Pfc Idle Learning And E85 ?

[discopotato03]

Hi all , firstly I did search but need a little info anyway .

It seems the humble PFC tries to relearn its idle suff if you alter (I think) the first 4 x 4 box area of the fuel or idle tables , or if you change your idle speed and decel fuel cut numbers .

It also tries to relearn of course if initaised .

Now it has been mentioned in other recent PFC threads that these computers use a form of closed loop to chase the target idle speeds as in with timing and narrow band closed loop 02 sensor feedback .

I'm curious to know what AFR it would chase with E70/E85 or if you'd have to untick the 02 feedback and set a suitable AFR some other way .

I'm beginning to think that this strange off idle retard I'm getting is because I change settings so often that the PFC keeps attempting relearns but doesn't get the right proceedure each time .

Something to try today , cheers A .

Edited October 28, 2012 by discopotato03

[rob82]

If it doesn't relearn the closed throttle tps then it will never know when to start/stop using delta ignition control, that is what sounds like your problem. Delta ignition control is not something that the PFC learns it is a way of quickly recovering or removing rpm by adding or subtracting ignition timing.

I still don't think you get what lambda is, no matter what ratio of ethanol you use it will alway target lambda 1 in closed loop mode that's all a narrowband sensor is designed to target.

It doesn't matter what values you have in the bottom most cells as it doesn't reference them when in "idle mode", only when you get enough tps on board ilk it revert to using the main map.

[GTSBoy]

Also, it will be very unlikely to manage to properly run closed loop mixture control at idle, because the sensor cools down and slows down.

The words "closed loop idle control" only ever refer to idle speed, not mixture. Speed control is achieved via an idle air motor/valve/whatever, and by the delta ignition control.

[rob82]

Also, it will be very unlikely to manage to properly run closed loop mixture control at idle, because the sensor cools down and slows down.

The words "closed loop idle control" only ever refer to idle speed, not mixture. Speed control is achieved via an idle air motor/valve/whatever, and by the delta ignition control.

The sensor shouldn't cool down as most 02 sensors have there own heating circuit. The main issues with closed loop at idle is the slowed exh gas velocity which makes the switching from rich to lean slower. If you have a good ecu with decent injectors there should be no reason you can't run closed loop mixture control at idle.

[GTSBoy]

In my experience, the heater in typical narrowbands is not sufficient to keep them hot enough to do a good job. Gas speed is certainly also a factor.

[discopotato03]

Yes I do understand what Lambda is but learing how a PFC learns is my learning curve .

The big disadvantage with Datalogit is that we still don't have enough access to whats going on within the silver box because things like this idle learning is something shat should be visible and recordable .

Now for the record I innitialised mine and reloaded my dat file even though some say all you have to do is change idle speed or FC settings to drive it back into "learn" mode . I was reading a thread from the RX7 board last night and I think they mentioned that changing fuel tables in the first 2-3 load and speed sections does the same think . Actually no that was from a Rex site somewhere . I think they said that the learn business starts at 60c coolant temp too .

Anyways I went to start the learn process and was about to stick the Tech Edge WB probe back in but though no stuff it leave the std narrow band one in and enable 02 feedback . I let it warm up to around 45 deg and reset and reloaded my dat and let it idle for 10 min , AC on for another 10 min , demist plus parkers on for another 10 min .

Yes everything is smoother now though it does take some seconds to drop back to my target idle speed at full running temp .

At that stage I started a log and went to the monitor page to double check the closed throttle TP voltage which was 4.15 . At a slightly cracked opening the timing stil retards from 15 to ~ 13 for no apparent reason and picks up slightly up the rev range . I moved the TPS to a closed setting of 4.5v and left it at that . I realise that I have made a change and possibly need to relearn again but I'm still not convinced that my TPS is 100 % so I'll enquire how much new ones cost .

A couple of things I learned from the moniter page are that giving the throttle a very shallow rapid tap , barely enough to make the engines note change , shows no increae in TPS voltage so this is either too fast for PFC to read or ?

Also this very slight throttle retard business only happens one way ie slightly opening the throttle not getting beyond the retard point and very slowly returning to idle .

I'm wondering if the second lower plugs idle and WOT switches , if that what these sensor use , are doing weird things its sendig squiff signals to the PFC .

Just lastly when I next do a idle reset I'll have the headlights on as well because I think it needs greater electrical load to give a better overall result . Datalogit shows from memory 14.3 volts with no load and 14v with AC demist and parker lights on .

Anyway idle light load and drivability is better now and I'll be staying away from parameters that upset this dedicated idle zone .

Surely this is something thats easier on later computers like the ViPec plug in , cheers A .

Edited October 29, 2012 by discopotato03

[paulr33]

just check when you are datalogging that the hand controller is not plugged in

they share the hand controller bandwidth so its the datalogit can miss values when fighting for bandwidth with the hand controller in use at the same time

from memory the only way to make it re-learn idle is DATA INIT

[discopotato03]

I never had much time for the hand controller and only had it plugged in with Datalogit when I first bought that . I think the hand controller froze and has been gathering dust ever since . I reckon some sort of Lap Top interface is the only way to go .

Cheers A .

[STATUS]

for idle the single most important feature is latency and voltage change. Generally i tune the whole car, then initialise then reload tune and do idle learn last... gives the best results.

Another thing is rbs dont like idling at 14.7 generally unless everything is spot on....... 13-13.6 will often give the best vacuum.

[discopotato03]

Interesting point , off hand whats the Lambda number for that . Damn I wish Datalogit would let me set target AFRs .

Lowest manifold pressure would indicate the strongest induction pulses and its probably as good an inicator of a "strong" idle as any .

Not being a pro tuner I'm always trying to find ways of making the fuel tables settings do most of the work and minimising the accell bit so having the idle a tad richer and the cells below load wise the same probably helps out a bit here .

I still run into that split second rich spike when the throttle is closed at times .

Yes I also wondered about doing the idle learning after the tune was complete to tie up loose ends .

Thanks , cheers A .

[GTSBoy]

13.6/14.7 = ?

/sheesh