Per-Gear Tuning

[iNVAR]

lol

[SeeeeeYa]

I have used it on TD04's with success. 17psi 2400rpm in 1st gear in manual and autos. 2400rpm launch got a 1.66 60ft with clutch slip

 

And, as mentioned, it was an LGT owner who provided my springboard for this. The performance of his SpecB is nothing short of amazing, and he can only implement per gear boost on his 09 SpecB.

 

You are leaving a pile of fine edibles on the table if you don't.

[_MFB_]

The smart guys looked long and hard for this logic in the Legacy ECU but didn't find it.

What legacy ecu mate?

[_MFB_]

Ive used this to have virtual boost target and wgdc control. Trying to mimic the crude method of the sti 03 jdm and its per ger boost compensation. How does the 1st and 6th gear row work for sti 08+?

 

All these tables IMHO are from subaru as both cobb and OS just reverse engineer what has been made.

 

Its stuff like SD , Learning view, maf blending, Merp mode, Carberry that make it original. Excited for DTC cel light flashing!

[SeeeeeYa]

Ive used this to have virtual boost target and wgdc control. Trying to mimic the crude method of the sti 03 jdm and its per ger boost compensation. How does the 1st and 6th gear row work for sti 08+?

 

All these tables IMHO are from subaru as both cobb and OS just reverse engineer what has been made.

 

Its stuff like SD , Learning view, maf blending, Merp mode, Carberry that make it original. Excited for DTC cel light flashing!

 

The IWSTI thread I linked in the first post of this thread explains.

 

In short, awesome results.

[fahr_side]

What legacy ecu mate?

 

4th gens. Of course 5th gen has per-gear... tune with this everyday. Thanks again Eric Minehart!

[_MFB_]

dschultz got the per gear torque request for my Forester :-) http://www.romraider.com/forum/viewtopic.php?f=25&t=7048

RR guys taught me how to use it then I came up with my own approach. I see you are specific with 2011 sedan. Ive read in the past it has been discussed that values are just interpolated. Most guys at RR dont really keep a secret. Just ask nicely ;-)

 

Yep, 5th gen got it good fahr_side :-)

 

Thanks SeeeeeYa, I'll fix up my tables now to implement per gear target and wgdc :-)

http://www.iwsti.com/forums/3795940-post23.html I agree with Airboy on the multiplier. Wouldve been easier though if the scaling could be edited to have the same value as the multiplier (per gear timing) so no conversion is needed.

[_MFB_]

I also see that you have posted as well on preemptive knock. Its been discussed here and defined in Ecuflash/romraider http://www.romraider.com/forum/viewtopic.php?f=15&t=8509

Here it is on my forester 2009 EDM http://i46.tinypic.com/35b7fig.jpg

[_MFB_]

It used to be like this up to 2005 (here's a 2003 sti JDM)

http://i47.tinypic.com/2b4r4k.jpg Simple :-)

[SeeeeeYa]

Thanks, _MFB_, it's good to know that when Cobb gets around to attending to preemptive knock in ATR there is plenty of background to go on.

 

To me, however, this is Greek, LOL. And, from looking through whatever ROMs I can get my hands on there is little directly usable data between USDM ECUs and those from other markets. Engines are different, etc. Data such as you've provided will benefit the OS knowledgeable, but has no impact on ATR-only users. I feel fortunate that my STi has all the necessary tables in ATR.

 

As mentioned in my Cobb thread, I'm good now with preemptive knock control. I'm glad it's there even if it is "hidden" from my direct view. I can evaluate knock's origin with Knock Sums, and I've mitigated the most egregious unnecessary preemptive FBKC by correcting my Boost Targets to match my post-per gear reality. Whatever's left is addressed by the timing.

 

It is interesting how tuning per gear has broadened my tuning experience from the process of implementing it, and my car's performance in power as well as day-to-day driving, but also how those have worked together to require ripple-effect tuning corrections well beyond what I first imagined. There's no such thing as simple tuning. Everything affects everything else.

 

The thought of going back, to mere Stage2, is as appealing as reverting to stock. I feel like I'm driving an STi now that has bite to match its bark.

[_MFB_]

I understand you on per gear timing and per gear boost control on the sti 2011. Waiting to uncap ours in EDM. ;-) Ive done this though for Legacy, forester and wrx edm (wrx has no per gear timing) Just watch out for total ignition timing. Does Cobb already allow you to log per gear timing? I previously asked Cobb for support for non USDM subarus. Just glad OS is here to help us out.

 

OS is pretty much complete now for us Non USDM subarus. Most of the alpha defs are for non stock set up (bigger injectors, etc) while some may be for advance tuning strategies (like pgtr and pgt).

 

The tau tables are one of big interest for you as that will get rid of the rich dip once load increases. You see this on the dyno too, more prevalent though in FMIC. Another one is the Front AF smoothing table to get rid of overly reactive ecu logic ;-) Bamofo of RR got this out, useful for those with cai. Then theres another Requested torque table recently found by Throttle Happy. ;-) All is good, hope both commercial and OS can work together in making our subaru work for us (end users)

[COBB Tuning]

Thanks, _MFB_, it's good to know that when Cobb gets around to attending to preemptive knock in ATR there is plenty of background to go on.

I've taken a quick look at this yesterday and we already have a table defined in our Accesstuner software to manipulate the alternate feedback knock correction being discussed here and elsewhere. The table is called "Feedback Knock Retard Decrement Alternate (Max. CL Target Rich)". However, on further analysis, this alternate correction can also be used with knock activity, so at this point, I would not zero out this table. We have this defined for the 07-12 LGT, 06-13 FXT, 07-13 STi, and 06-13 WRX.

 

Bill

[_MFB_]

09 WRX onwards have per gear Mark 08 missed out

 

Tables are Calculated Engine Torque, realised what they are from Audi/R35 tuning. Is interesting seeing the values used on WRX Vs STi, roughly 10% higher on STi...

 

Yeah, I had this checked a few times hahaha So when are you going to give me Calculated Engine Torque? Got some few lgt, wrx, sti, and my foz edm to "test". :-)

[fahr_side]

I know that the '08 ROM AZ1E400U can be flashed to the '09 WRX that stock had AZ1G400U, from Eric doing this. Seems logical the reverse would work. That would allow per-gear timing to be employed on the '08 eh?

[_MFB_]

It may give you the error write on protected rom areas. Maybe there's another step needed to do this. :-)

[SeeeeeYa]

I've taken a quick look at this yesterday and we already have a table defined in our Accesstuner software to manipulate the alternate "pre-emptive" feedback knock correction being discussed here and elsewhere. The table is called "Feedback Knock Retard Decrement Alternate (Max. CL Target Rich)". So, you simply enter zero in that table, save, reflash and it will disable the alternate correction. We have this defined for the 07-12 LGT, 06-13 FXT, 07-13 STi, and 06-13 WRX.

 

Bill

 

Thank you, Bill. I would never have figured that out. Will do it tomorrow!

 

Do you foresee any negatives to disabling this??

[COBB Tuning]

Thank you, Bill. I would never have figured that out. Will do it tomorrow!

 

Do you foresee any negatives to disabling this??

Actually, do NOT set this table to zero. Both feedback decrement values (main and alternate) are potentially active with knock activity. I think there may be another explanation for feedback knock with no knock sum increase, but we will have to set aside time in the future to take a look at it (the behavior may potentially vary with certain ECUs). In the meantime, if you have any logs that show the behavior, please post them here or in the thread at the cobb forums.

 

Bill

[SeeeeeYa]

Actually, do NOT set this table to zero. Both feedback decrement values (main and alternate) are potentially active with knock activity. I think there may be another explanation for feedback knock with no knock sum increase, but we will have to set aside time in the future to take a look at it (the behavior may potentially vary with certain ECUs). In the meantime, if you have any logs that show the behavior, please post them here or in the thread at the cobb forums.

 

Bill

 

Thanks for the follow up. It caught me in time.

 

It appears that my editing of boost targets has ended the FBKC initiated by negative TDBE. I'm relieved of that concern, and glad that the ECU continues its overwatch of knock without interference.

 

Combined with refinement of my per-gear timing, this and various other related tuning has culminated in datalogs, which include all my driving miles, without a single instance of knock. As temperatures increase and I further adjust to accommodate, my fingers will remain crossed.

 

Some things about this per-gear experience are just too good to be true... and we all know that story.

[spectre9]

I think there may be another explanation for feedback knock with no knock sum increase

Bill

 

I have read that knock sum only increases at the end of a sequence of knock events. This seems to match my observations of timing being added/pulled during streams of knocks... with knock sum only incrementing when knock has stabilized.

 

One should not assume that knock sum that increments slowly means low knock... nor does a high count == lots of knock.

 

A few intermittent knock events vs frequent knock that occurs faster than a countdown timer is my view of the situation. This appears close to reality in my observations.

[SeeeeeYa]

I have read that knock sum only increases at the end of a sequence of knock events. This seems to match my observations of timing being added/pulled during streams of knocks... with knock sum only incrementing when knock has stabilized.

 

One should not assume that knock sum that increments slowly means low knock... nor does a high count == lots of knock.

 

A few intermittent knock events vs frequent knock that occurs faster than a countdown timer is my view of the situation. This appears close to reality in my observations.

 

Yes, from the viewpoint of exceptions, you are correct. For example, very high knock sums can come from lots of low load/low rpm knock... knock that the ECU essentially creates on purpose. Stock tunes are full of it. Then one stops that non-damaging information and is left with much lower knock sums... only what is there now is real knock, the kind that hurts.

 

But this is not the norm, this is the product of modified tuning and specialized eyes. That is, it doesn't provide generalizations. As a general rule, high knock sums better be viewed, and analized, as indicating lots of knock events. I can see the possibility for what you're saying. But it is an exception to the vast combination of probabilities where higher knock sums means higher knock.

 

When coping with danger one is best served by focus on the negative probabilities instead of the positive ones. ANY indication of knock is sufficient for alarm, and ALL knock warnings are viewed as equally dangerous.

[GTEASER]

Subscribed to be schooled.

[BlazeRex]

Ohh boy, I'm loving how much more code is being discovered. Some of it has proven to be very helpful! I do question myself sometimes what they were thinking when they implemented some of this code..