HOWTO: One Man's Way of Scaling a MAF

[Rogan]

I'm trying that now..

* Registered users of the site do not see these ads.

[Rogan]

Ok, I'm really second-guessing this whole self-tuning thing.. Maybe I should stop. I can't even get the spreadsheet right! LOL Here's what I got, will..

[williaty]

PMed you.

[Rogan]

cool. email sent. thanks

[williaty]

You just had the spreadsheet set up wrong to make a graph from it. See how I added the MAFv's for your MAF Scaling Map into the MAFv column before your actual data and then pasted the Means and Modes beside the MAFvs to make it all line up in columns?



I didn't check your data itself, you're on your own there.

[Rogan]

sweet! I see my error in the spreadsheet, thank you for your help!
Now I can continue with the directions!

[Rogan]

ty - in the last formula (cor g/sec), do i understand that for the Cor% row, you enter the 'mode' number? Or is it more so a "guess" of what you want to increase/decrease, judging by what the mode value is?

I.E. In the MAFv cell of the ROM, say it's 1.4v with 7.2 g/sec. The mode calculated from the log is -10.94. Would you enter "-10.94" in the Cor% cell, resulting in a Corr g/sec from 7.2 to 6.41, effectively "leaning out" a little, at that MAFv?

[williaty]

Originally Posted by bchslap

ty - in the last formula (cor g/sec), do i understand that for the Cor% row, you enter the 'mode' number? Or is it more so a "guess" of what you want to increase/decrease, judging by what the mode value is?

A little of both. If I were telling a computer to do it, I'd program the computer to use the modal value. 99% of the time, I use the modal value. I use the mean as a sanity check to see another way to analyze the data. The only time I don't use the modal correction is if the modal correction is 0 but the mean isn't. For instance, if the mode is 0 but the mean is 1.56, I'll probably use 0.78. This is because a 0 correction is the most common, but the error isn't "balanced" around that error. If you adjust the g/sec just a little bit, you'll almost certainly still have 0 as the modal correction, but you'll bring the mean closer to 0.

I.E. In the MAFv cell of the ROM, say it's 1.4v with 7.2 g/sec. The mode calculated from the log is -10.94. Would you enter "-10.94" in the Cor% cell, resulting in a Corr g/sec from 7.2 to 6.41, effectively "leaning out" a little, at that MAFv?

Yup.

[Rogan]

awesome! LOL thanks for your patience with my n00b questions

[R4ND0M_AX3]

I plot out mode, median, and average(mean). Then I look at the three and decide what I want my correction to be (also plotting this). This last time I started to notice an offset across the board. So I had excel average my desired correction. Then applied that average correction to my injector scale and my proposed MAF changes. In this case I reduced the size fo the injectors (increased ST fuel scaler) by 1%. Then reduced my proposed MAF changes by the same amount.
I have a stock MAF + APS inlet + larger injectors. So I don't want to stray too far from the factory MAF settings.

Just got back from a short test drive. Things look good. Still a tiny dip or bump or two. But slowly getting better. I just need to make sure my wideband readings at WOT are still close to the fuel table like I had them many moons ago.

[savageblitz]

ty,

Using your method, I've got huge corrections, across the whole MAFv range. I've got an APS CAI 70mm, AVO turbo inlet, TurboXS TMIC.

MAFv 0.94 0.98 1.02 1.05 1.09 1.13 1.17 1.21 1.25 1.29 1.33 1.37 1.41 1.48 1.56 1.64 1.72 1.8 1.88 1.95 2.03 2.11 2.19 2.27 2.34 2.42
corr g/s 1.7052 1.9488 2.1888 2.4955 2.9106 3.2982 3.62004 4.1825 4.64508 5.15592 5.67248 6.2894 6.9266 7.48 9.32695 10.685 12.848 15.539 17.796 20.920 23.218 25.451 29.344 32.186 36.45 39.854

The logs were generated after I increased the MAFv scale by 1.1 from the stock values (since I read on the forum that the stock is 65mm - after changing to 70mm, just scale by 70/65 = 1.1).

Is that normal?

[savageblitz]

My correction values... corrections are mostly > 15

MAFv Mean Mode
0.94, ,
0.98, 16.41 ,
1.02, 14.26 ,
1.05, 15.31 ,
1.09, 15.14 , 18.75
1.13, 18.01 , 19.53
1.17, 18.54 , 16.4
1.21, 18.69 , 19.53
1.25, 18.52 , 18.75
1.29, 17.59 , 18.75
1.33, 16.79 , 17.18
1.37, 17.50 , 17.96
1.41, 16.42 , 17.96
1.48, 8.725 , 6.25
1.56, 12.54 , 11.72
1.64, 12.60 , 9.37
1.72, 13.98 , 13.28
1.8 , 17.77 , 18.75
1.88, 20.20 , 18.75
1.95, 20.78 , 22.65
2.03, 19.87 , 19.53
2.11, 17.26 , 14.84
2.19, 16.63 , 16.4
2.27, 13.91 , 12.5
2.34, 12.49 , 12.5
2.42, 10.47 , 9.37

[williaty]

I have no personal experience tuning for that specific setup. However, there's a lot of things that could be going on, not the least of which is an actual mechanical issue with the car. I would try to find someone who has tuned that specific setup and ask them how much they had to modify the scaling. While every car is different, knowing what someone else did will at least tell you if you're in the ballpark or WAY off base.

[williaty]

Originally Posted by savageblitz

The logs were generated after I increased the MAFv scale by 1.1 from the stock values (since I read on the forum that the stock is 65mm - after changing to 70mm, just scale by 70/65 = 1.1).

Is that normal?

Again with the not having done this personally, but I'm pretty sure I remember discussion of this on either romraider or LGT and computing the scalar wasn't that simple (70/65).

[Airboy]

Originally Posted by savageblitz

ty,

Using your method, I've got huge corrections, across the whole MAFv range. I've got an APS CAI 70mm, AVO turbo inlet, TurboXS TMIC.

The APS 70mm intake messes up the low flow area a lot. Some data and from another forum.

http://www.iwsti.com/forums/cobb-str...libration.html

[RaceFaceXC]

so im using williaty's method described in the OP and im getting correction %'s of 8% and 9% for the first few MAF's... am I way way off or is this typical for the first MAF rescaling with a SRI?
i havn't gone any further because im afraid im way off but maybe im not. Any thoughts/opinions? thanks.

[williaty]

My lowest attainable MAFv needed scaled +22%, but it came back under 10% shortly after that:


Which results in this:


I'm still trying to get idle dialed in. Idle is a freakish place. I'm trying to get it to be pulling about 5% fuel at idle to make sure when the O2 sensor hasn't warmed up yet and started working that it's running rich enough to actually get itself started.

[savageblitz]

ty,

how do u read your first graph? is "change: the changes made to "stock" line that results in the "corrected" line?

your stock and corrected line are very close, what does that indicate?

btw, how do you know that idle is still not dialed in?

some other questions: using your method, can I merge two log files together but adding the time column from the previous file to the next file I am adding? the reason is becoz my defogger switch always goes off after about 15-20mins and that stops the logging. So its quite impossible for me to get > 20mins worth of logging everytime. any implications?

[savageblitz]

Originally Posted by Airboy

The APS 70mm intake messes up the low flow area a lot. Some data and from another forum.

http://www.iwsti.com/forums/cobb-str...libration.html

airboy, thanks for the link.. its very useful. I've compared my corrections with the ones in the link and they are quite close, so I guess I am doing ok.

in fact, I just did a reflash with the new corrections and the car runs ok so far. I think I still need a few more iterations though.

[RaceFaceXC]

anyone using OpenOffice.Org Calc app? Im getting tons of #VALUE! errors when i try to use these formulas and I cant figure out why. OO.Org's forum is so far no help and google i can only find answers applied to Microsoft Office Excel and i tried those fixes and they aren't helping. I love openSource software like Enginuity/RomRaider, EcuFlash and OpenOffice.Org (not to mention Firefox) but im getting frusturated.

N/M, think i figured it out.. my universe just got 10x better.

[williaty]

Originally Posted by savageblitz

how do u read your first graph? is "change: the changes made to "stock" line that results in the "corrected" line?

The Change line is the percent difference between the stock scaling and the corrected scaling.

your stock and corrected line are very close, what does that indicate?

That they're not very far apart Seriously, I dunno. This intake isn't screwing up the scaling much and I think that sort of a plot makes the values seem closer than it really is.

btw, how do you know that idle is still not dialed in?

The trendline is still curvy over there. I'd like to get it to consistently pull about 5% over there, but I think that's a pipe dream. I've changed my AF Learning Airflow Ranges table to confine all of idle to AF Learning #1 A, so at least the chaos is confined.

some other questions: using your method, can I merge two log files together but adding the time column from the previous file to the next file I am adding? the reason is becoz my defogger switch always goes off after about 15-20mins and that stops the logging. So its quite impossible for me to get > 20mins worth of logging everytime. any implications?

The better way to merge log files would be to process them down to the point where the only columns you have are AFL, AFC, and MAFv. Then just paste the logs one after the other. Then sort by MAFv and create the column for Cor%.

You can also use the 774 and newer builds and just click the "Log to File" button for long logs, which is what I do.

[RaceFaceXC]

I finally finished all the computing for my first stab at scaling the MAF and got

MAFv g/sec % cor cor g/sec

1.17 2.48 8.7 2.7
1.21 2.72 8.5 2.95
1.23 2.85 7.8 3.07
1.25 3.04 8 3.28
1.29 3.46 7.5 3.72
1.3 3.57 1 3.61
1.33 3.89 1 3.93
1.37 4.35 3.1 4.48
1.38 4.5 1 4.55
1.41 4.83 3.1 4.98
1.45 5.44 4.5 5.68
1.48 6.03 5 6.33
1.56 7.27 5 7.63
1.64 8.59 4.5 8.98
1.72 10 3.8 10.38
1.8 11.55 4.5 12.07
1.88 13.25 3.9 13.77
1.95 15.16 3.8 15.74
2.07 18.31 4.5 19.13
2.11 19.55 1.6 19.86
2.19 22.27 3.1 22.96
2.27 25.3 1.2 25.6
2.34 28.66 4.2 29.86
2.42 32.36 3.6 33.52


Anyone who skims this see anything out of the ordinary? seems to be ok to me but i am a little worried about adding so much more fuel @ the lower end and getting a super rich condition. is a lower end correction that severe typical for the first round of rescaling of an SRI n/a? thanks for any input/comments.

I cant get the columns straight without making them worse... sorry.

[williaty]

Originally Posted by RaceFaceXC

I finally finished all the computing for my first stab at scaling the MAF and got

MAFv  g/sec  % cor  cor g/sec

 1.17	2.48	8.7	2.7
1.21	2.72	8.5	2.95
1.23	2.85	7.8	3.07
1.25	3.04	8	3.28
1.29	3.46	7.5	3.72
1.3	3.57	1	3.61
1.33	3.89	1	3.93
1.37	4.35	3.1	4.48
1.38	4.5	1	4.55
1.41	4.83	3.1	4.98
1.45	5.44	4.5	5.68
1.48	6.03	5	6.33
1.56	7.27	5	7.63
1.64	8.59	4.5	8.98
1.72	10	3.8	10.38
1.8	11.55	4.5	12.07
1.88	13.25	3.9	13.77
1.95	15.16	3.8	15.74
2.07	18.31	4.5	19.13
2.11	19.55	1.6	19.86
2.19	22.27	3.1	22.96
2.27	25.3	1.2	25.6
2.34	28.66	4.2	29.86
2.42	32.36	3.6	33.52

Anyone who skims this see anything out of the ordinary? seems to be ok to me but i am a little worried about adding so much more fuel @ the lower end and getting a super rich condition. is a lower end correction that severe typical for the first round of rescaling of an SRI n/a? thanks for any input/comments.

I cant get the columns straight without making them worse... sorry.

See, that's how you do it.

And look up a bit, I had to put the wammy on my scaling way more than you did.

[RaceFaceXC]

the wammy huh? you mean how you were up 22% at one point? I loved that gameshow. Big Money.. no wammy.. STOP!

[Rene2.5RS]

Sounds similar to this:

http://www.romraider.com/forum/topic1871.html

It is what I used and automated.

-Rene