My Flat Tappet Cam Wiped Out So I Decided To Go Roller

[Rustjunky67]

I like that blog. Think I'm gonna just use that Quaker state 5w30 full synthetic in my next build.

YEARS of testing with nothing to gain..... dang right i would take his opinion much more seriously than any person or company who stands to gain monetarily or otherwise from My decision. It's just like this forum, The guys who give sound advice don't stand to gain anything from it, they are more interested in helping others and spreading accurate information and knowledge

 

[L31MaxExpress]

Here is how to smooth out the VE learn from the EBL WUD. First I do the learn process with the EBL. Then I open the learned .bin file as a compare file. I copy the VE table from the learned .bin file to the current .bin file. I do not like the default color combination for the low/high values in Tunerpro. Makes the changed numbers in red difficult to see for me. First step is change the dumb color combination for me. I also remembered how to crop all the white area out of the image I wanted to post, so I fixed that as well. Hopefully these are easier to see what is going on.

Changed the default colors

You must be registered for see images attach

''

First step I use is to copy the newly learned VE table from the learned .bin file to the current one. The red numbers are the areas that the VE learning changed. The black numbers were not changed from lack of hits in those cells. Also notice how the area circled in orange has learned to a lower value than the unlearned area to the left of it at lower KPA. I used my VE table and your datalog for an example.

You must be registered for see images attach

First step I would use to smooth out this learned table is to multiply the far left table by 0.95 or reduce it 5%.

You must be registered for see images attach

Next I use horizontal interpolation to to smooth out the areas between the learned areas. I like smoothing between map points at the same RPM as it seems to give better results and ensures that the MAP values naturally increase as the KPA rises.

You must be registered for see images attach

This is what the finished product will look like after interpolating it.

You must be registered for see images attach

 

[L31MaxExpress]

Finally for the initial few learns, you need to grab the whole VE table and smooth it, using the smooth function. I like a 0.5 factor for the smoothing process to start with.

You must be registered for see images attach

Initial learned VE table

You must be registered for see images attach

Decel leaned out 5%, Interpolated and Smoothed VE table

You must be registered for see images attach

Follow the steps I have laid out, use the VE Learning, combined with interpolation and smoothing functions. Your VE tables will shape up much more quickly than trying to hand edit them or flat line them. When the fueling is closer I bet it runs much more smoothly. I would concentrate on getting the fueling correct rather than combating the surge for now. The surging will probably disappear on its own after several good learns and VE table smoothing operations as I have laid out for you in these posts.

 

[L31MaxExpress]

Finally lets say your VE table is twisted. It is 15% lean at low KPA and 5% rich at high KPA. How to easily straighten it out.

Multiply the Low KPA area by 1.15

You must be registered for see images attach

Then multiply the High KPA area by 0.95

You must be registered for see images attach

Then use Interpolate X function.

You must be registered for see images attach

Finally smooth the table

You must be registered for see images attach

This is the result of twisting the VE table compared to the previous table I posted.

You must be registered for see images attach

 

[L31MaxExpress]

I forgot to mention as well. Personally, I would pick one method of VE learning. Disable HWY mode until the VE and Timing tables are close to where they need to be. If a Wideband is available, I would be learning the VE tables off the wideband and ignoring closed loop. Disable closed loop, disable long term aka BLM fuel trims, then set the Initial BLM values to 128 Min/Max. Forget trying to use the slow narrow band sensor for dialing in the VE table. The wideband spits out data much more quickly than the narrow band, so use it!

 

[PlayingWithTBI]

Personally, I would pick one method of VE learning. Disable HWY mode until the VE and Timing tables are close to where they need to be. If a Wideband is available, I would be learning the VE tables off the wideband and ignoring closed loop. Disable closed loop, disable long term aka BLM fuel trims, then set the Initial BLM values to 128 Min/Max. Forget trying to use the slow narrow band sensor for dialing in the VE table. The wideband spits out data much more quickly than the narrow band, so use it!

Yeah, I'm using a WBO2 and disable those things. I'm wondering if, after adjusting my tables, when running normally, should I allow HWY until, I have more miles on it during break in. I was getting ~16 - 17 AFR with HWY. Maybe too lean for now, plus maybe more CL operation will help?

Thanks for all your posts, I've never used the "Interpolation" function before, something else to play with.

 

[L31MaxExpress]

Yeah, I'm using a WBO2 and disable those things. I'm wondering if, after adjusting my tables, when running normally, should I allow HWY until, I have more miles on it during break in. I was getting ~16 - 17 AFR with HWY. Maybe too lean for now, plus maybe more CL operation will help?

Thanks for all your posts, I've never used the "Interpolation" function before, something else to play with.

I would get the fueling and timing dialed in before enabling highway mode. Just one less thing to deal with while setting up the basics.

The interpolate function is a great help at filling in empty data spots.

I actually sent you a bin file with the low rpm fueling reduced and the heavy load timing reduced about 4* as well.

 

[PlayingWithTBI]

I actually sent you a bin file with the low rpm fueling reduced and the heavy load timing reduced about 4* as well.

Yes, I got that. I see you adjusted SA- Low and High Compensation too, thanks! I'll flash it and go for a drive tomorrow.

I would get the fueling and timing dialed in before enabling highway mode.

I'm going to do that, I was just wondering if HWY would be too lean while breaking in the engine.

 

[L31MaxExpress]

Yes, I got that. I see you adjusted SA- Low and High Compensation too, thanks! I'll flash it and go for a drive tomorrow.


I'm going to do that, I was just wondering if HWY would be too lean while breaking in the engine.

I would not want the added exhaust heat while breaking it in.

Also if the VE table is lean in areas, the lean cruise could make it even leaner.

I halved the base spark correction values. That was a pretty substantial timing swing.

Send me the datalog on that and I will look at it. If I remember correctly, double tap on the space bar if you find it doing anything weird at a specific point. I seem to remember the WUD program putting a marker in the datalog when the space bar is double tapped. You can put multiple markers in the same datalog, in the event that you run across surging or whatever at a specific point.

 

[L31MaxExpress]

One other thing, you want closed loop correction completely disabled while you are tuning the VE tables. Otherwise your closed loop function will be fighting your VE learning from the wideband. I shoot for my fueling being slightly on the rich side in open loop. Then when I enable closed loop, I let the ECM use the 02 sensor to pull the fuel it wants to run near stoich. If I see the engine running continously lean in closed loop, you can increase the 02 voltage switch point settings say 100mv and see where it gets you. Looking at averaged closed loop fueling my van was 2-3% lean in closed loop. I added 100mv and the closed loop function went to being 1-2% rich. My engine likes being on the richer side. My closed loop function pulls about 5% fuel on a normal day. I start slightly rich and let the PCM lean it back a bit.