Finessing a VVE table?

[JAY4SPEED]

In Simon's (Swingtan's) VVE tutorial he touched a bit on the subject of finessing a VVE table since it behaves a little differently than a typical VE table after you generate the coefficients and generating the VVE table. If I understood it correctly, He mentioned to add 4% for course adjustments and 2% to surrounding cells on peaks and valleys after using the paste and multiply function to bring up surrounding cells to have a greater impact on the areas you are trying to correct. How far should I go into the neighboring cells to add the 4% or 2% compensation?

I've pretty much realized that the VVE table will be extremely hard to dial it in exactly since it moves around so much (far reaching changes in zones you didn't intend to change) every time you paste and multiply BEN factors and every time you generate Coefficients and generate a VVE. Are there any other tips or words of wisdom that could be shared to help "finesse" the VVE table into where the BENs want it to be? I've played around a little bit with the Zone edge align and saw that the higher the number, the deeper into the neighboring zones become affected. Is there an ideal setting for Zone edge align?

Thanks Fellas!

Jay

[swingtan]

Hi Jay,
for starters, how long are the logs you have that you are using to correct the VVE? If you are road tuning, then I'd suggest you try doing long drives and applying the data after that. I've tried applying data from a 2.5 hour trip and the resultant VVE was very, very close.

Also, don't forget to try smoothing the spikes by selecting a group of cells around a spike ( or dip ) and averaging the values. Also, take note of the areas that the VVE is already very close and don't apply any BEN data for those areas. Only work on the areas that need adjusting.

Simon.

[GMPX]

Don't forget, before you start fine tuning the VVE to change your MAP zones to match the engine. In other words if you have a decent cam then there is not much point wasting all the cells below 35kPa, instead if you play around with the MAP zones you will give yourself more accurate results in the area's that matter.
This is particularly true for boosted engines, you really should extend the end MAP zones in to boost.
Keep in mind though that you must make these changes early in the tuning process as it makes a dramatic change to the VVE.

Cheers,
Ross

[JAY4SPEED]

Hi Jay,
for starters, how long are the logs you have that you are using to correct the VVE? If you are road tuning, then I'd suggest you try doing long drives and applying the data after that. I've tried applying data from a 2.5 hour trip and the resultant VVE was very, very close.

Thanks so much for the reply Simon. I was hoping you'd chime in. I'm using LTFT bens with both banks averaged together as the data source of my BEN factor, and I'm in closed loop speed density right now. I have had pretty good success on longer drive cycles getting consistent data. I've logged 120 miles at a time on a road trip and it got me very close to where I want to be. Mostly my drive cycles are around 15 miles (24 km) on the highway to and 15 miles from work on the return trip. I feel the data I'm getting is good data because my LTFT are getting very close to 0. I feel my data and my data filter is on the right track getting me the data I need. The questions I have are when you actually apply the data to the VVE table.


Also, don't forget to try smoothing the spikes by selecting a group of cells around a spike ( or dip ) and averaging the values. Also, take note of the areas that the VVE is already very close and don't apply any BEN data for those areas. Only work on the areas that need adjusting.

Simon.

Right, this is exactly the info I was looking for and had questions about. I'm not totally clear as to the process of selecting the surrounding cells around a spike or dip. So, you basically raise the of all the cells circling a spike for example by 4% (for a course adjustment). If I'm understanding it correctly, that's like basically exaggerating a spike by giving it more surface area, to have a greater influence once the coefficients and vve is generated, correct? I must have missed that part of the tutorial where you average the spike with the new higher surrounding cells. So after you make the spike's surface area bigger, you smooth the area as well before generating coefficients and vve?

Thanks for being patient with me, I am grasping this slowly but surely..:grin:


Don't forget, before you start fine tuning the VVE to change your MAP zones to match the engine. In other words if you have a decent cam then there is not much point wasting all the cells below 35kPa, instead if you play around with the MAP zones you will give yourself more accurate results in the area's that matter.
This is particularly true for boosted engines, you really should extend the end MAP zones in to boost.
Keep in mind though that you must make these changes early in the tuning process as it makes a dramatic change to the VVE.

Cheers,
Ross

Thanks Ross, I had Simon help me with my map and rpm zones a while back, so they are set up where I idle and cruise right in the middle of vve zones. I'm have stock LS2 (airflow wise) in a Trailblazer SS. I wanted to get the concepts down before I tried it with a cam swap. BTW I'm finding the factory VVE table to be out about 5 to 10% in most areas (but that may have been affected when I rezoned my VVE table)

This brings me to another question. If I were to change the map / rpm zones of the VVE table, would I have to start auto VVE from scratch throwing out all of my data and starting over? Or can i copy the table, make the zone changes, paste the vve values back in, close and restart the software (the same method as the setup procedure the AUTO VVE tutorial?

Thanks Guys for all the help!:cheers:

Jay

[swingtan]

Points as follows....


I think that 15M is not long enough to get a good representation of the cars fueling needs. It probably only relates to a 20minute trip and the first 5 min of that is the engine warming up. The oil may take another 5 min to really come up to temp so you only really have a short "usable" log to work with. You will be tuning the VVE for the short trips and not a full running temp environment. Stick with trips of at least 60 min and remove the data where the IVT is under 130'C.

With the spikes, you are getting close. The longer trips will help to reduce the size of the spikes as data will be averaged over more data cells. Also, don't forget to set up your filters to exclude data that really shouldn't be included. I filter on....




GM.ETCTP % is changing less than 0.50/100mS
SAE.ECT °C is greater than 85.00
SAE.RPM, is greater than 600.00
EXT.WO2AFR1 is less than or equal to 16.50
E38.INVLVTMP_DMA °C,greater than or equal to 133.50
GM.APP %,greater than 1.00
EXT.WO2AFR1 changing less than 0.10/100mS


This should reduce the transient data a lot and give BEN data for steady state running. Note that this filter will allow you to leave DFCO turned on as it simply filter out the data when it's active.

If you change the MAP / RPM boundaries you do not have to start from scratch. Just remember to copy and paste the VVE cells after re-opening the file so the values are placed over the new boundaries.


Simon

[JAY4SPEED]

Thanks Simon! I will give your filter set a try on my next road trip.:cheers: