how "smooth" should a VE table look like?
01-29-2013, 12:07 PM
#21
TECH Enthusiast
Thread Starter
iTrader: (1)
Join Date: Jul 2010
Location: CroSSby, TexaSS
Posts: 554
Likes: 0
Received 0 Likes
on
0 Posts
okay i will do this and do some more datalogging as time allows. Thanks for all the help everyone, I just want my car to run right and give the engine what it needs
01-31-2013, 09:33 AM
#22
The pcm itself INTERPOLATES the cells so technically it will never by "spiky" in the eyes of the processor code. Even if you have a high cell and a low cell right next to it, it will just do a blend of the surrounding cells.
An example of this.
Cell values:...................;Pcm uses this.
10,10,10,10;..................;13,15,14,12
10,20,10,10;..................;15,16,15,13
10,10,10,10;..................;13,14,14,13
Of course this is a simple section with just a single cell change. but you get the idea. And it was an extreme example of showing double the number in one single cell, which wont happen in real life, but you get the idea.
When you have a large table with lots of data points, all cells and values are manipulating each other to achieve a valid signal.
So Choppy in the table isn't really "Choppy" But it could still be wrong.
In EFIlive there is a button you can press that will smooth the ve tables to get rid of the high points and bring the surrounding points back up to copy close to what the cpu is doing anyways. I use it often to get the ve table smooth if it is in a speed density build.
One thing I see a lot in data logging is, if you watch, sometimes if you don't fix the large spikes ever time, the spike will get higher and higher and higher as you go because it isn't the cell that is the issue, it is the surrounding ones.
An example of this.
Cell values:...................;Pcm uses this.
10,10,10,10;..................;13,15,14,12
10,20,10,10;..................;15,16,15,13
10,10,10,10;..................;13,14,14,13
Of course this is a simple section with just a single cell change. but you get the idea. And it was an extreme example of showing double the number in one single cell, which wont happen in real life, but you get the idea.
When you have a large table with lots of data points, all cells and values are manipulating each other to achieve a valid signal.
So Choppy in the table isn't really "Choppy" But it could still be wrong.
In EFIlive there is a button you can press that will smooth the ve tables to get rid of the high points and bring the surrounding points back up to copy close to what the cpu is doing anyways. I use it often to get the ve table smooth if it is in a speed density build.
One thing I see a lot in data logging is, if you watch, sometimes if you don't fix the large spikes ever time, the spike will get higher and higher and higher as you go because it isn't the cell that is the issue, it is the surrounding ones.
Last edited by wait4me; 01-31-2013 at 09:38 AM.
01-31-2013, 11:00 PM
#25
FormerVendor
Join Date: Nov 2012
Location: Dyno Tuning in KY
Posts: 390
Likes: 0
Received 0 Likes
on
0 Posts
I'm with Ed, set those codes at least while tuning to verify they are tripped. You can verify this via the trouble codes. Some OS's work fine with the codes set to 2-no MIL or 3-no error reported but some do not. Just part of the different OS's being a little different than one another so it's best while tuning to make sure the codes are tripped before calibrating the VE.
Use your 3d graph view and right click and drag to rotate and right&left click and slide to zoom in or out. It is a very useful tool to use to get your VE dialed in. The smooth functions really help too.
Use your 3d graph view and right click and drag to rotate and right&left click and slide to zoom in or out. It is a very useful tool to use to get your VE dialed in. The smooth functions really help too.
01-31-2013, 11:15 PM
#26
TECH Enthusiast
Thread Starter
iTrader: (1)
Join Date: Jul 2010
Location: CroSSby, TexaSS
Posts: 554
Likes: 0
Received 0 Likes
on
0 Posts
Guys i made the changes for the DTC's tonight before i headed off to college, the car went rich a bit, (13.0-5) at idle and 14.0-2's cruising etc. I guess i had some sort of modifier or something skewing my AFR error %, I haven't attempted the WOT area yet as i had to attempt to get my cruising cells back to 0-3's Here's two Data logs, One on the way to school and one on the way home
02-01-2013, 10:49 AM
#30
Ok. so, you can change your 02 switching points to be lower in the computer. Do that to the table, in EFIlive I know the table name, not sure in yours, but it will be a table that has like .400 .400 in columns. Change them until it gets to stoich on your wideband. "hopefully you don't have an exhaust leak near your afr sensor and it is calibrated correctly"
Then turn off long term trims, but leave on your strims. There should be a setting to turn off ltrims. But allow the car to still go into closed loop. Then log your strims on the car while you are driving with no maf. It will tell you how far off you are from what the computer wants pretty quick in all the cells as you drive. Then stop the car, and multiply and paste the difference on the strims onto your ve table in the idle and non full throttle areas. For example, if it is +3 in 1000 row and 65 kpa, then go to that section on the ve table and ADD 3 % to that exact cell. Then do the rest of them. In EFIlive, I set up maps that allow it to be a simple one click process including smoothing to make it a nice table. Not sure if you can do it in your software or not. But probably is. It makes it really quick. The key though is to make sure your base matches actual, which is what those switch point tables should do. If not those, then maybe the stoich ratio of your fuel may be off a bit, so the stoich ratio value might need tweaked just a tiny bit.
Then turn off long term trims, but leave on your strims. There should be a setting to turn off ltrims. But allow the car to still go into closed loop. Then log your strims on the car while you are driving with no maf. It will tell you how far off you are from what the computer wants pretty quick in all the cells as you drive. Then stop the car, and multiply and paste the difference on the strims onto your ve table in the idle and non full throttle areas. For example, if it is +3 in 1000 row and 65 kpa, then go to that section on the ve table and ADD 3 % to that exact cell. Then do the rest of them. In EFIlive, I set up maps that allow it to be a simple one click process including smoothing to make it a nice table. Not sure if you can do it in your software or not. But probably is. It makes it really quick. The key though is to make sure your base matches actual, which is what those switch point tables should do. If not those, then maybe the stoich ratio of your fuel may be off a bit, so the stoich ratio value might need tweaked just a tiny bit.
02-01-2013, 11:45 AM
#31
TECH Enthusiast
Thread Starter
iTrader: (1)
Join Date: Jul 2010
Location: CroSSby, TexaSS
Posts: 554
Likes: 0
Received 0 Likes
on
0 Posts
Ok. so, you can change your 02 switching points to be lower in the computer. Do that to the table, in EFIlive I know the table name, not sure in yours, but it will be a table that has like .400 .400 in columns. Change them until it gets to stoich on your wideband. "hopefully you don't have an exhaust leak near your afr sensor and it is calibrated correctly"
Then turn off long term trims, but leave on your strims. There should be a setting to turn off ltrims. But allow the car to still go into closed loop. Then log your strims on the car while you are driving with no maf. It will tell you how far off you are from what the computer wants pretty quick in all the cells as you drive. Then stop the car, and multiply and paste the difference on the strims onto your ve table in the idle and non full throttle areas. For example, if it is +3 in 1000 row and 65 kpa, then go to that section on the ve table and ADD 3 % to that exact cell. Then do the rest of them. In EFIlive, I set up maps that allow it to be a simple one click process including smoothing to make it a nice table. Not sure if you can do it in your software or not. But probably is. It makes it really quick. The key though is to make sure your base matches actual, which is what those switch point tables should do. If not those, then maybe the stoich ratio of your fuel may be off a bit, so the stoich ratio value might need tweaked just a tiny bit.
Then turn off long term trims, but leave on your strims. There should be a setting to turn off ltrims. But allow the car to still go into closed loop. Then log your strims on the car while you are driving with no maf. It will tell you how far off you are from what the computer wants pretty quick in all the cells as you drive. Then stop the car, and multiply and paste the difference on the strims onto your ve table in the idle and non full throttle areas. For example, if it is +3 in 1000 row and 65 kpa, then go to that section on the ve table and ADD 3 % to that exact cell. Then do the rest of them. In EFIlive, I set up maps that allow it to be a simple one click process including smoothing to make it a nice table. Not sure if you can do it in your software or not. But probably is. It makes it really quick. The key though is to make sure your base matches actual, which is what those switch point tables should do. If not those, then maybe the stoich ratio of your fuel may be off a bit, so the stoich ratio value might need tweaked just a tiny bit.
02-01-2013, 03:04 PM
#33
Correct. The setting to turn off ltrims should just be a single on off switch. You still want it to go into closed loop though. So leave those settings be. Just change the ltrim off. The o2 switch points table should be able to skew them to real vs commanded in the problem areas.
02-01-2013, 05:15 PM
#34
TECH Enthusiast
Thread Starter
iTrader: (1)
Join Date: Jul 2010
Location: CroSSby, TexaSS
Posts: 554
Likes: 0
Received 0 Likes
on
0 Posts
Correct. The setting to turn off ltrims should just be a single on off switch. You still want it to go into closed loop though. So leave those settings be. Just change the ltrim off. The o2 switch points table should be able to skew them to real vs commanded in the problem areas.
02-01-2013, 05:34 PM
#35
FormerVendor
Join Date: Nov 2012
Location: Dyno Tuning in KY
Posts: 390
Likes: 0
Received 0 Likes
on
0 Posts
Depending on year and OS for the LTFT switch... some older stuff only has enable ECT for LTFT. I normally set the LTFT enable ECT to something north of what it could ever see to prevent them from ever enabling as well as disable the switch if it is present. (These are just the observations I've made with HPT) EFILives mapped parameters may be different.
02-01-2013, 09:33 PM
#36
TECH Senior Member
Jesse is referring to the OL STFT Enable...
when enabled and in OL, it performs STFT trimming on any stoich cell in OLFA...
you can see this on the wideband, it looks like very tight short oscillations around stoich.
when enabled and in OL, it performs STFT trimming on any stoich cell in OLFA...
you can see this on the wideband, it looks like very tight short oscillations around stoich.
02-02-2013, 03:44 PM
#37
TECH Enthusiast
Thread Starter
iTrader: (1)
Join Date: Jul 2010
Location: CroSSby, TexaSS
Posts: 554
Likes: 0
Received 0 Likes
on
0 Posts
Well guys i haven't tried the STFT method yet but i went to the 1/8th mile last night in sd open loop and i went from 7.54 @ 93.26 1.76 60ft to 7.43 @ 93.91 1.72 60ft just by datalogging the runs and changing only the WOT areas, I'm starting to see where the VE graph needs to be raised now, I guess i can see the "trend" now more or less
