Prediction Coefficients

[ttz06vette]

Thanks...do you think twin turbo will be same as ZL1 blower settings?

[schpenxel]

Stock ZL1 is 30-40% higher than your stock setting, which is pretty close to what I thought you should try

If you need any help setting your histograms/charts/graphs back up just let me know. You've got my email address

Taking screenshots of your current setup will make it easier to get the same histograms back

[ttz06vette]

Stock ZL1 is 30-40% higher than your stock setting, which is pretty close to what I thought you should try

If you need any help setting your histograms/charts/graphs back up just let me know. You've got my email address

Taking screenshots of your current setup will make it easier to get the same histograms back

Thanks Carson.

[JBZ]

OK so ive read through this thread a few times and think that my 2010 CTS-V may suffer from Prediction Coefficients but not quite sure. there is a lot of great information in this thread and getting a definitive answer and reading through the lines is a bit tough to say the least.

so here is what im up against, not only do I have an initial lean spike (not real high but there) but I get a lean bit between 8700 and 9900 htz almost every pull I do, which typically is still under my dynamic disable RPM. I have my VVE and MAF dialed in as good as my novice self can get except for these two areas. I have been increased my manifold volume by 15% as of yesterday before I took the logs (screenshot attached) and it has improved the lean tip in but I have more to go. I will increase it by 20% (from factory value) today and see how that goes. but I am wondering if this lean spike rise after I have already entered PE (8700-9900 htz) may have something to do with the prediction coefficients, but really need some input from you guys who are not only smarter than I am but have messed with these coefficients a lot.

another thing to note is it could be possible that the initial PE lean (sometimes its flat lined not lean) spike could be sensor latency? tried to back calculate it and it appears to be about 150 ms.



so here is my current coefficients, everything else is zeroed out. what I don't get is that like the OP said the ZL1's (same engine) are completely different.







So I read time and time again that the MAF table should be smooth, no sharp edges and spike. well if I multiply by% in this 8700-9900 htz lean anomaly it makes my MAF all crazy and look like this. note that I did not save this, this is just an example and I don't think it is the right direction, hopefully to find out what is actually causing this.







here are some screenshots of logs since the 15% increase in my manifold volume calc. notice the lean spike when first entering PE (which has decreased with the 15% increase) but also the 8700-9900 range lean spike and the rich tip out

[GHuggins]

Yours is modified enough, so about time to zero them out ; )

[JBZ]

Yours is modified enough, so about time to zero them out ; )

so should I zero out all of them, base, current and corrected?

[GHuggins]

Everything under the prediction coefficient header...

[barum]

Yours is modified enough, so about time to zero them out ; )

Why is that? Unless DynAir is doing unwanted things, which it isn't in the print screens above, there's no need to zero everything out.

I've been running with
- base 0.9 except cells 11, 17, 23, 29 where base is 0.95
- current 0.1 except the cells listed above where current is 0.05
- old 0.0
- corrected 0.1

Yesterday I changed corrected to 0.5 like it is in LS9 and noticed immediately that some of the problems at idle/during cold start with pure SD mode came partly back. This led me to think that the corrected gain may actually be some sort of weighting factor between MAF based part and GMVE based part.

The engine seemed to like this change on the general level, so I may experiment more by
- using more MAF based values (lower) at idle
- using more GMVE based values (higher) above the idle/low load

Edit. This is going in next based on the above assumption and the actual ave/min/max MAP vs speed vs TPS data in normal traffic:





Edit2. This is a comparison of the stock LS7 parameters shown in HPTuners and EFIlive, pretty much the same jsllc showed earlier, but with variable names (as interpreted by each company):



Btw, note how GM uses higher corrected gain in cells 2 & 3 where the very low speed cruising and tip-in + clutch off takes place.

[JBZ]

I had terrible luck with zeroing out my prediction coefficients and went back to the factory values after one drive. it was twitchy at low speed and I would get these lean spikes any time I blipped the throttle. even in the 25% range. you can look at these circled areas of the throttle blips and lean out. these were in stop and go traffic and not anything unusual as far as throttle is concerned. the only other thing I changed in the entire tune was manifold volume...that I have kept the same since going back to factory coefficients and I do not get this leanout.

[barum]

Intake volume alone has an effect on transient fueling, so unless you haven't swapped the stock supercharger to something else, there's no need to change the volume. To my understanding there's (yes, yet another) mathematical model calculating how the fuel travels/vaporizes within the intake and by changing one of the primary parameters (volume) you cause major discrepancy between the model and real world.

[schpenxel]

But.. I've also read different takes on whether putting it in SD screws up the transient model anyways. I read something from I believe Chris@HPT saying disabling the MAF messes with the model, so at that point I'm not sure it matters since you're just trying to get it back to working correctly.

I've read others say it uses dynamic airflow, not MAF airflow, so being in SD doesn't make any difference

Personally I think it's the former

[JBZ]

Intake volume alone has an effect on transient fueling, so unless you haven't swapped the stock supercharger to something else, there's no need to change the volume. To my understanding there's (yes, yet another) mathematical model calculating how the fuel travels/vaporizes within the intake and by changing one of the primary parameters (volume) you cause major discrepancy between the model and real world.

I increased it off of recommendations by others here. I don't know the better . but honestly it did resolve a lean spike on tip-in that I was getting while entering PE, with no ill effects seen on the scanner anyways. I increased it by 15% and trying rationalize that the tip-in lean spike was due to maybe my increased boost and possibly changing the actual volume. but again I really don't know if that is right or wrong.

regarding the prediction coefficients I was under the assumption that I was getting a lean area (not spike) after being in PE at the same MAP/MAF hts regardless of being in MAF only or blended. of course that makes me think that it is predicted incorrectly. and going to maf only didn't change that issue. for me to tune my MAF around that problem was making it ugly and jagged in that area. I may have to say ooh well and make it look like what it looks like to properly fuel that region.

[cc-rider]

I have my prediction coeff's zero'd out, and my manifold volume increased, and have not had any issues at all with odd fueling or lean spikes. But I have been running SD tune. These settings all seem to be interconnected to each other in a wild equation/algorithm that can only be figured out via trial and error.

[GHuggins]

Intake volume alone has an effect on transient fueling, so unless you haven't swapped the stock supercharger to something else, there's no need to change the volume. To my understanding there's (yes, yet another) mathematical model calculating how the fuel travels/vaporizes within the intake and by changing one of the primary parameters (volume) you cause major discrepancy between the model and real world.

Problem is that although the physical displacement of the manifold hasn't changed, it flows much more air - so you need to increase manifold volume... This is the quickest and easiest way to dial in transient fueling, prediction coefficients and well - all of it... No one to my knowledge has the ability to dial these in - if we did - I would Another example is when someone installs turbos - intake is still the same right, but we are flowing a far larger amount of air through it... You would think that dynamic air being dialed in would correct for this, but it doesn't... We can go in a screw with transient tables and on and on, but nothing will get things right like manifold volume will for deceleration and tip in fueling...

[JMsquared]

Ok, I'm going to take the bait, I'm having this same problem with my 2.4 SC Cobalt, the tip in lean out and pop. What version of HPT 3### has the manifold volume parameter?
I am using 3.2## now, there is no manifold volume listed as indicated in a previous post.

This has been on spellbinding thread....

[schpenxel]

It'd been added on a case by case basis as best I can tell. I've only seen it on E38 controllers personally.. not sure if it exists or not for that application

[JBZ]

I had to request it from hpt support as well.

[barum]

After a few more tests with "corrected" table and zone boundaries, these seem to work best (smoothest DynAir, reasonably fast response time), at least for NA applications:
- base 0.9
- current 0.1
- old 0.0
- corrected 0.0 <= almost any change to this will lead to worse (more noise in output signal and/or unexpected/goofy prediction result during transitions) DynAir

[barum]

These have been added to HPT:

[ECM] 34971 - MAP Estimated Max: Estimated maximum MAP value.
[ECM] 34972 - MAP Estimated Min: Estimated minimum MAP value.
[ECM] 12003 - Intake Manifold Volume: The volume of the intake manifold. Used as an input into the transient fueling model.

You have to request them though. Info you need is above and tell them they can find them on a 2012 ZL1 E67

Good to know, so this is where they will show up:

[Michael_D]

I tried to zero out the predicted tables once. Don't recall the exact details, but it made burst knock go bonkers. I happen to like the burst knock feature (I know some just kill it), so ended up just going back to the OEM settings.

Looks like I have something else to experiment with..... Does it ever end? Will I ever not have a laptop in the passenger's seat?