RomRaider

The faster the better on the freeway.

Also, some 55 mph constant driving is good too.

You do need some street driving. Cruising on the boulevard would be the best.

Mix it up but try to keep your throttle movements smooth and constant if possible. No sharp jabs (unless you need to of course).

The logs that you sent me were fine, but I need more high speed driving on the freeway.

mickey,
sent you an email a little bit ago, just an FYI..

mickeyd,

I've spent some time logging my prodrive cai for my 02 wrx with stock injectors.
I have attached my graph data tab for a little review.

thanks in advance,
dm

Let me throw a little "shout out", if you will, to mickey..

Mike has been helping me for the past couple weeks, to help dial in my MAF issue (3" MAF, modded injectors, same-time-swap)..
During this period of time, I started out (after driving the car for about 1600mi), at getting about 20.2-21.5 MPG. Using this spreadsheet, the first couple of iterations, the scatter plot looked like a buckshot blast from 100yds. I mean, it was all over the place, literally.
In the last two weeks, he's helped me with 2 latency changes, and about 3 or 4 MAF scalar changes. each time, the car gets A LOT better and better.

when I filled up with fuel on Monday, I'd went 338mi, and 14.6gal. = 23.2mpg. That's about 1mpg less, than I ever got with the stg2 2.0L.

Dyno Log Lab shows I'm making around 350whp/340Tq @ 15psi (wastegate spring only). After his latest iterations for the MAF, I filled up yesterday evening..
12.2gal, 329mi. = 26.96MPG... @ $3.39/gal, I'm really liking that!

He's now got me very close to where my data is quite stable, with only a few flashes. My biggest outstanding problem, now, is a real rich spike when I let off the throttle, and unless I kick to neutral, blip the throttle, and roll to a stop, the car WILL die, as it goes pig-rich and struggles/takes too long to recover, thus, killing the engine. An immediate restart, and it's fine..

At any rate, wanted to publicly (forum-ly) thank mike for all his assistance and getting my car much closer to perfect, and much more enjoyable to drive, via this spreadsheet! (although, it still confuses the hell out of me! )

That sounds like Mike.

He has helped me a lot to. I try and repay by helping others when I can.

what do you do if you do not have "injector pulse width"? I can only see "injector duty cycle" on my 02 WRX ECU? I'm sure I have the latests defs too.

I have been doing some rescaling myself. What should you do if you have AFRL at say positive 5, and AFRC at -5? I believe the total correction is then 0, but still, it would be nice to have everything at 0.

By definition, AFR Correction will never stay at -5%. If that was a steady state condition, AFR Learning would move to 0 by itself. You can not just look at one data point. Short term fuel trim oscillates. The larger the error, the more the oscillation.

You must have injector pulse width. The ecu does not have an injector duty cycle parameter. If you look inside the defs, you will see that IDC is just a function of IPW and RPM. I would recommend that you not log IDC. IPW is the important parameter. You can always calculate IDC later the same way the logger would.

For this spreadsheet, I use IPW direct because it has higher resolution but does not include latency. You need to download the latest defs.

my defs do not have IPW...only IDC. I'll check to make sure I have the latets defs.

I call it IPW as shorthand, but in the latest defs, it should be called:

Fuel Injector #1 Pulse Width (Direct)* (ms)

mickey,

just wondering - why did you take out the MAF Table tab out of the latest version of your spreadsheet "Closed Loop Fueling Rev 6.zip" ? I tot it was useful as I have the new values that I need to input to my MAF table directly from the MAF table tab.

Or should I just use the older version "CL AFR analysis updated 4.zip" ? Thanks

I took it out because you really need to use some judgment as to how accurate the data is. I didn't want people to blindly use the numbers especially since it was based on a curve fit. The curve fit can cause the data near the ends to become funky because of the polynomial fit.

Mickey,

Ok, I see why now. I'm not a statistical guru, so let me know if the following judgements are correct when determing if the column "S" in Graph Data should be applied to my MAF table (based on your notes in the spreadsheet):

1. MAF Check 1
If the correction (y-axis) for MAFv (x-axis) is within +-2% I can use that MAFv in column "M" in Graph Data and its corresponding value in column "S"

2. MAF Check2
You mentioned that R2 needs to be very close to 1.00. If my value is not close to 1.00, does that mean I should discard this set of logs that I have used for the spreadsheet?

How close to 1.00 is close? Is 0.98 considered close enough?

3. MAF Check3
If the particular MAFv has a correction value between +-5%, I can use the the corresponding value in column "S" in Graph Data.

For just MAF Scaling, I do not need to use Injector Raw Data, right? It just shows whether my injectors are scaled correctly, right?

In the tab labeled graph data, I use column S for my scaling.

However, before I use column S, I visually check the scatter in the graph labeled MAF Check 3. I also visually check AFR Learning Ranges A-D.

MAF check 2 and Injector Raw Data tells you if you have a problem with injector latency or scalar. A wide cloud in MAF check 2 can not be corrected using MAF scaling. In MAF check 2, a straight line indicates good scaling. A R value of 0.98 is good.

When everything is said and done, your MAF Check 3 scatter should be mostly between -5 and +5 percent with a small amount of data beyond that.

BTW, MAF check 1 is just a visual representation of column S.

When doing a visual check on AFR Learning Ranges A-D, what should I be looking out for?

After doing several rounds of logging, I've input the data into the spreadsheet. I noticed that for areas in MAF Check 3 where the scatter is quite sparse and not within +-5%, the corresponding values in column S is null. So am right to say that actually the values in column S already takes into account (to some degree) that only valid data is displayed in Column S?

What if the values in column M (MAFv) in Graph Data does not match my MAF Table.

For example:
My MAF Table is
1.21 1.25 1.29 1.33

But the Column M has:
1.20 1.22 1.24 1.26 1.28 1.30 1.32 1.34

Do I average out Column M 1.20 and 1.22 to get the correction for 1.21 in my MAF table?

Column S is based on relative percentages. So, if you are looking for what percentage change to make at 1.33 volts then you would look at what is going on around 1.32 and 1.34 volts. It's not going to be that different.

If column S is blank then you don't have enough data. Column S is filled in only if there is enough data.

The very low MAFv values are difficult to adjust. You need to find a very large empty parking lot and drive at minimum speed with a constant throttle.

The AFR Learning values should roughly correspond to the shape of the curve in MAF check 1 and 3.