I have a DBW close loop idle issue

[motor]

I have a DBW close loop idle issue,When I gently repeatedly step on the accelerator pedal, the“E throttle ISC CL Trim” add a little value about 2.5%.What kind of control logic is this? This automatically added value adversely affects closed-loop idle speed control.
please see log  for details, thanks!(This happened  file time 0:03.630 and 0:26.654 )

 

 

[Adamw]

Hi Motor, yes it seems we have a bug or something unexpected going on in CL E-throttle idle, it was reported here last week:  

I have been able to reproduce it here and have passed all the info on to our engineering team so they can have a look at it.  Hopefully we will have a fix soon but in the meantime you might have to switch to open loop mode.

[motor]

Thank you very much！

When will the next update be released？

[Adamw]

On 6/9/2018 at 6:54 PM, motor said:

When will the next update be released？

I dont know sorry but will update these threads once they have found the issue.

[motor]

Hi Adamw,

Have you found a way to solve the issue?

[motor]

 This issue has not been solved, I have tested new firmware, exactly the same as before!

[Adamw]

11 hours ago, motor said:

 This issue has not been solved, I have tested new firmware, exactly the same as before!

I tested the fix for the problem that "a6-quattro" reported above using his map and it seemed to work correctly.  I assumed you were reporting the same problem but now after looking at your map I think it maybe just because your E-throttle target and base position tables are set up wrong.  The top row of the E-throttle target should not have zero in it.  Usually this should have most of the idle throttle requirement "%" for normal hot idle in the top row of the target table, the base position table should only have small numbers in it to allow for fast idle warm up and small correction to normal idle. 

Can you try to set up your base position and target tables like this and give us a new log and map if it is still not right for you.

 

[motor]

Already tested as you said, but still the same as before。please see log file，thanks!

 

 

On 8/16/2018 at 11:02 AM, Adamw said:

I tested the fix for the problem that "a6-quattro" reported above using his map and it seemed to work correctly.  I assumed you were reporting the same problem but now after looking at your map I think it maybe just because your E-throttle target and base position tables are set up wrong.  The top row of the E-throttle target should not have zero in it.  Usually this should have most of the idle throttle requirement "%" for normal hot idle in the top row of the target table, the base position table should only have small numbers in it to allow for fast idle warm up and small correction to normal idle. 

Can you try to set up your base position and target tables like this and give us a new log and map if it is still not right for you.

 

On 8/16/2018 at 11:02 AM, Adamw said:

I tested the fix for the problem that "a6-quattro" reported above using his map and it seemed to work correctly.  I assumed you were reporting the same problem but now after looking at your map I think it maybe just because your E-throttle target and base position tables are set up wrong.  The top row of the E-throttle target should not have zero in it.  Usually this should have most of the idle throttle requirement "%" for normal hot idle in the top row of the target table, the base position table should only have small numbers in it to allow for fast idle warm up and small correction to normal idle. 

Can you try to set up your base position and target tables like this and give us a new log and map if it is still not right for you.

 

Already tested as you said, but still the same as before。please see log file，thanks!

 

 

 

[Adamw]

Thanks for testing and the files.  I havent had enough time over the last couple of days to look at what is going on but I will try to test myself in the next couple of days and will update you then.

What were you using the timer 1 with the idle target trim function to achieve?  Is this to give a dashpot type effect or something else?

[motor]

Yes, I use Timer 1 to assist in the dashpot of  idle,but in the test it was disabled.

[cj]

It looks like that while CL idle is active, whenever you press the throttle to the point where you exceed the AP lockout threshold, it adds 0.4 to the CL idle trim immediately. I cant see where its getting this 0.4 from so it might be hardcoded, or it might be some multiple of your proportional gain or ethrottle base numbers. I assume the idea here is to have it gently return to idle RPM rather than try to "catch the stall" that happens with low ethrottle targets where the RPM can drop really quickly below where you want it.

You can see that in the 8 times you blip the throttle, it adds 0.4% for each of the first 7, and between them when you're at 0% AP, the CL logic is actually doing its thing to reduce the CL % as soon as you get off the gas pedal (note at 32.2 sec that the CL number drops from 1.0 to 0.9 and status is DEC IDLE). On the 8th blip of the throttle something different happens though because the RPM is already so high from the current throttle angle, that you are above the RPM lockout for CL idle, and this time the CL% doesnt move at all. In the later part of your log it start off the RPM is too high for CL to kick in and reduce the CL% (so catch 22 here), but eventually overrun ignition trim kicks in and pulls 10 deg of timing (35-36sec) because you are under 0.5% AP and over 1500RPM, which drags the RPM down enough for CL to kick in and start removing the CL trim% until you get back to normal idle at CL 0.1-0.2%. I'm not sure if you set this up on purpose to catch this or its just a co-incidence that it does a pretty good job of fixing the catch 22 of too high an idle RPM for CL to fix itself.

As far as how you fix this, you could try changing your open loop trims, proportional gain table, or ethrottle target table by a bit and see if you can influence that 0.4% number that is getting added. You could also try adding quite a bit the the proportional gain table to get it to drop CL% quicker once off the throttle. The link guys might have some info on the background logic at work here as to where this 0.4% comes from and why its being added every time you cross the AP throttle lockout number rather than just the first time like I would expect if its to ensure a smooth return to idle. You could work around it by just not blipping the throttle this many times this quickly so it has time to return to normal idle. I dont think increasing any of the CL lockout parameters will help much - increasing the lockout RPM you are just trading off how many times you can blip the throttle against how bad a high idle you have once you hit the problem. You might somewhat fix it by lowering the RPM lockout by quite a lot - it will still have the same CL % getting added but if the idle RPM exceeds lockout on the first or seconds throttle blips, its only going to be say 200RPM above target anyway and from this RPM you can probably tune your idle ignition control to pull back  timing and get you back to where CL idle will work correctly. Its all a bit of a workaround though.

[motor]

17 hours ago, cj said:

It looks like that while CL idle is active, whenever you press the throttle to the point where you exceed the AP lockout threshold, it adds 0.4 to the CL idle trim immediately. I cant see where its getting this 0.4 from so it might be hardcoded, or it might be some multiple of your proportional gain or ethrottle base numbers. I assume the idea here is to have it gently return to idle RPM rather than try to "catch the stall" that happens with low ethrottle targets where the RPM can drop really quickly below where you want it.

You can see that in the 8 times you blip the throttle, it adds 0.4% for each of the first 7, and between them when you're at 0% AP, the CL logic is actually doing its thing to reduce the CL % as soon as you get off the gas pedal (note at 32.2 sec that the CL number drops from 1.0 to 0.9 and status is DEC IDLE). On the 8th blip of the throttle something different happens though because the RPM is already so high from the current throttle angle, that you are above the RPM lockout for CL idle, and this time the CL% doesnt move at all. In the later part of your log it start off the RPM is too high for CL to kick in and reduce the CL% (so catch 22 here), but eventually overrun ignition trim kicks in and pulls 10 deg of timing (35-36sec) because you are under 0.5% AP and over 1500RPM, which drags the RPM down enough for CL to kick in and start removing the CL trim% until you get back to normal idle at CL 0.1-0.2%. I'm not sure if you set this up on purpose to catch this or its just a co-incidence that it does a pretty good job of fixing the catch 22 of too high an idle RPM for CL to fix itself.

As far as how you fix this, you could try changing your open loop trims, proportional gain table, or ethrottle target table by a bit and see if you can influence that 0.4% number that is getting added. You could also try adding quite a bit the the proportional gain table to get it to drop CL% quicker once off the throttle. The link guys might have some info on the background logic at work here as to where this 0.4% comes from and why its being added every time you cross the AP throttle lockout number rather than just the first time like I would expect if its to ensure a smooth return to idle. You could work around it by just not blipping the throttle this many times this quickly so it has time to return to normal idle. I dont think increasing any of the CL lockout parameters will help much - increasing the lockout RPM you are just trading off how many times you can blip the throttle against how bad a high idle you have once you hit the problem. You might somewhat fix it by lowering the RPM lockout by quite a lot - it will still have the same CL % getting added but if the idle RPM exceeds lockout on the first or seconds throttle blips, its only going to be say 200RPM above target anyway and from this RPM you can probably tune your idle ignition control to pull back  timing and get you back to where CL idle will work correctly. Its all a bit of a workaround though.

Hi CJ,Thank you very much for your input, thank you for reading my log carefully, you are right and very detailed! Forgive my English。
Regards!

[Adamw]

Hi Motor,

I have finally had some time to test this today.  I too see the same 0.4% CL trim added everytime the pedal is pressed.  Although this is similar to the bug that "A6 Quattro" reported in the linked thread above we have fixed that one (i tested again just to be sure) but we have never noticed this one that you have found.  So sorry we missed that when you reported it earlier, I thought it was linked the same problem that we were already fixing.

I dont know where that 0.4% trim comes from it seems to be hardcoded and not related to any user adjustable setting or table.  I basically zeroed out every table, setting & gain and it still got 0.4% added everytime...

I have passed this on to the engineering team to look at now, in the meantime you may be able to get acceptable idle control using some of CJ's ideas.