Recalibrating MAF transfer curve
#1
Staging Lane
Thread Starter
iTrader: (1)
Join Date: Nov 2002
Location: Louisville, KY
Posts: 59
Likes: 0
Received 0 Likes
on
0 Posts
![Default](https://ls1tech.com/forums/images/icons/icon1.gif)
AS to not beat the subject...
To recalibrate the STOCK MAF Table (I have ported end, CAI, Lid) would I just do some partial throttle runs all the way through the RPM range and based off the recorder MAF frequency, would i adjust that corresponding air flow(g/s) only, in order to adjust that moment in times LTFT up/down.
So if for example at some point in time my maf frequency is X and my LTFT are say +10% would i scale that indivdual cell (X) on the maf table by 110% to recieve a new calculated value for that frequency only. do that for say every maf frequency say in 100(hz) steps or equiv, and layout those new points on a graph. Then do a curve regression to best fit a smooth curve through thes new calculated points.
From this new smooth regressed curve update each cell in the MAF table to the new value.
IS this correct? I know is it a lot of work, but is is the right way to do it with running ported maf ends, lid and CAI? <img border="0" title="" alt="[Smile]" src="gr_stretch.gif" />
To recalibrate the STOCK MAF Table (I have ported end, CAI, Lid) would I just do some partial throttle runs all the way through the RPM range and based off the recorder MAF frequency, would i adjust that corresponding air flow(g/s) only, in order to adjust that moment in times LTFT up/down.
So if for example at some point in time my maf frequency is X and my LTFT are say +10% would i scale that indivdual cell (X) on the maf table by 110% to recieve a new calculated value for that frequency only. do that for say every maf frequency say in 100(hz) steps or equiv, and layout those new points on a graph. Then do a curve regression to best fit a smooth curve through thes new calculated points.
From this new smooth regressed curve update each cell in the MAF table to the new value.
IS this correct? I know is it a lot of work, but is is the right way to do it with running ported maf ends, lid and CAI? <img border="0" title="" alt="[Smile]" src="gr_stretch.gif" />
#2
TECH Regular
iTrader: (1)
Join Date: Aug 2002
Location: Scottsdale, Az
Posts: 429
Likes: 0
Received 0 Likes
on
0 Posts
![](https://ls1tech.com/forums/images/ranks/ls1tech10year.png)
![Default](https://ls1tech.com/forums/images/icons/icon1.gif)
That is precisely how you tune speed density. Since each load/rpm cell is a look up value each individual cell MUST be accurate, so you do as you suggest. In fact Joe Georger has written a rather elegant program to take TTS Datamaster logging data and correlate it to individual load/rpm values and automatically calculate new values for each cell and output a new .bin file with corrected VE tables.
Now MAF is more forgiving since a change in the motor typically will flow more or less air and that new airflow is measured not looked up. So it can handle changes better. But to get spot on tuning than the proper procedure should be to do it in individual cells. However I get the feeling that most people don't bother. That may be because with a MAF system you don't need to or simply that folks have gotten lazy due to the flexibility of the system.
My LS1Edit is on order and I will know more after my h/c/lt install on the 98. But on speed density we call the injector constant the "big ****" and the VE table the "fine tuning". You start out after a motor change by playing with the inj constant. But at some pt you must get each cell correct and you also want to work your way back to an absolute and a correct inj constant.
Now MAF is more forgiving since a change in the motor typically will flow more or less air and that new airflow is measured not looked up. So it can handle changes better. But to get spot on tuning than the proper procedure should be to do it in individual cells. However I get the feeling that most people don't bother. That may be because with a MAF system you don't need to or simply that folks have gotten lazy due to the flexibility of the system.
My LS1Edit is on order and I will know more after my h/c/lt install on the 98. But on speed density we call the injector constant the "big ****" and the VE table the "fine tuning". You start out after a motor change by playing with the inj constant. But at some pt you must get each cell correct and you also want to work your way back to an absolute and a correct inj constant.
#3
Staging Lane
Join Date: Feb 2002
Location: Newport, Michigan
Posts: 97
Likes: 0
Received 0 Likes
on
0 Posts
![Default](https://ls1tech.com/forums/images/icons/icon1.gif)
Chris,
Basically, what you said is the way I went about it (modifing the MAF table based on the LTFT values). However, I did 1/2 of the change on my first change. Your example of 110% (LTFT +10%), I would multiply by 105%, and then go out allow the PCM to learn, then take more scans. BUT, you can not make one multiplication thru the entire MAF table and think that you got it right, because you do not!
Concerning the scans: I try to visit as many fuel cells as possible. One type of scanning I conduct is to get on an Xway, short shift each gear up to 4th (I have an M6 car), then while I'm in 4th at a very low speed slowly accelerate, take the car speed up to what is responsible under your condition, staying in 4th gear. (For an A4 car, I would put the tranny in Drive, not overdrive.)
I also take scans at several steady state conditions, the first is idle (when I return from the street scanning session) and others at like 25mph, 45mpg, 65mph (whatever)
Now I got the scans for analysis. I do record MAF frequency, Engine RPM, LTFT, never STLT, fuel cell #, timing,,,,.
After analyzing the scans, I locate the different portions of the MAF table for changes. Typically, I break the table down into the idle and just off idle region (1500-3000Hz), mid-range region (5000-7000Hz), and upper end region (8000Hz and above).
Note that there are gaps. I average the LTFTs for these respective regions and multiply by 1/2 of the LTFT avg value for that region. (I use 1/2 the value because I do not want to make to big of a change at a time and cause other problems. So if the LTFT value is +15%, I would multiply by 107.5% for that region.) In the gap areas of the MAF table, I prorate the change to the flow rate so that I maintain a smooth looking curve. The MAF table is a non-linear curve which gets more steep at higher frequencies. Always check your work in the graphics mode in LS1Edit and make your changes to a new file name so that if you have to you can always go back to the original MAF table values. Also, if you saved the stock PCM stuff, and left it as is, their is a compare file option in LS1Edit which is nice.
The objective is to get all of the LTFT values between 0% and -5%. The -5% is a little arbitary.
If you re-calibrate the MAF table for a ported MAF sensor, it will take many scan/LS1Edit cycles. If you get it done properly in 5 cycles, you are doing great, IMHO. I took about 8 to 10 cycles mainly because this was my first tune.
In some other post, a tuner (who does not like this method) stated that an LS1Edit user changed the entire MAF table by 120% and then had drivibility problems. Well, DAH, no wonder, 120% is too much. With my ported MAF installed, I scanned LTFTs at 18% up to 22%. In order to re-calibrate the MAF table, no cell was changed more than 15%. So I conclude that the change to the MAF table and the LTFTs are not a one to one correlation.
Basically, what you said is the way I went about it (modifing the MAF table based on the LTFT values). However, I did 1/2 of the change on my first change. Your example of 110% (LTFT +10%), I would multiply by 105%, and then go out allow the PCM to learn, then take more scans. BUT, you can not make one multiplication thru the entire MAF table and think that you got it right, because you do not!
Concerning the scans: I try to visit as many fuel cells as possible. One type of scanning I conduct is to get on an Xway, short shift each gear up to 4th (I have an M6 car), then while I'm in 4th at a very low speed slowly accelerate, take the car speed up to what is responsible under your condition, staying in 4th gear. (For an A4 car, I would put the tranny in Drive, not overdrive.)
I also take scans at several steady state conditions, the first is idle (when I return from the street scanning session) and others at like 25mph, 45mpg, 65mph (whatever)
Now I got the scans for analysis. I do record MAF frequency, Engine RPM, LTFT, never STLT, fuel cell #, timing,,,,.
After analyzing the scans, I locate the different portions of the MAF table for changes. Typically, I break the table down into the idle and just off idle region (1500-3000Hz), mid-range region (5000-7000Hz), and upper end region (8000Hz and above).
Note that there are gaps. I average the LTFTs for these respective regions and multiply by 1/2 of the LTFT avg value for that region. (I use 1/2 the value because I do not want to make to big of a change at a time and cause other problems. So if the LTFT value is +15%, I would multiply by 107.5% for that region.) In the gap areas of the MAF table, I prorate the change to the flow rate so that I maintain a smooth looking curve. The MAF table is a non-linear curve which gets more steep at higher frequencies. Always check your work in the graphics mode in LS1Edit and make your changes to a new file name so that if you have to you can always go back to the original MAF table values. Also, if you saved the stock PCM stuff, and left it as is, their is a compare file option in LS1Edit which is nice.
The objective is to get all of the LTFT values between 0% and -5%. The -5% is a little arbitary.
If you re-calibrate the MAF table for a ported MAF sensor, it will take many scan/LS1Edit cycles. If you get it done properly in 5 cycles, you are doing great, IMHO. I took about 8 to 10 cycles mainly because this was my first tune.
In some other post, a tuner (who does not like this method) stated that an LS1Edit user changed the entire MAF table by 120% and then had drivibility problems. Well, DAH, no wonder, 120% is too much. With my ported MAF installed, I scanned LTFTs at 18% up to 22%. In order to re-calibrate the MAF table, no cell was changed more than 15%. So I conclude that the change to the MAF table and the LTFTs are not a one to one correlation.
#4
Staging Lane
Thread Starter
iTrader: (1)
Join Date: Nov 2002
Location: Louisville, KY
Posts: 59
Likes: 0
Received 0 Likes
on
0 Posts
![Default](https://ls1tech.com/forums/images/icons/icon1.gif)
Thanks guys for all the info.
The first non rainy dry day I get I will be working on this.
I did scale the entire table at first upto 10%, Which helped across the board, but I needed more and it threw off the high and low end quite a bit.
So i knew i had to redo the curve in pieces. <img border="0" alt="[driving]" title="" src="graemlins/gr_driving3.gif" />
The first non rainy dry day I get I will be working on this.
I did scale the entire table at first upto 10%, Which helped across the board, but I needed more and it threw off the high and low end quite a bit.
So i knew i had to redo the curve in pieces. <img border="0" alt="[driving]" title="" src="graemlins/gr_driving3.gif" />