Using ChrisB's Analyzer, my methodology.
07-15-2004, 10:01 AM
#1
Using ChrisB's Analyzer, my methodology. I'm sharing this for comments, corrections, or to help anyone who's doing VE tuning.
Car is an '82 Z28 with an '01 LS1 in it. TR 224 cam, aftermarket cats, big 3.5" exhaust...the engine's breathing characteristics are very much not stock.
I'd been using LTFT's and the Law of Averages to tune the stock VE table and gotten it about as good (LTFT's in the zero to slightly negative everywhere) as it will get. I started with a 5% reduction in the IFR table, then manually tuned the VE table based on LTFT data.
The car runs fine, throttle response feels crisp, but two things are nagging me:
1.) tailpipes stay sooty. My '02 Silverado is a light grey inside, the car tips are black black black.
2.) when transitioning from steady cruise to acceleration (non-PE), the exhaust smells rich stinky even thru the cats.
So I'm using ChrisB's log file analyzer to build my VE table, IFR table back to stock.
Step 1: I accumulated all my months of EFILive data into a single CSV file, removed all the rows where trim cell = 21 and all cells where TPS = 0. Got over 55,000 rows of useable data.
Step 2: I read this data into ChrisB's analyzer, set the target LTFT to zero and the scalar to 3000 as I read here. Unchecked the VE percentage box.
Step 3: Starting with a virgin copy of my PCM's bin, I imported the VE table where it had generated non-zero data and hand-smoothed it. I bumped those data by 103.58% as I read here, then hand-smoothed the boundaries between LFA-generated data and stock points.
NOTE: smoothing meant modifying the stock data points upward or downward to maintain a straight/smooth line down each RPM column, no modifications made to the LFA-generated data.
Observations:
1.) The resulting VE table it generated is very dissimilar to the hand-tuned VE table I have now...even with a 5% reduction in the IFR table "helping" the hand-tuned VE table add fuel. Much lower VE numbers everywhere.
2.) The resulting VE table is lower in the low RPM (1200-3600RPM), high vacuum areas than the stock VE table, and higher in the high rpm (> 3600rpm) areas as well as the low rpm, low vacuum areas (>50 KPa).
NOTE: This is what I'd expect to see with a cam that shifts the torque curve upward in the RPM range...they move the volumetric efficiency curve.
Next step is to upload this bin to the car and see how it responds next week. In the interim, I'd welcome any corrections or comments.
Car is an '82 Z28 with an '01 LS1 in it. TR 224 cam, aftermarket cats, big 3.5" exhaust...the engine's breathing characteristics are very much not stock.
I'd been using LTFT's and the Law of Averages to tune the stock VE table and gotten it about as good (LTFT's in the zero to slightly negative everywhere) as it will get. I started with a 5% reduction in the IFR table, then manually tuned the VE table based on LTFT data.
The car runs fine, throttle response feels crisp, but two things are nagging me:
1.) tailpipes stay sooty. My '02 Silverado is a light grey inside, the car tips are black black black.
2.) when transitioning from steady cruise to acceleration (non-PE), the exhaust smells rich stinky even thru the cats.
So I'm using ChrisB's log file analyzer to build my VE table, IFR table back to stock.
Step 1: I accumulated all my months of EFILive data into a single CSV file, removed all the rows where trim cell = 21 and all cells where TPS = 0. Got over 55,000 rows of useable data.
Step 2: I read this data into ChrisB's analyzer, set the target LTFT to zero and the scalar to 3000 as I read here. Unchecked the VE percentage box.
Step 3: Starting with a virgin copy of my PCM's bin, I imported the VE table where it had generated non-zero data and hand-smoothed it. I bumped those data by 103.58% as I read here, then hand-smoothed the boundaries between LFA-generated data and stock points.
NOTE: smoothing meant modifying the stock data points upward or downward to maintain a straight/smooth line down each RPM column, no modifications made to the LFA-generated data.
Observations:
1.) The resulting VE table it generated is very dissimilar to the hand-tuned VE table I have now...even with a 5% reduction in the IFR table "helping" the hand-tuned VE table add fuel. Much lower VE numbers everywhere.
2.) The resulting VE table is lower in the low RPM (1200-3600RPM), high vacuum areas than the stock VE table, and higher in the high rpm (> 3600rpm) areas as well as the low rpm, low vacuum areas (>50 KPa).
NOTE: This is what I'd expect to see with a cam that shifts the torque curve upward in the RPM range...they move the volumetric efficiency curve.
Next step is to upload this bin to the car and see how it responds next week. In the interim, I'd welcome any corrections or comments.
07-15-2004, 12:12 PM
#4
FormerVendor
iTrader: (1)
Join Date: Feb 2002
Location: Waldorf, MD
Posts: 3,059
Likes: 0
Received 0 Likes
on
0 Posts
After loging two short trips (going and comming) I see the program wants me to severly lean out the values that are allready in the table and this is throughout the rpm range. I can understand idle but cant understand anything after that since when the car comes on the cam the VE should improve and not get worse.
Isnt soot caused by incomplete combustion? :shrug:
Isnt soot caused by incomplete combustion? :shrug:
07-15-2004, 12:19 PM
#5
Originally Posted by HumpinSS
if tps =0 wouldnt that imply idling. Why did you remove those cells?
Edit
Target LTRIM doesnt work
Edit
Target LTRIM doesnt work
1.) MAF accuracy at low airflow is recorded here as not so good, better to use a WB-02 and tune in open-loop mode.
2.) Deceleration data are useless...the PCM kicks into fuel cutoff mode under defined conditions. Too much opportunity to introduce variation, and the data are all captured during a closed-throttle condition anyway. I want data when I'm making power, not coasting.
07-15-2004, 12:22 PM
#6
[QUOTE=HumpinSS]After loging two short trips (going and comming) I see the program wants me to severly lean out the values that are allready in the table and this is throughout the rpm range. I can understand idle but cant understand anything after that since when the car comes on the cam the VE should improve and not get worse.
QUOTE]
Just to make sure I understand what you're doing, you're saying ChrisB's program returned you lower VE numbers than what was in your stock VE table?
QUOTE]
Just to make sure I understand what you're doing, you're saying ChrisB's program returned you lower VE numbers than what was in your stock VE table?
07-15-2004, 12:55 PM
#7
FormerVendor
iTrader: (1)
Join Date: Feb 2002
Location: Waldorf, MD
Posts: 3,059
Likes: 0
Received 0 Likes
on
0 Posts
yes, i can send you dumps of what happened. But you wont get it till after 5:00 pm EST
Few examples to get you started
Stock VE table
4000 rpm all map values from 20-100
67,70,75,78,79,81,82,84,87
Calculated values from LFA
4000 20,60,80
36,67,77
Few examples to get you started
Stock VE table
4000 rpm all map values from 20-100
67,70,75,78,79,81,82,84,87
Calculated values from LFA
4000 20,60,80
36,67,77
Trending Topics
07-29-2004, 08:09 AM
#11
I hit a snag and lost ground on troubleshooting, but I have progress to report.
The VE table ChrisB's program generated, based on sensor outputs from my car, was wrong. The math he uses is empirical, and I double-checked my procedure to be sure I was using it right.
I regression tested the process by capturing data from my bone-stock '02 Silverado and repeating the process...the VE table ChrisB's program generated was nearly identical to the stock VE table. I knew at this point I had bad sensor data on the car.
The Intake Air Temp and MAF sensors are the two big contributors, so I went after them. IAT verified against my infrared pyrometer, so it was ruled out. The MAF had given me problems previously (posted here) so I replaced it with a new stock unit. This cleared up the nagging problem of my LTFT's being near-zero +/- 1%, but still blacking up my tailpipes and rear bumper with soot. LTFTs suddenly went negative to 5-12% as the PCM yanked fuel. Bear in mind, at this point I was running a bin with IFR reduced to 95% and a VE table that had been increased to varying extents across the board as I used LTFT data to tune by specific Fuel Trim Cell areas.
My MAF was the problem. It wasn't completely failed, but was under-reporting airflow enough to throw off the LTFTs and ChrisB's LFA math. This accounts for the P0101 errors I got previously. The powertrain came from a wrecked '01 Trans Am, and the MAF had some surface corrosion from environmental exposure.
Last step I did was roll back to an old bin with stock IFR, VE, spark, pretty much everything but my baseline mods {!CAGS, !COT, 3.73 gears, fan temps, etc.} A good 30min drive w/ data capture showed me at 0% +/- 1% LTFTs pretty much everywhere (only mod is a TR 224 cam). So I'm back to square one, all the data I captured to date using the old/bad MAF is worthless, but I'm making progress.
The VE table ChrisB's program generated, based on sensor outputs from my car, was wrong. The math he uses is empirical, and I double-checked my procedure to be sure I was using it right.
I regression tested the process by capturing data from my bone-stock '02 Silverado and repeating the process...the VE table ChrisB's program generated was nearly identical to the stock VE table. I knew at this point I had bad sensor data on the car.
The Intake Air Temp and MAF sensors are the two big contributors, so I went after them. IAT verified against my infrared pyrometer, so it was ruled out. The MAF had given me problems previously (posted here) so I replaced it with a new stock unit. This cleared up the nagging problem of my LTFT's being near-zero +/- 1%, but still blacking up my tailpipes and rear bumper with soot. LTFTs suddenly went negative to 5-12% as the PCM yanked fuel. Bear in mind, at this point I was running a bin with IFR reduced to 95% and a VE table that had been increased to varying extents across the board as I used LTFT data to tune by specific Fuel Trim Cell areas.
My MAF was the problem. It wasn't completely failed, but was under-reporting airflow enough to throw off the LTFTs and ChrisB's LFA math. This accounts for the P0101 errors I got previously. The powertrain came from a wrecked '01 Trans Am, and the MAF had some surface corrosion from environmental exposure.
Last step I did was roll back to an old bin with stock IFR, VE, spark, pretty much everything but my baseline mods {!CAGS, !COT, 3.73 gears, fan temps, etc.} A good 30min drive w/ data capture showed me at 0% +/- 1% LTFTs pretty much everywhere (only mod is a TR 224 cam). So I'm back to square one, all the data I captured to date using the old/bad MAF is worthless, but I'm making progress.
07-29-2004, 08:48 AM
#12
FormerVendor
iTrader: (1)
Join Date: Feb 2002
Location: Waldorf, MD
Posts: 3,059
Likes: 0
Received 0 Likes
on
0 Posts
Damn that is dirty. At least you figured out the problem. I also think i am having MAF problems and that is why I get messed up values when using LFA. No matter what i do to the ve table i am getting +LTRIMS. I swapped out my o2's last night and that helped bring them down a little but still hasnt corrected the issue. I may have to buy a MAF just to eliminate that sensor being a problem
07-29-2004, 12:26 PM
#13
TECH Addict
Join Date: Feb 2004
Posts: 2,826
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by crainholio
I hit a snag and lost ground on troubleshooting, but I have progress to report.
The VE table ChrisB's program generated, based on sensor outputs from my car, was wrong. The math he uses is empirical, and I double-checked my procedure to be sure I was using it right.
I regression tested the process by capturing data from my bone-stock '02 Silverado and repeating the process...the VE table ChrisB's program generated was nearly identical to the stock VE table. I knew at this point I had bad sensor data on the car.
The Intake Air Temp and MAF sensors are the two big contributors, so I went after them. IAT verified against my infrared pyrometer, so it was ruled out. The MAF had given me problems previously (posted here) so I replaced it with a new stock unit. This cleared up the nagging problem of my LTFT's being near-zero +/- 1%, but still blacking up my tailpipes and rear bumper with soot. LTFTs suddenly went negative to 5-12% as the PCM yanked fuel. Bear in mind, at this point I was running a bin with IFR reduced to 95% and a VE table that had been increased to varying extents across the board as I used LTFT data to tune by specific Fuel Trim Cell areas.
My MAF was the problem. It wasn't completely failed, but was under-reporting airflow enough to throw off the LTFTs and ChrisB's LFA math. This accounts for the P0101 errors I got previously. The powertrain came from a wrecked '01 Trans Am, and the MAF had some surface corrosion from environmental exposure.
Last step I did was roll back to an old bin with stock IFR, VE, spark, pretty much everything but my baseline mods {!CAGS, !COT, 3.73 gears, fan temps, etc.} A good 30min drive w/ data capture showed me at 0% +/- 1% LTFTs pretty much everywhere (only mod is a TR 224 cam). So I'm back to square one, all the data I captured to date using the old/bad MAF is worthless, but I'm making progress.
The VE table ChrisB's program generated, based on sensor outputs from my car, was wrong. The math he uses is empirical, and I double-checked my procedure to be sure I was using it right.
I regression tested the process by capturing data from my bone-stock '02 Silverado and repeating the process...the VE table ChrisB's program generated was nearly identical to the stock VE table. I knew at this point I had bad sensor data on the car.
The Intake Air Temp and MAF sensors are the two big contributors, so I went after them. IAT verified against my infrared pyrometer, so it was ruled out. The MAF had given me problems previously (posted here) so I replaced it with a new stock unit. This cleared up the nagging problem of my LTFT's being near-zero +/- 1%, but still blacking up my tailpipes and rear bumper with soot. LTFTs suddenly went negative to 5-12% as the PCM yanked fuel. Bear in mind, at this point I was running a bin with IFR reduced to 95% and a VE table that had been increased to varying extents across the board as I used LTFT data to tune by specific Fuel Trim Cell areas.
My MAF was the problem. It wasn't completely failed, but was under-reporting airflow enough to throw off the LTFTs and ChrisB's LFA math. This accounts for the P0101 errors I got previously. The powertrain came from a wrecked '01 Trans Am, and the MAF had some surface corrosion from environmental exposure.
Last step I did was roll back to an old bin with stock IFR, VE, spark, pretty much everything but my baseline mods {!CAGS, !COT, 3.73 gears, fan temps, etc.} A good 30min drive w/ data capture showed me at 0% +/- 1% LTFTs pretty much everywhere (only mod is a TR 224 cam). So I'm back to square one, all the data I captured to date using the old/bad MAF is worthless, but I'm making progress.
I have my stock tune, pulled the numbers with HPtuners. Then I pulled the numbers from a vehicle that is the same year stock tune, etc., with Edit and translated it to HPTuners. Needless to say the translation did not go so well. The numbers aren't right. Which sucks. They are off by like a hair at the bottom of the scale and a full afro on the top of the scale. Ugh. I hate VE.
07-29-2004, 03:37 PM
#14
FormerVendor
iTrader: (1)
Join Date: Feb 2002
Location: Waldorf, MD
Posts: 3,059
Likes: 0
Received 0 Likes
on
0 Posts
i thought the conversion what editve * 5.12/178.33
http://www.hptuners.com/forum/YaBB.p...8997;start=2#2
http://www.hptuners.com/forum/YaBB.p...9793;start=2#2
http://www.hptuners.com/forum/YaBB.p...8997;start=2#2
http://www.hptuners.com/forum/YaBB.p...9793;start=2#2
Last edited by HumpinSS; 07-29-2004 at 03:44 PM.
07-29-2004, 04:50 PM
#15
TECH Addict
Join Date: Feb 2004
Posts: 2,826
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by HumpinSS
i thought the conversion what editve * 5.12/178.33
http://www.hptuners.com/forum/YaBB.p...8997;start=2#2
http://www.hptuners.com/forum/YaBB.p...9793;start=2#2
http://www.hptuners.com/forum/YaBB.p...8997;start=2#2
http://www.hptuners.com/forum/YaBB.p...9793;start=2#2
07-29-2004, 04:52 PM
#16
TECH Addict
Join Date: Feb 2004
Posts: 2,826
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by Another_User
Yep. Pull a VE table up on Edit and convert it. Then compare it to one of the same type/year from HPTuners. I need to pull my formula up...I think what they said is actually slightly different...but I tried that one too.
CylinderVolume being (I would figure...) the calculated cylinder volume in HPTuners.
07-29-2004, 06:14 PM
#18
TECH Addict
Join Date: Feb 2004
Posts: 2,826
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by HumpinSS
.708 for the 346
07-29-2004, 08:22 PM
#19
Originally Posted by HumpinSS
After loging two short trips (going and comming) I see the program wants me to severly lean out the values that are allready in the table and this is throughout the rpm range. I can understand idle but cant understand anything after that since when the car comes on the cam the VE should improve and not get worse.
Isnt soot caused by incomplete combustion? :shrug:
Isnt soot caused by incomplete combustion? :shrug:
Guys - I calculated my VE table with the LFA.
Below 2400 RPM ALL New VE cells are less than Stock VE. At 2800 RPM to 7200 RPM -> if you draw a line from 70 kPa/2800 RPM to 25 kPA/7200 RPM you will divide my new VE table into values less than stock and values greater than stock VE. Above the Diagonal all values are less than Stock VE; below the Diagonal all values are greater than stock. FWIW.
joel 
