BRZedit Fuel Trims, Closed to Open loop transiton

[mad_sb]

[Updated and corrected 5/5/2013]

Greetings,

First off, Those of you who are familiar with late model subaru ecu tuning won't be surprised by anything i found or have to say.. so may bee a quick "hey post some numbers" in the innovate supercharger thread would be a better use of your time than this thread will be

For those who are still with me....
Let me start with what lead me down the path i am about to explain. After working with BRZedit for a week or so i made my first trip to the dyno and did my initial all stock tune... Results were not bad.. I posted about it here.

After driving on the new tune for a couple of days I noticed that my WOT AFR's were not consistent... high 12's would become low 12's .. sometimes for a portion of the rpm range, sometimes all the way through redline. I would make some adjustments, flash it, drive it, felt great, next day back to nasty rich again wot.

So, I took a closer look at my logs and noticed my fuel trims looked horrible. Keep in mind this is a completely stock car with less than 1,000 miles on it at the time. Also, long term fuel trims DO apply under WOT. The ECU seems to have 4 trim zones, each can have a different value. The zones are broken down by grams/sec as measured from the maf.

Here is one of my logs displaying average short term fuel trims over the entire log (about 20 minutes of mixed driving from fully warm):


and from the same log, the average long term fuel trims:



Yep, that's nasty. Add 10% fuel when your AFR is tuned for 12.8:1 and you get richer than 12.1:1.. NICE As I did some more logging i noticed that I could induce big long term trims very easily by holding part throttle about 4,000 rpm... like i was waiting for a break in traffic to shove the go pedal through the floor.. not that i have ever done that. So, I cam to two conclusions:

1. The stock ecu, like many other late model subaru ECU's like to hold Closed loop way longer than i want it to.

2. The stock maf scaling does not work worth a crap for my car. (I had already checked for post maf leaks, filter issues, maf housing issues etc etc)

* A quick side note, there may be some variations in the stock maf housings causing some cars to like the stock maf scaling more than others, or maybe the sensors themselves have some wide tolerances form the factory.. I don't know. But I do think the issue with my car is why some people have horrible luck with after market intakes. It would probably only take a day for me to throw a CEL if i put an intake on that leaned out the car any more without re scaling the maf.

Closed to Open loop transition:
Here is an excellent write up of how subaru handles the closed to open loop transition on late model WRX's. I used this as the basis for the changes i made after I read it several times and matched up the tables talked about to the ones in BRZedit. Since we all love pictures.. here are most of the tables in the stock rom that control transition out of closed loop an into open loop:


A few important things to note:

• It is reported that the closed loop delay interval units are 10 milliseconds each (so a delay of 100 is about 1 second)
• The LAST thing checked before dropping into open loop is the requested AFR from the fule map

Long story short, you can bypass all the timers and force the transition to be based completely on your fuel map by setting the delay timers to 0, since the last thing the ecu checks when deciding to transition from closed to open loop is the requested AFR. If the requested AFR is richer than the Minimum Open Loop AFR the ecu drops into open loop mode.


[New info 5/5/2013]
With continued logging, testing, and tuning.. I noticed that many times the target AFR in closed loop mode was not 14.7:1 but actually a bit richer like 14.3 or 14.5 for example. It appears the ecu uses the Closed Loop AFR A & B tables to determine the closed loop target AFR. Here they are next to the stock fuel map.

Interestingly, it does not seem to target the exact value in either table (I assume one table is cold, the other is hot) but rather some where in the middle between the 14.7 and the values in these tables. It's possible that what i was seeing was interpolation between the two tables based on coolant temp and the target value does 100% come from these two tables.

If you set the Closed Loop Target A & B table values to 14.7:1 you end up turning off long term fuel trims all together. I only did this AFTER setting the transition timers to zero. In my case, i can set the entire closed loop target maps to 14.7 without worry because my transition to open loop is only based on the fuel map target afr. In other words, I will never be in closed loop with a target AFR lower than the Minimum Open Loop AFR value. If you want to keep your timers in place but still want to turn off long term trims, i would suggest you make the Closed Loop Target A & B tables mirror your fuel map. You will need to change some of the x and y axis values to do so.

When i started testing, i had hopes that the Long Fuel Trim Minimum and Maximum would actually allow you to set the max and min values for the long term trims. In theory this would allow you to use enough long term trim to compensate for changes in ethanol content, supplier to supplier variation etc, without destroying your tuned WOT AFR. Sadly, these two values do not seem to have ANY effect, at least with my rom. The stock value for the factory rom from my car was +/- .40 which i assumed was +/- 40%. To date, regardless of what value i use in these tables, if the long term trims are enabled, they get set without regard to the max and min values.
[End New info]

here is how i currently have my transition tables setup (anywhere the fuel map is leaner than 14.5 I'm in closed loop, anywhere it is richer than 14.5 i am in open loop:


Re Scale the Maf:
I still find it hard to believe i had to do this on a stock car but i did. I'm going to reserve the second post for a how to on maf scaling because i think may people would benefit from a re scale.

I would like to than @epifan for making a few changes per my request so that we can now log maf voltage directly. This makes re scaling the maf very simple to do... but I will save the how to for another day. For now the results:

current map's average short term fuel trims from a 30 minute log:


Long term from same log:


As an added bonus, my fuel map is now very close to the actual logged AFR.. I'm typically within 2.5% of the target AFR at WOT now. There are still a few places that need cleaning up and i think the few places you still see trims over +5% are due to needing a little bit more acceleration enrichment... it seems the trims are kicking in to provide the enrichment with smallish throttle changes.

[mad_sb]

Reserved for Maf Scale How To

[2forme]

This offers a lot of insight into why some cars hate some intakes while others do not. Awesome info @mad_sb!

[Visconti]

Nice write up.

Thanks!

[Yruyur]

Ditto

Sent from my SCH-I535 using Tapatalk 2

[Calum]

x4

Awesome work man! And a huge thanks on the transparency.

[Mars2]

LOve that!!!
Makes me understand why I feel like I drive a different car everyday :-)

[mad_sb]

Glad folks are getting some use out of my digging around

I have some more data to post in a bit, Currently I am working on allowing the long term trims to work in the closed loop areas without getting in the way of the open loop target AFR's

Some of the tuners i have spoken with have confirmed this is a common issue and one of the reasons most of the mail order tunes require fine tuning for the individual car.

[whtchocla7e]

I don't understand much of this but could it possibly explain why my stock car runs like crap for a while when started from cold?

[Doborder]

Please keep us updated... this is very interesting

[Visconti]

Originally Posted by mad_sb

Glad folks are getting some use out of my digging around

I have some more data to post in a bit, Currently I am working on allowing the long term trims to work in the closed loop areas without getting in the way of the open loop target AFR's

Some of the tuners i have spoken with have confirmed this is a common issue and one of the reasons most of the mail order tunes require fine tuning for the individual car.

Why don't you just adjust the tune so there aren't fuel trims everywhere ..

John

[mad_sb]

Originally Posted by Visconti

Why don't you just adjust the tune so there aren't fuel trims everywhere ..

John

Because then your fuel map looks like it was put together by a crack head And since the ecu uses the target AFR from the fuel map as the last check before going open loop (if the target afr from the fuel map is lower than the minimum open loop afr) that can further complicate the transition.

Ecuteck may have re coded some of the logic but in stock form it is very much like the other late model subaru ecu's (holding closed loop way too long).

Honestly though, once you zero out the timers so that you transition based only on the fule map target ARF, you can just tune the fuel map in the areas that are outside of closed loop... just personal preference.

[mad_sb]

So, I have some additional information i have been meaning to post for a while now.

Attached are the closed loop target AFR maps along side the stock fuel map. These target AFR maps are the AFR the ecu shoots for while in closed loop mode. However, the main fuel map still controls fueling.

So, for example at 2400 rpm 60 g/rev load the target AFR is 14.35:1 (14.55 depending if the car is cold or warm) while the fuel map is set to 14.7:1.... The min afr for open loop is 14.5 so you really would not want to set the fuel map any richer than 14.6 or so. As a result the only way to reach the desired AFR is with fuel trims.

I previously posted that by setting the long term fuel trim min and max to a low value i was able to disable the long term fuel trims... this was not the case though. As of today I still don't know what those parameters do, but they DO NOT control the max and min values of the long term trims... at least not on my rom. In fact, they don't seem to do anything at all that i can tell.

However, you can disable the long term fuel trims by setting the closed loop AFR tables to 14.7:1. When i made my original post, i had set these two tables to be equal to the fuel map... as a result they were 14.7 in the closed loop areas. The table change combined with zeroing out the timers (so closed to open loop is only controlled by the target afr) ment that my long term trims were disabled.

Currently, i just have the closed loop AFR tables flood filled with 14.7 and it behaves the same way, no long term trims, transition based only on target AFR.

Now, some of you may be looking at this and thinking LEAN CRUISE MODE..... Well, that may very well be possible. I have not tested this yet. I would assume you would want to raise the AFR in the desired lean cruise range on ALL 3 TABLES... Just setting the main fuel map to say 15.4:1 will NOT give you lean cruise since the closed loop target table would still be set much richer. This may not work at all though. It's possible the ecu will not target anything leaner than 14.7 and just ignore the value. I did do a test one day where i just set the closed loop target tables to 15.0 but when i drove it, the ecu did not target 15.0 but rather 14.7. I think that the closed loop target needs to either be the same as or richer than the main fule map for the ecu to actually use the closed loop target table.

I'll leave that testing to someone else because i honestly just don't care about lean cruise right now

[Visconti]

Originally Posted by mad_sb

Because then your fuel map looks like it was put together by a crack head And since the ecu uses the target AFR from the fuel map as the last check before going open loop (if the target afr from the fuel map is lower than the minimum open loop afr) that can further complicate the transition.

Ecuteck may have re coded some of the logic but in stock form it is very much like the other late model subaru ecu's (holding closed loop way too long).

Honestly though, once you zero out the timers so that you transition based only on the fule map target ARF, you can just tune the fuel map in the areas that are outside of closed loop... just personal preference.

I certainly don't have any problems with that.

I'm able to set the fuel targets correctly all while having a correct looking MAF curve with very small trims if any across the board.

Looks like you just need to spend a little more time on it or brzedit doesn't offer the control that ecutek does.

John