EQ <> VE Version 4

[wrenching1]

Quick question, Was just doing some testing and didn't understand why when you lets say input all stock factory data into the bluecat, generate the ve table then without changing anything in the ve, use the blue cat program to regenerate all zones, coefficients ...etc do you get different values from stock without changing anything. Should it spit out the same values? Or is it that I'm missing something?

[wrenching1]

I increased the precision of all the tables before importing, selected the ve format that I wanted to use , which I found created a false ve if you don't chose before generating the ve. After using bluecat to generate the ve I used the program to generate the equations and the vastly differ from the factory ones. Without changing anything should these match up somewhat? I know there is a error in the calculation but this has a wide margin.

[LSxpwrdZ]

Depends on the VE table... most of the stock E38/67 VE tables are not smooth at all. His tool tries to calculate a curve based on the data it is inputed. If you uncheck the automatic zone boundaries it'll be closer however his tool wants a smooth VE table and will manipulate the data to best fit the curve. With the cliffs and valleys in the stock VE table it is nearly impossible to re-create it using the tool. It's best to just start off with stock then get in the editor and manually smooth the VE table out on the rich side and start tuning. A VE table should be smooth anyway.

[waltjrUM]

I'm new to the virtual VE stuff (read 48 hours of a lot of reading), so take my input with a grain of salt.

I see what wrenching1 is talking about though. If you go through the workflow of this process with the tool, you can input a set of virtual VE data, generate the VE table, generate the equations directly from the VE table that was just created, and then check them against the set of input equations. I also found that the coefficients for example can be a lot different than those that were input (read more than the max error % shown in the generation status), even with max precision and single cell set to max.

I was a bit alarmed at first, but after messing with the settings a bunch of times and retrying the same process above, I'm leaning toward thinking that with all the polynomial equations, etc my guess is that the coefficients and squared values can vary so much because it is after all just a model and there are probably a large amount of combinations of constants/coefficients/multipliers that result in the same (read very similar) final VE value calculated by the PCM.

I could be completely off base, but perhaps someone can "check" my understanding. Thanks.

[waltjrUM]

So, after 1 iteration I'm not so sure about my previous post. The attached pics show my scan from the base tune (stock virtual VE) and my scan after 1 iteration of only half % multiplying the histogram cells that were 3 or greater into the generated VE table. The resulting virtual VE seemed significantly different than stock, but I ran it anyway to see what would happen. The 2nd scan looks worse, but the numbers are still in a pretty tight group.

I'm wondering if it's just that the generated virtual VE model is simply different from factory in some fundamental way (not necessarily good or bad). I noticed that the RPM boundaries were different from stock, but remained the same between iteration 1 and 2, so I'm going to try another iteration and see if things get closer to 0. If not I might give up for now lol, but I thought it would be cool to start learning this stuff while I'm still stock. Trying to keep myself busy so I can fight the itch to mod while I give my trans time to prove that the issue I was having was fixed by a repair the dealer did.

[ZO6_Bryan]

Quick question. I've done alot of VE tuning in SD mode on gen 3's. The virtual VE stuff aside. Do you put the PCM into SD mode the same in a gen 4 as gen 3. Fail the MAF and copy high octane table to low octane table? Anything different that you need to do in a gen4. Specifically an E67 in a 2012 ZL1.

[BigDaddyCool]

I increased the precision of all the tables before importing, selected the ve format that I wanted to use , which I found created a false ve if you don't chose before generating the ve. After using bluecat to generate the ve I used the program to generate the equations and the vastly differ from the factory ones. Without changing anything should these match up somewhat? I know there is a error in the calculation but this has a wide margin.

Just to annoy that ddecker fella, this program does exactly the same as the last one.... It does seem improved, I did read the original description of the 7 digits thing......

But here's the thing. You take a stock VE out, just equate it and copy it back....... how does a column go from eg. + 2100.00 etc to like -1500.00 in the constant area......... when you changed nothing......

And technically the software in a sense does less practical stuff and more tedious stuff....... VVT would be useful for those who need it but .......

meh. I'm trying to simply 'work with it', the positives are the extra smoothing options etc.....

I'm actually using a mix of the old version and the new version ....... both have cons and pro's.

Still I dont see how you take a stock calibration, equate it and when it goes back it doesn't produce the same figures......

Technically that means the whole 'i think i worked out the numbers' in a previous post was full of it.

------
The Simple:

Copy it out, equate it and put it back in without making any changes = SHOULD BE THE SAME.......
...

DStecker or whatever you can argue all you want with me it's a simple concept and that simple concept = NOT WORKING.

Make your custom VE table, put it in pull it out....... never the same.

--------

Frankly I certainly wouldn't pay for this software...... HP Tuners, look at their system..... great, working, gold......... their bugs are nothing to worry about if they ever have any........ but this, no.

...... The software is better in ways, but it cant do the first basic thing it was meant to do and that to me = stop now, it's a failure.

[DSteck]

I think somebody doesn't understand the nature of regression fitting. The program works 100% the same way GM's DYNA_Tools does. The sample area used to generate each section is slightly different.

Kill edge matching and see what happens.



Funny that this works perfectly fine for people who understand it, and have successfully SD tuned cars with it.

[waltjrUM]

Well, whether the coefficients are wildly different between iterations or not, the software appears to work for me. My latest scan is attached. It looks better than my original scan. I do have to note it was about 10* cooler than it was yesterday, but I let the car get up to temp fully before I started scanning. Also, the pattern of higher error at lower RPM from the original scan is reversed, so I'd say the weather probably couldn't have been a huge factor else I would see higher error across the board.

[LSxpwrdZ]

Just to annoy that ddecker fella, this program does exactly the same as the last one.... It does seem improved, I did read the original description of the 7 digits thing......

But here's the thing. You take a stock VE out, just equate it and copy it back....... how does a column go from eg. + 2100.00 etc to like -1500.00 in the constant area......... when you changed nothing......

And technically the software in a sense does less practical stuff and more tedious stuff....... VVT would be useful for those who need it but .......

meh. I'm trying to simply 'work with it', the positives are the extra smoothing options etc.....

I'm actually using a mix of the old version and the new version ....... both have cons and pro's.

Still I dont see how you take a stock calibration, equate it and when it goes back it doesn't produce the same figures......

Technically that means the whole 'i think i worked out the numbers' in a previous post was full of it.

------
The Simple:

Copy it out, equate it and put it back in without making any changes = SHOULD BE THE SAME.......
...

DStecker or whatever you can argue all you want with me it's a simple concept and that simple concept = NOT WORKING.

Make your custom VE table, put it in pull it out....... never the same.

--------

Frankly I certainly wouldn't pay for this software...... HP Tuners, look at their system..... great, working, gold......... their bugs are nothing to worry about if they ever have any........ but this, no.

...... The software is better in ways, but it cant do the first basic thing it was meant to do and that to me = stop now, it's a failure.

Read my previous post... this isn't 2+2=4.... 4-2=2. This is a complex polynomial curve in various regions of the virtual surface. You will simply never get a 100% match anytime you go from VVE to VE and then back. There is always some fudge factor in there at specific points. This is why when the program generates it shows you error for each zone.

[ZO6_Bryan]

Quick question. I've done alot of VE tuning in SD mode on gen 3's. The virtual VE stuff aside. Do you put the PCM into SD mode the same in a gen 4 as gen 3. Fail the MAF and copy high octane table to low octane table? Anything different that you need to do in a gen4. Specifically an E67 in a 2012 ZL1.

Never mind. Found my own answer. No need to worry about the spark tables.

[ZO6_Bryan]

My first try at this. Below is the Stock ZL1 VVE map and my new VVE map after smoothing out the problems areas with the extrapolate function. I will try this out in the car tomorrow but I think simply smoothing out the two big dips will help alot. As that is the part of the map that the engine jumps right into after putting the bigger pulley on.

[Brice]

VB.NET has a certain power to it that I like, but its an erratic, buggy, flawed piece of crap compared to VB6

You can't say that. You're trying to code .Net with the VB6 logic. .Net is oriented object.
So moving to VB.NET and thinking VB6 is a bad move ;-)

vb.net?? AHHHH!!!! Looks good though.

Keith ... I think you and guys @ hptuners should use bluecat program and incoparate his work on hptuners. (and pay him for it and his work).

Hptuners is in .Net technology too.
You can translate it from VB.NET to C# using reflector to adapt it for hptuners.

*edit *
Congrats Bluecat, it's a lot of equations ! Massive work... Just looked to your source code ... just wooooooo.... I feel your pain and I think you had long nights on it (Read_EQ fonction ;-)).
You can be pride of your work. Thanks for what you do for the community.
paypal on the way.

Keith ... you can definitively intregrate easily the bluecat functions to hptuners. make it happens!

[wrenching1]

Read my previous post... this isn't 2+2=4.... 4-2=2. This is a complex polynomial curve in various regions of the virtual surface. You will simply never get a 100% match anytime you go from VVE to VE and then back. There is always some fudge factor in there at specific points. This is why when the program generates it shows you error for each zone.

Jamie

Thanks for the response, i figured the complexity of the program with the smoothing function attributed to the deviation, just wanted to make sure i wasn't missing something.
I will be trying this on a ctsv with twin 62s

[edcmat-l1]

Just to annoy that ddecker fella, this program does exactly the same as the last one.... It does seem improved, I did read the original description of the 7 digits thing......

But here's the thing. You take a stock VE out, just equate it and copy it back....... how does a column go from eg. + 2100.00 etc to like -1500.00 in the constant area......... when you changed nothing......

And technically the software in a sense does less practical stuff and more tedious stuff....... VVT would be useful for those who need it but .......

meh. I'm trying to simply 'work with it', the positives are the extra smoothing options etc.....

I'm actually using a mix of the old version and the new version ....... both have cons and pro's.

Still I dont see how you take a stock calibration, equate it and when it goes back it doesn't produce the same figures......

Technically that means the whole 'i think i worked out the numbers' in a previous post was full of it.

------
The Simple:

Copy it out, equate it and put it back in without making any changes = SHOULD BE THE SAME.......
...

DStecker or whatever you can argue all you want with me it's a simple concept and that simple concept = NOT WORKING.

Make your custom VE table, put it in pull it out....... never the same.

--------

Frankly I certainly wouldn't pay for this software...... HP Tuners, look at their system..... great, working, gold......... their bugs are nothing to worry about if they ever have any........ but this, no.

...... The software is better in ways, but it cant do the first basic thing it was meant to do and that to me = stop now, it's a failure.

Have you tried using it to tune a car? Have you ever used EFIlive's VVE? By the way you post, it sure seems like you've never actually tuned a car using it. I've tuned a shit ton of them, and it works great. I've got a couple ZL1s running around in SD and they run flawlessly.

[motorheadmike]

I am incredibly excited about trying this new version out.

Bluecat, you are a credit to this community for your generosity and support in creating and distributing this program. It is rare to see this kind of act in this day and age; thank you for your efforts!

[LSxpwrdZ]

I actually just tried it out the other day for the first time as I've been using the older 3.5 version for a while now. I really like his updated version alot! Seems to me it calculates faster as well which is a huge plus IMO. Great work Phil!

Sent you some $$$ also for your hard work!

[BigDaddyCool]

I am still messing with the new version to see what I think........

It is better then the last at this point I'd say by far.... But still.... I use them both, the old version for it's big ve map and editing ...... and then use the newer one to equate the ve and export to hp tuners.

Prob i find is tuning ve or maf is easy in open loop with wideband, but when you put it back into close loop with the mix of ve and maf enabled it just all goes to ............ Maybe it.s just my ramjet being a real ........... to tune, it's mostly sorted but...

eg. stock ve vs eq ve........ i find my o2 sensors are sticking in certain spots because the ve table gear has been changed by me with eq........ wideband may also report that area is the right atf ratio........... i'm working to resolve this ......... my prob is that exactly tho if the stupid eq program is out 1-5% when it equates the ve in that spot it can stuff my setting for that area.

eq works......... and at the end of the day bugs or what not 'yes' it will do the job. I still think it needs a tidy up and a bit of refining.

Im half thinking buying a 2.5BAR map sensor and just upgrading to the custom os to run it mafless etc may be the way to go for me.....

[MikeOD]

Im half thinking buying a 2.5BAR map sensor and just upgrading to the custom os to run it mafless etc may be the way to go for me.....

You can run the 2.5 bar OS with a stock MAP sensor...just set the MAP calibration the same as it is in the stock OS, then set your VE table KPA axis to whatever resolution you want for that MAP sensor range.

[LSxpwrdZ]

You can use whatever VE axis you want with the new version... He made it customizable so you can generate whatever VE table you want.