Page 1 of 3
Idle Control Algorithms - Brainstorming
Posted: Mon Jul 09, 2012 11:35 pm
by Fred
Go nuts!
I don't mean controlling air flow, that's the obvious one, be it by solenoid, PWM valve, stepper valve, drive by wire or wax stat.
I mean:
Tuning the idle, boxing it
Conditionals to determine that you are or aren't idling
Adjusting RPM with timing changes
Inputs and outputs from and to the code
etc.
Re: Idle Control Algorithms - Brainstorming
Posted: Mon Jul 09, 2012 11:46 pm
by johntramp
What is the target of an idle alg? To hold a specific RPM, or MAP value or lambda value or ...?
Re: Idle Control Algorithms - Brainstorming
Posted: Mon Jul 09, 2012 11:49 pm
by BenFenner
Go ahead and let those of us out of the loop know what idle stuff is incorporated right now (none? just idle valve PID?) and let us know of anything you've seen you consider a band-aid or bad practice. I don't want to suggest things just to get them shot down.
Also, I might consider trying to revisit the PID idle valve control stuff in another thread when that time comes.
Re: Idle Control Algorithms - Brainstorming
Posted: Tue Jul 10, 2012 12:46 am
by Fred
John, in terms of tuning, keeping a stable AFR is the goal, and the core goal is stability of RPM, which comes from stable AFR and stable timing.
Given an engine at rest, you can feed it more or less air, more or less fuel, more or less timing, in order of responsiveness.
It may be subjected to more or less load from the clutch/tq-converter, A/C pump, alternator (headlights, fans, pumps, stereos), power steering, etc.
For some of these you may or may not get an input. For some of these you may or may not want to react. For example it pisses me off NO END that this Seat Ibiza TDI tries to stop me labouring the engine below nominal idle by applying more throttle, it's borderline dangerous and totally unnatural. Many cars do this due to over enthusiastic and overly stupid algorithms and lack of inputs. Also for example, you don't want to fall to 600rpm when the air con compressor fires up.
Ben, nothing what-so-ever. The first thing that could help with idle was the GP output with hysteresis. It could be used with an on/off solenoid valve based on temperature. Nothing else is plumbed up. The next thing that I add will be basic open loop PWM outputs with 2D tables to drive them, but I'm in no rush to do that.
What I want is everything, logic, tables, values, configurables, options, etc, between the inputs and the outputs. Don't be shy, I have very few thoughts on this matter, hence it's a brain storming thread, not a "this is how it is" thread or a "what do you think about this detail that I haven't quite sussed" thread, the latter two of which are more common. :-)
Re PID: Imagine that all of the airflow control mechanisms are available and tuned properly, what inputs to those outputs would you give? etc.
Fred.
Re: Idle Control Algorithms - Brainstorming
Posted: Tue Jul 10, 2012 11:39 pm
by Bangbug
And temperature of engine for load input. (as all engines will be effected (or affected if you like) by this)
Unless you have that sorted as part of the cold start up?
Actually won't this more or less BE part of the start up?
Re: Idle Control Algorithms - Brainstorming
Posted: Wed Jul 11, 2012 8:56 am
by Fred
"Re PID: Imagine that all of the airflow control mechanisms are available and tuned properly, what inputs to those outputs would you give? etc."
Right, the airflow would be set based on engine temperature :-)
Re: Idle Control Algorithms - Brainstorming
Posted: Thu Jul 12, 2012 12:35 am
by BenFenner
Fred wrote:Re PID: Imagine that all of the airflow control mechanisms are available and tuned properly, what inputs to those actuators would you give? etc.
Fixed?
First things first. I'd like to see ignition timing trim versus idle error/delta. 2D table.
Next is obvious; fuel trim versus idle error/delta. 2D table.
Those are the clear, obvious ones. To me anyway.
Re: Idle Control Algorithms - Brainstorming
Posted: Thu Jul 12, 2012 1:10 am
by Fred
Not fixed, broken. Outputs go to actuators. Inputs go to code, and output from code goes to outputs in hardware. What inputs to that code/logic/tables that generate the outputs for the actuators, do you want?
Please elaborate on the form and purpose of your two tables, in detail, such that a noob could understand :-)
Fred.
Re: Idle Control Algorithms - Brainstorming
Posted: Thu Jul 12, 2012 5:37 am
by Bangbug
Aka If I don't understand, you're not explaining very well.
Re: Idle Control Algorithms - Brainstorming
Posted: Thu Jul 12, 2012 8:46 am
by Fred
What he said! :-)