Tuning using LTFT
03-21-2007, 10:20 AM
#1
Staging Lane
Thread Starter
Join Date: Aug 2002
Location: Raleigh,NC
Posts: 86
Likes: 0
Received 0 Likes
on
0 Posts
Tuning using LTFT I have EFI live scan and tune.There is a map for LTFT with rpm vs map just like the VE table. Is there a way to copy and paste these vaules into the VE table-like you do with a ben factor using a wideband.For example say a particle cell has a LTFT of -4.6,how could i paste this to the VE table and reduce the VE number by 4.6%.
03-21-2007, 11:46 AM
#2
Fuel trim tuning takes a LONG time because you have to wait for full -relearns between applying changes. You can't just log the trims though, because you have the MAF in the equation also. You must fail out the MAF and verify SD operation to begin this excercise. Turn off fuel adders like COT, disable air pump, etc. Start up in SD and RESET trims (the current learned value reflects the MAF) and ride around for a LONG time until all logged stfts are pretty much 0. If, at that point, your logged ltfts are no more than 10% out, you apply them over your VE table. In HPT it's paste special %. Now you disable ltfts, reset trim learning, and repeat the process with the values from the stfts a few times until things are in-line. At this point, when the MAF is re-enabled it will screw the trims back up, so it's time to calibrate that against your proper VE table (or stay SD)....
I am typing this in a hurry at lunch; there is a writeup floating somewhere around here of the full process if you are unfamiliar.
I am typing this in a hurry at lunch; there is a writeup floating somewhere around here of the full process if you are unfamiliar.
03-21-2007, 03:17 PM
#3
TECH Addict
iTrader: (10)
Join Date: Jun 2004
Location: Commerce Twp, MI
Posts: 2,918
Likes: 0
Received 0 Likes
on
0 Posts
If you get a little creative with Excel, it's not that hard to tune via the fuel trims. I have a spreadsheet that I posted a while back that allows you to do this. The other option would be to create a custom .pid that you could log directly in the scanner. But, I've found copying/pasting to Excel a little easier.
As mentioned before, you will need to disable the MAF sensor, turn off DFCO, and copy the high octane spark table to the low octane table to get the VE table to respond appropriately. However, you will only be able to tune areas where the STFTs are still actively switching. If you give it enough gas to go into open loop, let off immediately as open loop tuning must be done with a WB.
The first pass through, you're going to want to make corrections based on LTFTs and STFTs. Then, adjust the VE, turn off the LTFTs, and flash the tune (or leave the LTFT's enabled). Then, you can either tune off of the STFTs alone or continue the LTFTs+STFTs tuning.
It really doesn't matter if the LTFT's have fully learned over 100 miles....I think that's a myth started in the sticky at the top of the forum. As long as you have the average of the total fuel trim corrections made to the final fueling calculation (ie avg LTFTs + avg STFTs), you'll be fine. Rinse and repeat the process several times to get it close and call it quits. Re-enable the MAF and restore the low octane spark table, DFCO, and LTFTs back to factory. Then, do the same thing you just did with the VE table to the MAF table.
As mentioned before, you will need to disable the MAF sensor, turn off DFCO, and copy the high octane spark table to the low octane table to get the VE table to respond appropriately. However, you will only be able to tune areas where the STFTs are still actively switching. If you give it enough gas to go into open loop, let off immediately as open loop tuning must be done with a WB.
The first pass through, you're going to want to make corrections based on LTFTs and STFTs. Then, adjust the VE, turn off the LTFTs, and flash the tune (or leave the LTFT's enabled). Then, you can either tune off of the STFTs alone or continue the LTFTs+STFTs tuning.
It really doesn't matter if the LTFT's have fully learned over 100 miles....I think that's a myth started in the sticky at the top of the forum. As long as you have the average of the total fuel trim corrections made to the final fueling calculation (ie avg LTFTs + avg STFTs), you'll be fine. Rinse and repeat the process several times to get it close and call it quits. Re-enable the MAF and restore the low octane spark table, DFCO, and LTFTs back to factory. Then, do the same thing you just did with the VE table to the MAF table.
03-21-2007, 05:33 PM
#4
I don't like it... If you keep rinsing and repeating you will find that you will never get it as close as with a WB. So basically, once you re-enable the MAF, since the error isn't all MAF table, now your adjustments made to it will bring trims back in line, but you aren't really calibrated, just cheating one set of errors against another until things mostly line up. I can't see how are an accurate MAF table can really be derived from this, but I guess it gets ya close. I've had reasonably good luck SD tuning like this, but still have a hard time with PE really obeying commanded.
What spreadsheet are you using? I use one to solve for secondary VE after the fact, but that's it... I go straight from histogram to VE.
What spreadsheet are you using? I use one to solve for secondary VE after the fact, but that's it... I go straight from histogram to VE.
03-21-2007, 06:02 PM
#5
eh, i've tuned cars entirely without WB and i've come within <7hp of what later i was able to correct on a dyno.
i'm not recommending it, just making a point that it's not impossible to come decently close. it does however take a cubic shitton of time more than WB.
i'm not recommending it, just making a point that it's not impossible to come decently close. it does however take a cubic shitton of time more than WB.
03-21-2007, 07:05 PM
#6
TECH Addict
iTrader: (10)
Join Date: Jun 2004
Location: Commerce Twp, MI
Posts: 2,918
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by Frost
I don't like it... If you keep rinsing and repeating you will find that you will never get it as close as with a WB. So basically, once you re-enable the MAF, since the error isn't all MAF table, now your adjustments made to it will bring trims back in line, but you aren't really calibrated, just cheating one set of errors against another until things mostly line up. I can't see how are an accurate MAF table can really be derived from this, but I guess it gets ya close. I've had reasonably good luck SD tuning like this, but still have a hard time with PE really obeying commanded.
What spreadsheet are you using? I use one to solve for secondary VE after the fact, but that's it... I go straight from histogram to VE.
What spreadsheet are you using? I use one to solve for secondary VE after the fact, but that's it... I go straight from histogram to VE.
I made my own spread sheet. It has some built in filters that let you
** "target" a certain LTFT% (say you wanted to shoot for -1.6%)
** apply only a percentage of the correction (25%, 30%, 60%, whatever)
** weight the log data (apply a greater percent correction for cells hit more often)
It's pretty sweet. But like Marcin said, it's time consuming.
03-21-2007, 08:36 PM
#7
Staging Lane
Join Date: Jan 2004
Location: Calgary, Alberta, Canada
Posts: 84
Likes: 0
Received 0 Likes
on
0 Posts
[QUOTE=SSpdDmon]I got my '02 Z within +/- 1~2% with the NBO2's. /QUOTE]
... and when you checked that tune with a wideband, was the VE still within +/- 1~2%?
... and when you checked that tune with a wideband, was the VE still within +/- 1~2%?
Trending Topics
03-22-2007, 08:50 AM
#8
TECH Addict
iTrader: (10)
Join Date: Jun 2004
Location: Commerce Twp, MI
Posts: 2,918
Likes: 0
Received 0 Likes
on
0 Posts
[QUOTE=Billf6531] Yup. It's all about getting quality data. Filter your logs appropriately and you'd be surprised what you can do. 
Originally Posted by SSpdDmon
I got my '02 Z within +/- 1~2% with the NBO2's. /QUOTE]
... and when you checked that tune with a wideband, was the VE still within +/- 1~2%?
... and when you checked that tune with a wideband, was the VE still within +/- 1~2%?
03-24-2007, 02:02 AM
#9
I set up some pids and tables in efilive that just use strims. Like a BEN filter without the wideband. With the pcm calibration you just shut off ltrims, then allow the strim system alot faster control. With the Roadrunner pcm, it tunes itself completly within minutes in the non wot areas. Then i just verify wot with the wideband on the other channel and the tunes are within 1% everywhere. Then turn on ltrims and send them on thier way....
03-24-2007, 08:51 AM
#10
Moderator
iTrader: (11)
Join Date: Mar 2002
Location: East Central Florida
Posts: 12,605
Likes: 0
Received 6 Likes
on
6 Posts
On some vehicles the FTC bins are really too coarse to be
very useful, cover too much of a changing space and tend
to drift with driving as a result. It's hard to imagine that a
0-2500RPM span, for example, is useful for adjusting VE at
800 and at 2400 (by the same single amount).
Tightening it up and making all the cells used is a freebie
of course and can make this more accurate. But still I prefer
working off the real-time NBO2 voltages rather than a rolling
average of yesterday's news.
very useful, cover too much of a changing space and tend
to drift with driving as a result. It's hard to imagine that a
0-2500RPM span, for example, is useful for adjusting VE at
800 and at 2400 (by the same single amount).
Tightening it up and making all the cells used is a freebie
of course and can make this more accurate. But still I prefer
working off the real-time NBO2 voltages rather than a rolling
average of yesterday's news.
03-24-2007, 12:15 PM
#11
Since there are only 22(?) fuel trim cells, and only 16 for actual closed loop driving, each LTFT cell will cover many VE table cells. (Plus if you haven't changed the default ftc boundries, you've basically got one ftc cell for all closed loop driving.) So using LTFT will be very coarse at best, might make some VE cell values worse. Using ST fuel trims should work a lot better, just use the average STFT within each VE cell.
