Unstable Idle, TPS doing unexpected things

[tineira]

Hi,

I'm having issues with unstable idle. I'm using open loop, ethrottle idle, with ignition correction. The problem I have is that sometimes, specially when the engine is hot, it seems that the throttle closes and then opens without relation to the expected open loop values.

Here you can see that the idle base position should be 1.4%, but the TPS is 0.4%. then it auto corrects but go past 1.4 straight to 2% and then again to 0.4%.  this causes the idle to oscillate and sometimes even stall the engine.

20240725 1700.pclx ECU Log 2024-07-25 5;51;30 pm.zip

[Adamw]

None of the relevant channels are logged.  Can you do a PC log please.  

[tineira]

PC Datalog - 2024-07-28 5;27;00 pm.llgx

Sorry about that, attached is the new log.

what channels are you looking for? 

 

Thanks for your help

[Adamw]

39 minutes ago, tineira said:

what channels are you looking for?

For assessment of a control loop the bare minimum would be "set point" (e-throttle target), "process variable" (TPS) and "controller output" (Motor DC), two of those where not in the ECU log.  Then depending on what clues those give you, then you may also need to look at other variables that can influence the control loop such as Aux 9/10 supply voltage, 5V output, sometimes even odd stuff like MAP might influence throttle control. 

In this case the issue looks like mostly friction/stiction related.   In the snip below you can see at the left hand end the TPS was sitting slightly above target, the motor DC was at about 20%, the ecu starts reducing motor DC to bring TPS down to target, it needed to go all the way to -30% before it moved.  So, a 50% change in motor power to get the throttle to move 0.1%...    

At about 39sec the target changes by 0.4% and again you see the motor DC had to wind down another 20% before the TPS started to move - but this time all of a sudden the TPS "unsticks" and jumps to 0%. 

 

Your Batt voltage and Aux 9/10 voltage are low so I would suggest fixing those first before doing much more control loop tuning as those have quite an influence.  

If the issue is still apparent with normal voltage restored, then you can often improve this stiction effect by introducing some "noise" using derivative.  A little bit of deadband will also prevent the windup affect when you are close to target and you will probably want a bit more P & I to improve response.  

I would try the P = 5.0, I = 0.1, D = 80, and 0.1% deadband as the first test.  

[tineira]

Thanks, I will try with another alternator and get back with news 

[tineira]

Hi,

I changed the alternator for another one, but for some reason I'm seeing the same voltages, I will have to figure that out. 

On 7/28/2024 at 7:26 PM, Adamw said:

I would try the P = 5.0, I = 0.1, D = 80, and 0.1% deadband as the first test.  

Anyway I tried this values and they seem to have fixed the issue!

Thanks for your help.

[tineira]

Hi,

I fixed the voltage problem, but strangely this made the problem worse, now with the old and new settings, I'm seeing the same issue with the e-throttle.

Can you help me with some idea on how to configure the values? Also is this something normal to deal with or it might be a signal that there is something wrong with the throttle itself? I bought it new from amazon, supposedly directly from Bosch.

 

https://drive.google.com/file/d/1yfWIRYEREG-nqZDXb7HsJC4m_ewNt6N7/view?usp=sharing

 

Thanks for your help.

20240918 1230 New e-throttle.pclx

[Adamw]

I suspect there may be some mechanical issue with this throttle, but can you try this combo and do another log:

• PWM Freq 4000Hz
• Proportional 6.0
• Integral 0.1
• Derivative 100

And in PID set up, set this to E-throttle 1 so we get a bit more data.  

 

[tineira]

Thanks, the changes seem to have helped a lot. I could not replicate the issue. Anyway if its of any help, I attached a log file with the setting that you suggested.

https://drive.google.com/file/d/1WDzEViIlZvn5bdRDAsKAdoPS5QIi_Ztm/view?usp=sharing

https://drive.google.com/file/d/1LiDOprWIqEn3K_48tdyBjOsrrU-p46--/view?usp=sharing

 

I'll update if something else changes.

 

Thanks for your help!

 

[Adamw]

It looks like it might be close to usable now but it's still not as good as I would like.  The extra derivative is causing a bit of a bounce when it is travelling towards target at a high rate.  You could try increasing the proportional another point or so to see if that helps. It will start oscillating when you have too much proportional so it looks like it will still take more.