Deviation in logs for LTFT's
Thread Starter
TECH Regular
Joined: Dec 2005
Posts: 437
Likes: 9
From: Indianapolis
Deviation in logs for LTFT's I just logged two different runs on my way to work accidentally split them into two different logs when I hit the space to keep the laptop from sleeping. The issue is the begining of my log is showing that I need to add fueling to the VE cells on most of the table while the other log I have is showing removing fuel in most of the VE tables. Now the first one I started recording when the car was almost up to temp while the second run was on the highway going at speed already. Which one will be more accurate?
Just to give a little more info on the logs I was recording on my way to work which is 95% highway constant speed logging (I fill most of the tables with low rpm and high MAP by throttling back and forth in the different gears). Is this correct or should I change my logging conditions to make them more accurate?
Just to give a little more info on the logs I was recording on my way to work which is 95% highway constant speed logging (I fill most of the tables with low rpm and high MAP by throttling back and forth in the different gears). Is this correct or should I change my logging conditions to make them more accurate?
Did you log the FTCs? If you ran on the xway at nearly the same speed with little deviation, you might have been mainly in one FTC. The 1st log, I guess, was in several FTCs. I would say that both are correct for the FTCs that each log was in.
Thread Starter
TECH Regular
Joined: Dec 2005
Posts: 437
Likes: 9
From: Indianapolis
I try to cover all the FTCs in HP Tuners less then 4k RPM. The problem was that the same cell in the table on the first run was showing lean then that same cell on the second scan was showing rich. I'm not sure which to make my changes to the VE table with.
Well, maybe it is simply the learning business in the PCM. You stated that the engine was at first coming up to temperature. I have had changes like this. I always took the most recent session as Gods true. If there are any doubts like you have, simply continue logging until you are satisfied.
It seems to me that the accurate overall LTFT's would be the average of the two. I'm assuming you reset the LTFT's before you started logging.
I would copy the histograms into Excel or another spreadsheet program. Copy the actual percents of each histogram into 2 seperate charts (lets call them A and B) and then copy the cell counts of each histogram into 2 more seperate charts (lets call these Ac and Bc). Multiply the percent by the count for each cell of the two histograms by using formulas (A*Ac and B*Bc). Then add both new histograms (A*Ac + B*Bc) and divide by the total number of counts for each cell (Ac + Bc). This is way harder to explain in words than to actually do. Should take about 30 seconds from start to finish.
I would copy the histograms into Excel or another spreadsheet program. Copy the actual percents of each histogram into 2 seperate charts (lets call them A and B) and then copy the cell counts of each histogram into 2 more seperate charts (lets call these Ac and Bc). Multiply the percent by the count for each cell of the two histograms by using formulas (A*Ac and B*Bc). Then add both new histograms (A*Ac + B*Bc) and divide by the total number of counts for each cell (Ac + Bc). This is way harder to explain in words than to actually do. Should take about 30 seconds from start to finish.
