Open loop at ~4500 RPM is useful for setting primary injector scaling.
If you can get that range to stoich with scaling, you can then establish the MAF curve much more accurately, using LTFT at idle.

 

What type of percentages should the primaries be scaled by when making adjustsments?

Is it ever necessary to make adjustments to the injector latency?

 

start with about 1/2 of the amount of deviation from the target lambda.


Only if you change the injectors.

 

thats what i thought however every time i do the math and compare it to the load the pcm is spitting into my AP - they do not match

 

...and BARO comp and IAT comp and ECT comp..
etc.

BTW - I know this is condescending and I know it will be taken as some sort of nefarious suggestion aimed at quashing "competition" or something or other, but if you don't have an innate "feel" for what is going on in these calculations and how the scaling and curve shaping affects engine output, you probably are not going to ever "get it".
You will just find yourself running in ever tighter circles, trying to wrap your head around the numbers and getting more and more frustrated as seemingly simple and carefully crafted and calculated inputs yield wildly varying and often contradictory outputs from the motor.

Mazda (Denso) gave us a wonderful tool with which to correlate sensor inputs and produce actuator outputs. It is very precise and has so many overlapping layers that, like Photoshop, is able to produce a very nuanced result.
However, unlike Photoshop, there are no "filters" or "actions" that do some of the magic for you and crafting a decent calibration is a bit like trying to edit in Photoshop using a 10-year-old CTX CRT monitor with an intermittent red gun and a ton of gamma error and barrel distortion.
You have to do a LOT of creative mental correction to the absolute numbers to produce a result that will be correct in the "real world".

 

^^^^wow thats a lot of words...

 

231 words, to be precise.
Were there any in particular with which you have a contention?

 

 

Hey - There are a half million words in the English language.
You should be thankful that I don't try to explain this stuff in single syllables.
The posts would be four times as long and ten times as ignored.

 

Okay. Once I have that sorted, do I only adjust the idle portion of the MAF, or the entire MAF itself?

 

Set the target lambda to 1 = 14.7 AFR.

But use lambda makes math easier. Then go find a level place where ou can load your car to 4500 rpm or so and hold it for 20 seconds. Note the actual lambda and adjust the maf scale by the % deviation.

So if you load it to 4500 and your average lambda is 1.1 and you are aiming for 1 - then just add 10% to your MAF scale at that voltage point.

Repeat for as many scales on the MAF as you can.... if they all trend the same way you can make some educated guesses about how to scale the MAF... IE if 1.1,1.2,1.3..... all the way to 1.6 volts are all off by 10% you can be pretty confident that scaling the whole MAF 10% will get you closer than your stock calibration.

 

Why use a small word when a diminutive one would do?

 

one thing i have done which I think MM has alluded to (and seems to work) is not to try and scale the maf with an equal % at all points . I only do the ltft correction up to about 20g/s and apply a different correction from then on ...

 

Well yeah they dont always scale together.... but you can test as many as you can in the cruise range. You can keep scaling all the way from 0-5v - but for us boosted folks that can be a bit dangerous..... LOLZ

Ultimately you will scale the whole MAF; but to get the higher portions closer to actual you want to start down low where it is safer.

 

So here is something to ponder, in a tuner discussion/learning kinda thing...

When do the MAF adjustments end and the injector adjustments start??

Cheers,
Hymee.

 

I took ltft in closed loop and adjusted up to 20 g/s by that % then took what was happening at high loads and adjusted that by the % that was out by and just kinda smoothed it out in between .
seemed to work well .

 

In a perfect world - you would have your injectors flow tested.... barring that I would use the idle area to determine if your injectors are off; the P1 anyway...

I havent really decided how to do it otherwise.

 

You start with the injector scaling at 4200 RPM and 60 g/sec or so.
Then you observe the LTFT and scale the MAF.

Injector and MAF granularity at idle is to broad.

 

Do you think you should add STFT to LTFT to get an overall trim? I know the dude-logger doesn't record it (yet).

Cheers,
Hymee.

 

This is all nice discussion, BTW.

 

If you are scaling the injectors first; how do you know your MAF is accurate in order to scale it?

I was recommending idle because every RX8 I have ever plugged into all idle at 5.5-6 g/sec.... so if your MAF reading is accurate (in that range) and your fuel trims are still off; I would chalk it up to an injector scale issue.

 

mm has turned over a new leaf

 

Hardly.

I'm still a firm believer in giving a man only enough rope to hang himself with.

 

Recently I installed a AEM intake, RB header, and catless midpipe with this setup I would think that my STFT and LTFT would be way out to lunch but they are not. My STFT bounces around from -11 to +4 while my LTFT is 0 throughout the rpm band. Does this sound right? I'm running the Cobb stage 1 map because it seems to run better than the stock map. The only thing that I see that isn't right are that my AFR's are in the 11's at high rpm's.