MAF Scaling - Open/Closed Loop Fueling info

[troek]

awesome, thanks man. yeah im getting 11.2-11.4 so id like to see it back up to 11.8 but ill keep an eye on it. i use to get closer to 12, but i think on one of my cl maf scaling i smoothed out the whole curve and messed up some of the voltages over 3.0

[Turdinator]

So why do we have an Engine load compensator MP table? Why would Subaru/Toyota need to adjust the engine load based on the MAP sensor if we have a fully functioning MAF sensor? I was going to zero it out to remove some variables from my MAF scaling logs but I am thinking it might be best left in there. What are peoples thoughts?

[steve99]

Originally Posted by Turdinator

So why do we have an Engine load compensator MP table? Why would Subaru/Toyota need to adjust the engine load based on the MAP sensor if we have a fully functioning MAF sensor? I was going to zero it out to remove some variables from my MAF scaling logs but I am thinking it might be best left in there. What are peoples thoughts?

I'll take a stab in the dark at it,

I would imagine the MAP (manifold pressure) sensor is used for barometric pressure and to make adjustments to the MAF (air flow) calculations when the barometric(atmospheric) pressure changes mainly to account for altitude changes.

It may also be used in failsafe or limp home mode if the MAF is detected as faulty or out of range.

Except if your using Speed Density (which I know nothing about) the MAP sensor becomes the primary sensor I believe.
@jamesm would be the best person for that.

[jamesm]

The map sensor is used during cranking. As far as the load comp, I wouldn't recommend messing with it. You shouldn't have to to achieve minimal fueling error across the board.

[Td-d]

The load comp tables are used to deal with resonance issues in the fueling lines - I'd expect they are less pronounced in the ft86 compared to the older Subarus. Yes, they are used as a failsafe (a very crude SD basically), but manifold pressure is referenced extensively in the code.

And yes - don't mess with it unless you have fundamentally changed the fueling system.

[jamesm]

The overrun feature doesn't actually seem to work from what i can tell. Load still just flat-lines regardless of what you put in there. Maybe i'm just doing it wrong... but i didn't waste too much time on it as i use SD for anything that moves enough air to max it.

[Td-d]

Originally Posted by jamesm

The overrun feature doesn't actually seem to work from what i can tell. Load still just flat-lines regardless of what you put in there. Maybe i'm just doing it wrong... but i didn't waste too much time on it as i use SD for anything that moves enough air to max it.

I suspect it's just a left over from previous code - the developers often leave code in there, and just disable it de facto through the calibration (which is why the RPM delta is -10,000 as opposed to -17 normally). Makes sense, given the duel fueling.

[Turdinator]

My missus is out of town and instead of catching up on my porn watching or gaming i thought i would do some serious logging and get my MAF scale sorted. I made a tune based on Shiv's stage one where I have forced open loop and have a flat command AFR on the direct injectors only. After a drive around the block to confirm nothing was going to blow up I noticed my fuel system status hasn't moved from 1 to 4 as I would expect. Has running richer made the AFR sensor slower to heat up or have I changed something I shouldn't have?

[s2d4]

Originally Posted by Turdinator

My missus is out of town and instead of catching up on my porn watching or gaming i thought i would do some serious logging and get my MAF scale sorted. I made a tune based on Shiv's stage one where I have forced open loop and have a flat command AFR on the direct injectors only. After a drive around the block to confirm nothing was going to blow up I noticed my fuel system status hasn't moved from 1 to 4 as I would expect. Has running richer made the AFR sensor slower to heat up or have I changed something I shouldn't have?

Can you screencap your changes?

[Turdinator]

Originally Posted by s2d4

Can you screencap your changes?

I am pretty sure this is all the tables (sorry new computer so i am struggling here)
https://drive.google.com/file/d/0B8k...it?usp=sharing

[s2d4]

Thanks for the screencaps.

This is only speculation and you may need people like td-d to help.
From my understanding, warm up uses PI then switches over to DI when idle.
Could it be that your PI was never on so it just gets stuck in the warm up cycle, status wise?

Another question out of curiosity, what are you hoping to achieve?

[steve99]

Originally Posted by s2d4

Thanks for the screencaps.

This is only speculation and you may need people like td-d to help.
From my understanding, warm up uses PI then switches over to DI when idle.
Could it be that your PI was never on so it just gets stuck in the warm up cycle, status wise?

Another question out of curiosity, what are you hoping to achieve?

I was going to suggest something similar.

since he has disabled any transition to closed loop ECU just stays in open loop.

As I understand things at startup it in open loop warmup mode status =1, as ECU cannot transition to CL at any point its probably just staying in that mode ie 1 as its effectively been stopped from using CL mode ever.

Suppose it depends how its coded, probably never get to routine to change fuel status ram address.

I suspect he is trying to balance his DI & PI injectors ? I think @mad_sb has done this so he may remember

[jamesm]

Balancing PI/DI is much easier, imho, using closed loop. I've tried both ways extensively, and get much better (and certainly faster) results by just doing the balancing in closed loop with the old-school spreadsheets and/or scripts. you don't need any data above ~2.9v to get it perfect.

Also i'd caution against changing the DI flow rate values or anything of that nature. It's extremely important, in my view, to follow the principle of least intrusion. If you didn't change the DI hardware, the corresponding calibration values don't need to change. This is particularly important because the DI (when left in it's completely stock state) serves a very useful purpose: it becomes a 'rock' throughout the rest of the tuning/calibration process. It's a known-good (enough) value that you can compare other things to.

My process is this: disable port injection. scale maf to center mass against the DI only (process separation). switch to 100% port... adjust port scaler/latency against the now-known-good MAF. recombine and fine-tune in the MAF scale.

This, while perhaps not resulting in the absolute lowest fueling error imaginable (there is pre-existing DI fueling error that isn't being corrected directly), makes for a quick, repeatable and reliable way of getting things in line. This process will produce a vehicle that runs better than a stock car at every load, every rpm, every time regardless of the vehicle's specific hardware configuration. There's no real value in doing any more than necessary to achieve that, and will often create more problems than it solves.

This is all assuming that you've changed port injectors. If not, then principle of least intrusion would suggest that you shouldn't be mucking with the port configuration to begin with, and just doing regular old WRX-style closed loop MAF adjustment (ignoring port/di balance entirely).

[Turdinator]

Thank you for your input guys. I was trying to do similar to what Jamesm was talking about as far as getting a good known maf scale on the DI to then scale the PI against. My car is pretty much stock so what I am doing is overkill but I figured it was a good way to learn more tuning etc.


I deliberately didn't change the cold port injection ratio table as I didn't want to mess with the normal warm up procedure.