero this AM, and it really had little effect. In addition, OP problem is CL not engaging, rather than performing poorly (though that may be the case as well).
Reverant has been lobbying for P/S input for IAC increase. NB already has sensor and wire. He has not prevailed.
On MS3 Basic, the battery control is so good, that you can't use voltage drop for anything. Once the engine is running, my voltage at the MS is always 14.0 +/- about 0.1V.
Richy, the issue may be less lights and more MAT. I error my targets for a little high with cooler air, and then don't have sag with hotter. IIRC, the actual idle at fully warmed up is 24% at MAT=60, but 27% for MAT=90. So my target is 27, and in colder air, the RPM pauses about 200 RPM higher until CL drives it down.
With A/C adder, I found that when CLT is low, the adder, together with the RPM adder would put me in a high idle situation. But, at the same time, the idle speed setting at low CLT is 1100 RPM. When you add the fan adder of 50 RPM and the A/C adder of 300 RPM (my settings) you get pushed to 1450 RPM. I simply moved the 100*, 1500 RPM setting of the target table down unitl the total IAC% target was good for the lower temperature.
tldr: With the target table, you can get creative and match actual conditions for uniform behavior.
Reverant has been lobbying for P/S input for IAC increase. NB already has sensor and wire. He has not prevailed.
On MS3 Basic, the battery control is so good, that you can't use voltage drop for anything. Once the engine is running, my voltage at the MS is always 14.0 +/- about 0.1V.
Richy, the issue may be less lights and more MAT. I error my targets for a little high with cooler air, and then don't have sag with hotter. IIRC, the actual idle at fully warmed up is 24% at MAT=60, but 27% for MAT=90. So my target is 27, and in colder air, the RPM pauses about 200 RPM higher until CL drives it down.
With A/C adder, I found that when CLT is low, the adder, together with the RPM adder would put me in a high idle situation. But, at the same time, the idle speed setting at low CLT is 1100 RPM. When you add the fan adder of 50 RPM and the A/C adder of 300 RPM (my settings) you get pushed to 1450 RPM. I simply moved the 100*, 1500 RPM setting of the target table down unitl the total IAC% target was good for the lower temperature.
tldr: With the target table, you can get creative and match actual conditions for uniform behavior.
#17
Elite Member
iTrader: (1)
Join Date: Jun 2006
Location: Warrington/Birmingham
Posts: 2,642
Total Cats: 42
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
I just went out and had a play, I must be loosing it as I can't find a voltage compensation table other than the one for the igntition dwell.
DNMakinson, thanks for that. I'll have another play with the initial duty table
As an aside I've just fitted an AEM to replace an LC-1 and I'm not that impressed. it's really slow! EGO corrections completely mess the idle stability up as it's so slow to respond to changes.
Sensor location in the car is unchanged btw.
DNMakinson, thanks for that. I'll have another play with the initial duty table
As an aside I've just fitted an AEM to replace an LC-1 and I'm not that impressed. it's really slow! EGO corrections completely mess the idle stability up as it's so slow to respond to changes.
Sensor location in the car is unchanged btw.
#18
Retired Mech Design Engr
![](/mt2011/images_td/ranks/supporter1.gif)
iTrader: (3)
Join Date: Jan 2013
Location: Seneca, SC
Posts: 5,011
Total Cats: 859
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
As to my guess to the lack of catching CL, I don't see anything in the settings that is way out of line (I admit that I don't fully understand all of the lock-outs), and so I would suggest working with the target Duty Cycle table, tied to CLT. Increase all the values up by about 3 over the logged, typical control values, and then tune them down individually until RPM's hang at the right place and allow CL to engage.
The other thing to look at is a log of acutal TPS readings at closed throttle. If sometimes greater than 1.0, then CL will not engage.
#19
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
I just went out and had a play, I must be loosing it as I can't find a voltage compensation table other than the one for the igntition dwell.
DNMakinson, thanks for that. I'll have another play with the initial duty table
As an aside I've just fitted an AEM to replace an LC-1 and I'm not that impressed. it's really slow! EGO corrections completely mess the idle stability up as it's so slow to respond to changes.
Sensor location in the car is unchanged btw.
DNMakinson, thanks for that. I'll have another play with the initial duty table
As an aside I've just fitted an AEM to replace an LC-1 and I'm not that impressed. it's really slow! EGO corrections completely mess the idle stability up as it's so slow to respond to changes.
Sensor location in the car is unchanged btw.
PWM Idle Voltage Compensation
It's right before A/C Idle up.