VVE table after calculating coefficients

[Dadrunr1726]

I'm in the process of tuning the VVE and MAF on my 2015 Silverado L83 using the SilverSurfer's VVE assistant tool and Dyn Air calculations. I believe I have everything working as it should however my VVE table goes to shit once I hit calculate coefficients after bringing my data in from the assistant tool. I can interpolate the extreme negative values afterwards but I'm getting data in those areas so I'm not confident in the values once they are smoothed. I know this issue has been brought up a few times, but I wanted to see if anyone has found a way around this or what you guys recommend doing. Thank you


VVE calc coef.png

[tonn0297]

you need to smooth it some more, if theres a too large spread in the data points it shits the bed and will do what you show. So be sure to smooth it as much as possible within a "box". HPT do not care to add some error handling for it, so trial and error.

[Alvin]

I don't use that tool but typically when I see that I adjust boundaries to fix it.

I could fix the file for you if you post it.

[Dadrunr1726]

I don't use that tool but typically when I see that I adjust boundaries to fix it.

I could fix the file for you if you post it.

Okay thank you. Here is the tune file with the error values entered (multiplied by %-1/2), smoothed and coefficients calculated in to the Manifold Switch Open VE table. Problem is I plan on doing a few drives to get it dialed in so I'm thinking it will keep doing this unless that area of the table is addressed.

2015_Silverado_MAF_VVE_CAL-6_TEST.hpt

[edcmat-l1]

Did you try interpolating between the normal columns and then calculating the coefficients?

[Alvin]

Here you go.

2015_Silverado_MAF_VVE_CAL-6_TEST - Alvin small edit.hpt

[Dadrunr1726]

Here you go.

2015_Silverado_MAF_VVE_CAL-6_TEST - Alvin small edit.hpt

Thank you Alvin. Is it okay if I email you to discuss your process?

[Slow6.2]

When you interpolate and smooth the data, try to stay within these boundaries before calculating. The VVE assistant tends to go catawampus if theres too much of a change over them. Usually what I do is use the manifold switch closed table to make changes and calculate first to make sure everything plays nice, that way if it goes goofy I can just copy and paste the open table into the closed table to start over.
GEN V VVE ZONES.jpg

[Cringer]

There are just some combinations of data as well as the size of the zone itself that make the regression algorithm flake out like this some times. It is just a scenario where the math breaks, and I have seen it happen in my tool and the HPT Editor. The easiest thing to do is interpolate across the top of the excessive high or low area and re-calc the coeff. That is usually enough to work around it and usually won't happen again.

[Alvin]

There are just some combinations of data as well as the size of the zone itself that make the regression algorithm flake out like this some times. It is just a scenario where the math breaks, and I have seen it happen in my tool and the HPT Editor. The easiest thing to do is interpolate across the top of the excessive high or low area and re-calc the coeff. That is usually enough to work around it and usually won't happen again.

Now that the boundaries are redone.. unless the shape changes a lot again it should stay pretty good to go.

[Cringer]

Now that the boundaries are redone.. unless the shape changes a lot again it should stay pretty good to go.

It should. I find the larger the zone, the more reliable the regression.

[Dadrunr1726]

Now that the boundaries are redone.. unless the shape changes a lot again it should stay pretty good to go.

Alvin- when I first opened the file you sent back, I could see the boundary zones and pressure ratio axis changes you made. For whatever reason, the pressure ratio has gone back to the factory scaling after opening my previous file, even when I try to open it again from here. Seems like when I first opened it the pressure ratio started at around .25 and went to 1.12. Now its back to .150 to .930 but the boundary zones are still modified.

[Cringer]

Alvin- when I first opened the file you sent back, I could see the boundary zones and pressure ratio axis changes you made. For whatever reason, the pressure ratio has gone back to the factory scaling after opening my previous file, even when I try to open it again from here. Seems like when I first opened it the pressure ratio started at around .25 and went to 1.12. Now its back to .150 to .930 but the boundary zones are still modified.

HPT does not associate the RPM or MAP (PR) row and column breakpoints to the tune file so they are not transferable between different laptops or users. So if you manually change the breakpoints, then you have to manually change it back on your machine.

[Dadrunr1726]

HPT does not associate the RPM or MAP (PR) row and column breakpoints to the tune file so they are not transferable between different laptops or users. So if you manually change the breakpoints, then you have to manually change it back on your machine.

Okay that 's what I figured. Just strange how I could see the MAP (PR) row edits the first time I opened his file and now its back to the originally scaling but the boundary zones are still modified. I don't recall changing any of the breakpoints or atleast not intentionally.

[Abilor]

There are just some combinations of data as well as the size of the zone itself that make the regression algorithm flake out like this some times. It is just a scenario where the math breaks, and I have seen it happen in my tool and the HPT Editor. The easiest thing to do is interpolate across the top of the excessive high or low area and re-calc the coeff. That is usually enough to work around it and usually won't happen again.

First off, thank god for the tool you developed. I've spent all evening learning to use it, and smoothing my tables. I find that whereas MAF is very straightforward, especially with Dynamic Air calculus, the VVE remains mysterious. I did a long Speed Density run tonight using the Dynamic Air calcs, and noticed that it improved *slightly*. The wideband confirmed things are a lot more dialed in. As I sit here typing though, it occurs to me in my haste that I didn't copy the adjusted Closed Manifold table over to the Open Manifold table; I'm going to re-run with the same tune tomorrow now that I've cloned the tables (is this desirable?).

In the meantime, my VVE looks like a Klingon Ski Slope:

VVE_03_25_2025.png

Does this look insane? How many runs does it take to really dial this in? When it is dialing in, should I see these values moving closer to 0 over time? If so, how quickly? MAF changes seem linear and immediate; VVE changes feel like wishes and dreams.

SD_Logs.png

Also, what conditions do you set for SD Tuning aside from the Dynamic Air calcs and filters? Turn off DFCO, Cat Protection, disable MAF only Air Calc mode? Leave timing stock?

I've never met anything like the VVE man, can't tell if it's a bona fide PITA or I'm still inept or both!

Thanks!

[Cringer]

First off, thank god for the tool you developed. I've spent all evening learning to use it, and smoothing my tables. I find that whereas MAF is very straightforward, especially with Dynamic Air calculus, the VVE remains mysterious. I did a long Speed Density run tonight using the Dynamic Air calcs, and noticed that it improved *slightly*. The wideband confirmed things are a lot more dialed in. As I sit here typing though, it occurs to me in my haste that I didn't copy the adjusted Closed Manifold table over to the Open Manifold table; I'm going to re-run with the same tune tomorrow now that I've cloned the tables (is this desirable?).

In the meantime, my VVE looks like a Klingon Ski Slope:

VVE_03_25_2025.png

Does this look insane? How many runs does it take to really dial this in? When it is dialing in, should I see these values moving closer to 0 over time? If so, how quickly? MAF changes seem linear and immediate; VVE changes feel like wishes and dreams.

SD_Logs.png

Also, what conditions do you set for SD Tuning aside from the Dynamic Air calcs and filters? Turn off DFCO, Cat Protection, disable MAF only Air Calc mode? Leave timing stock?

I've never met anything like the VVE man, can't tell if it's a bona fide PITA or I'm still inept or both!

Thanks!

You have begun your quest! Stick with it, you will figure it out. Glad I can help along the way.

That VVE shape still needs a lot of work. It is important to remember that you are not creating a VE shape, you are creating the shape for each zone...or maybe tuning 30 different VE tables and gluing them together to form one larger shape. I made this video that explains how the nuts and bolts of VVE are implemented, it should help you understand the bigger picture (https://youtu.be/pZ7i9c0sCN8) and thus make the concept of VVE and how to work with it much easier.

I do not go into SD (or OL) to tune VVE. I have developed my own method and formulas to tune MAF and VVE at the same time during normal hybrid and closed loop driving. I initially published them here (https://youtu.be/xRZaTblSQ_0) forgive the audio quality as this was one of my earlier videos.

I disable Desoot and COT (for WOT tuning), LTFT, DFCO and CFCO (for CL tuning). And when I drive I use the lightest and steadiest of throttle inputs to as to now invoke transient fuel. Then for everything else I use filters in the scanner to remove all other transients (some of the filters need to be tweaked a little for Gen5 stuff and I think GHuggins uses a tweaked version of the user defined math for Gen5 too). But this will get you going.

[Alvin]

First off, thank god for the tool you developed. I've spent all evening learning to use it, and smoothing my tables. I find that whereas MAF is very straightforward, especially with Dynamic Air calculus, the VVE remains mysterious. I did a long Speed Density run tonight using the Dynamic Air calcs, and noticed that it improved *slightly*. The wideband confirmed things are a lot more dialed in. As I sit here typing though, it occurs to me in my haste that I didn't copy the adjusted Closed Manifold table over to the Open Manifold table; I'm going to re-run with the same tune tomorrow now that I've cloned the tables (is this desirable?).

In the meantime, my VVE looks like a Klingon Ski Slope:

VVE_03_25_2025.png

Does this look insane? How many runs does it take to really dial this in? When it is dialing in, should I see these values moving closer to 0 over time? If so, how quickly? MAF changes seem linear and immediate; VVE changes feel like wishes and dreams.

SD_Logs.png

Also, what conditions do you set for SD Tuning aside from the Dynamic Air calcs and filters? Turn off DFCO, Cat Protection, disable MAF only Air Calc mode? Leave timing stock?

I've never met anything like the VVE man, can't tell if it's a bona fide PITA or I'm still inept or both!

Thanks!

Only the manifold open tables are used. The rest could be tables of zeros.

That VVE shape looks way off and your error is still significant.

I'm more in favor of deliberately running the car in SD mode to get VVE right.

[Cringer]

Only the manifold open tables are used. The rest could be tables of zeros.

That VVE shape looks way off and your error is still significant.

I'm more in favor of deliberately running the car in SD mode to get VVE right.

Alvin, please refresh my memory...do the Gen5 ECM's suffer from the MAP^2 VVE bug? I thought Verlon said they did, but I cannot remember.

[Alvin]

Alvin, please refresh my memory...do the Gen5 ECM's suffer from the MAP^2 VVE bug? I thought Verlon said they did, but I cannot remember.

Not that I've seen. Even if it does.. quite ok to suffer though the once in a blue moon stumble to get it right IMHO.

[Cringer]

Not that I've seen. Even if it does.. quite ok to suffer though the once in a blue moon stumble to get it right IMHO.

OK thanks. IDK about once in a blue moon though...on my Gen4 I can trigger it at will and it will stumble every few seconds until I shut off the car.