Rear O2 sensor closed loop fueling target limits

td-d · **Posted:** Thu Jul 12, 2012 4:57 am

As promised in this thread, I've scratched around for the Rear O2 sensor closed loop fueling targets, and after some testing with MFB, I'm quite confident that I have found them. Zeroing these should remove the effect of the rear O2 sensor on closed loop fueling - so it's part of the arsenal of taking the rear O2 sensor out of the loop.

To demonstrate the point, see the attached graph, with stock values and with these zeroed - the fairly regular spikes to the CL fueling target are removed, creating a much more consistent target. These are limits to the compensation added to the CL fueling target - i.e. lambda.

For AZ1G101N, here are the definitions (excuse the messy scaling):

Code:

 <scaling name="AF3 Lambda" units="Lambda Additive" toexpr="x" frexpr="x" format="%.4f" min="0" max="255" inc="1" storagetype="float" endian="big"/>

         <table name="Closed_loop_fueling_target_comp_RearO2" category="Experimental" address="2E698" type="2D" level="1" scaling="AF3 Lambda">
      <table name="X" type="Static X Axis" elements="2">
         <data>High</data>
         <data>Low</data>
      </table>
   </table>

This will give you high and low range of 0.5 and -0.5 lambda (i.e. 7.35 and - 7.35 AFR) - as ridiculously high as the AF3 correction limits.

To find them in your rom

In IDA, do a search for 3F 00 00 00 BF 00 00 00 (all results). You will get 4 or 5 results depending on your Rom. These are the -0.5 and 0.5 sequences.

One of these sequences is the AF3 correction limits, as documented in my other thread. One of them (the first in this rom) will be the the compensation limits. If you follow the search, you should find a location with nearby references to closed loop compensation tables (ECT), e.g.:

Code:

ROM:0002E662 ; ---------------------------------------------------------------------------
ROM:0002E664 off_2E664:      .data.l unk_FFFF6580    ; DATA XREF: sub_2E43C+4r
ROM:0002E668 off_2E668:      .data.l Ram_CoolantTemperature ; DATA XREF: sub_2E43C+8r
ROM:0002E66C off_2E66C:      .data.l E_Engine_Load_4ByteExt_E32 ; DATA XREF: sub_2E43C+Cr
ROM:0002E670 off_2E670:      .data.l unk_FFFF8A7E    ; DATA XREF: sub_2E43C+10r
ROM:0002E674 off_2E674:      .data.l Table_CL_Fueling_Target_Compensation_Imm_NonCruise_ECT
ROM:0002E674                                         ; DATA XREF: sub_2E43C+16r
ROM:0002E678 off_2E678:      .data.l Table_CL_Fueling_Target_Compensation_Imm_Cruise_ECT
ROM:0002E678                                         ; DATA XREF: sub_2E43C+1Ar
ROM:0002E67C off_2E67C:      .data.l Fun_GetMapValue3D ; DATA XREF: sub_2E43C:loc_2E458r
ROM:0002E680 off_2E680:      .data.l unk_FFFF7324    ; DATA XREF: sub_2E43C+22r
ROM:0002E684 off_2E684:      .data.l flt_C7668       ; DATA XREF: sub_2E43C+28r
ROM:0002E688 off_2E688:      .data.l unk_FFFF7078    ; DATA XREF: sub_2DEFE+58Ar
ROM:0002E688                                         ; sub_2E4FA+18r
ROM:0002E68C off_2E68C:      .data.l unk_FFFF73A0    ; DATA XREF: sub_2DEFE:loc_2E48Cr
ROM:0002E690 off_2E690:      .data.l flt_C766C       ; DATA XREF: sub_2DEFE+5AAr
ROM:0002E690                                         ; sub_2E4FA+C6r
ROM:0002E694 off_2E694:      .data.l Fun_check_high_low_thresholds
ROM:0002E694                                         ; DATA XREF: sub_2DEFE+5B8r
ROM:0002E694                                         ; sub_2DEFE+5E6r
ROM:0002E698 Closed_loop_fueling_target_comp_RearO2_max:.float 0.5
ROM:0002E698                                         ; DATA XREF: sub_2DEFE+5BAo
ROM:0002E698                                         ; sub_2DEFE+5E8o
ROM:0002E69C Closed_loop_fueling_target_comp_RearO2_min:.float -0.5
ROM:0002E69C                                         ; DATA XREF: sub_2DEFE+5BEo
ROM:0002E69C                                         ; sub_2DEFE+5ECo
ROM:0002E6A0 off_2E6A0:      .data.l flt_C7674       ; DATA XREF: sub_2DEFE+5C2r
ROM:0002E6A4 off_2E6A4:      .data.l flt_C7678       ; DATA XREF: sub_2DEFE+5D8r
ROM:0002E6A8 off_2E6A8:      .data.l Func_Check_Max  ; DATA XREF: sub_2DEFE+5DEr
ROM:0002E6AC off_2E6AC:      .data.l Ram_Coolant_Temp_ ; DATA XREF: sub_2E4FA+6r
ROM:0002E6AC                                         ; sub_2E4FA+44r
ROM:0002E6B0 off_2E6B0:      .data.l unk_FFFF744A    ; DATA XREF: sub_2E4FA+14r
ROM:0002E6B4 off_2E6B4:      .data.l flt_C7680       ; DATA XREF: sub_2E4FA:loc_2E53Ar
ROM:0002E6B8 off_2E6B8:      .data.l byte_C7302      ; DATA XREF: sub_2E4FA+90r
ROM:0002E6B8                                         ; sub_2E4FA+B8r
ROM:0002E6BC off_2E6BC:      .data.l flt_C7670       ; DATA XREF: sub_2E4FA+9Er
ROM:0002E6C0 flt_2E6C0:      .float 64.0             ; DATA XREF: sub_2E4FA+EEo
ROM:0002E6C4 off_2E6C4:      .data.l word_8688C      ; DATA XREF: sub_2E4FA+F2r
ROM:0002E6C8 off_2E6C8:      .data.l sub_BE8C0       ; DATA XREF: sub_2E4FA+F4r
ROM:0002E6C8                                         ; sub_2E4FA+122r
ROM:0002E6CC off_2E6CC:      .data.l word_C7306      ; DATA XREF: sub_2E4FA+104r
ROM:0002E6D0 off_2E6D0:      .data.l word_86880      ; DATA XREF: sub_2E4FA+120r
ROM:0002E6D4 off_2E6D4:      .data.l word_C7304      ; DATA XREF: sub_2E4FA+132r

The code looks like this:

Heide264 · **Joined:** Sat May 07, 2011 11:22 am **Posts:** 144

Thanks =).

How would this relate to disabling AFL3? I am having a bit of trouble piecing the two together in their differences. One seems like a 'cause', one seems like an 'effect' to me.

td-d · **Posted:** Thu Jul 19, 2012 12:40 pm

In terms of AF learning 3 - I've found only a few roms that actually have it active (and later ones) - which would explain why many people who have logged it find it to be zero (the SSM routine doesn't pull a parameter). In those where it is evident, I would assume that like AF1 learning, the learning is based on historical behaviour of AF3 Correction - which I have found in basically all the roms I've opened up so far.

So AF3 correction, in testing, has seen the biggest impact - in that zeroing it completely removes short term AF corrections based on the rear O2 sensor.

In testing, zeroing these rear O2 compensation limit also made fueling in closed loop less variable. So I assume - but have not had the chance to test - that if you completely remove the rear O2 sensor, you would have to both zero the AF3 corrections and zero these limits (and of course remove the CELs) to completely get rid of the rich closed loop conditions that were noticed with the rear O2 delete.

I've confirmed with Bill as well (on another thread in IW) that these tables can indeed be used to completely remove the rich closed loop condition caused by removing the rear O2.

It would be interesting to define these for someone on here who a) has completely removed the rear O2 sensor, and b) has experienced the rich condition.

throttlehappy · **Posted:** Thu Jul 19, 2012 12:54 pm

td-d wrote:

In terms of AF learning 3 - I've found only a few roms that actually have it active (and later ones) - which would explain why many people who have logged it find it to be zero (the SSM routine doesn't pull a parameter). In those where it is evident, I would assume that like AF1 learning, the learning is based on historical behaviour of AF3 Correction - which I have found in basically all the roms I've opened up so far.

So AF3 correction, in testing, has seen the biggest impact - in that zeroing it completely removes short term AF corrections based on the rear O2 sensor.

In testing, zeroing these rear O2 compensation limit also made fueling in closed loop less variable. So I assume - but have not had the chance to test - that if you completely remove the rear O2 sensor, you would have to both zero the AF3 corrections and zero these limits (and of course remove the CELs) to completely get rid of the rich closed loop conditions that were noticed with the rear O2 delete.

I've confirmed with Bill as well (on another thread in IW) that these tables can indeed be used to completely remove the rich closed loop condition caused by removing the rear O2.

It would be interesting to define these for someone on here who a) has completely removed the rear O2 sensor, and b) has experienced the rich condition.

AZ1G101M is in same location as AZ1G101N

OK, so what is references by the other 3F 00 00 00 BF 00 00 00 values? Wondering if I can just zero them all out?

td-d · **Posted:** Thu Jul 19, 2012 12:58 pm

throttlehappy wrote:

AZ1G101M is in same location as AZ1G101N

OK, so what is references by the other 3F 00 00 00 BF 00 00 00 values? Wondering if I can just zero them all out?

Not sure - haven't check them out - there's only between 3 to 5 of these in each rom (depending on the make and year). I wouldn't just zero without knowing what the other addresses are :lol:

Heide264 · **Joined:** Sat May 07, 2011 11:22 am **Posts:** 144

td-d wrote:

So AF3 correction, in testing, has seen the biggest impact - in that zeroing it completely removes short term AF corrections based on the rear O2 sensor.

In testing, zeroing these rear O2 compensation limit also made fueling in closed loop less variable. So I assume - but have not had the chance to test - that if you completely remove the rear O2 sensor, you would have to both zero the AF3 corrections and zero these limits (and of course remove the CELs) to completely get rid of the rich closed loop conditions that were noticed with the rear O2 delete.

It would be interesting to define these for someone on here who a) has completely removed the rear O2 sensor, and b) has experienced the rich condition.

My mistake, I meant AFC3, not AFL3. Regardless, I appreciate the explanation.

I actually locked my car in open loop fueling months ago due to the richness. It made it near impossible to get my fuel comps down and it also botched my idle up pretty good. I'm just getting tired of tweaking the MAF scaling every 6 months, so I'm looking to leave my car in closed loop at idle... just to account for fuel/atmospheric variables. I don't have my rear o2 sensor removed - I just took plan b and ignored it.

I moved around both the AFC3 limits as well as the AFL airflow thresholds into my 32bit base and ROM definition. I pretty much stick to ECUFlash anymore for editing my ROMs. I will try out at least the AFC limit set to zero and post up my experiences. I will bump myself back into closed loop before trying it for a before/after.

I don't have IDA personally. If somebody could toss me an address for the fueling target limit, I'd be happy to experiment with that afterwards. My ECUID is an AE51910V (2011 usdm STi).

I need to read into the CL compensation tables a bit as well. I'm not as clear as I should be on those and their practical function. I may just set the whole thing to zero for now.

I appreciate all the work. Thanks again.

td-d · **Posted:** Thu Jul 19, 2012 4:36 pm

Here you go - for AE5I910V:

ROM:000328A4 Closed_loop_fueling_target_comp_RearO2_max:.float 0.5
ROM:000328A8 Closed_loop_fueling_target_comp_RearO2_min:.float -0.5

mrf582 · **Posted:** Tue Jul 31, 2012 3:46 pm

Can I request the same for a 2006 WRX ROM? A8DH201X
download/file.php?id=686

If set these to zero, is there any point in setting the AF3 correction limits to zero, like in your other thread?

td-d · **Posted:** Tue Jul 31, 2012 4:37 pm

mrf582 wrote:

Can I request the same for a 2006 WRX ROM? A8DH201X
download/file.php?id=686

Sure - I'll dig them out for you.

mrf582 wrote:

If set these to zero, is there any point in setting the AF3 correction limits to zero, like in your other thread?

I've not been able to verify whether AF3 correction alone would deal it - but from a short conversation on IWSTI with Bill, I get the feeling that zeroing both these would be the best approach to completely switch off any rear O2 corrections (although AF3 correction has by far the most direct impact). I unfortunately don't have my car at the moment to test this by unplugging the rear O2...

td-d · **Posted:** Tue Jul 31, 2012 6:10 pm

mrf582 wrote:

Can I request the same for a 2006 WRX ROM? A8DH201X
download/file.php?id=686

Alright - I'm relatively confident these are the rear O2 comps (but the code is quite different between the 08's up in terms of indirect references, and cruise / non cruise logic)

Code:

ROM:0002E020 Closed_loop_fueling_target_comp_RearO2_max: ; DATA XREF: sub_2C730+1798o
ROM:0002E020                 .float -0.5
ROM:0002E024 Closed_loop_fueling_target_comp_RearO2_min: ; DATA XREF: sub_2C730+179Co
ROM:0002E024                 .float 0.5

mrf582 · **Posted:** Tue Jul 31, 2012 6:44 pm

Damn, you are quick! Thank you much!

dj-phoenixx · **Posted:** Mon Aug 13, 2012 2:00 pm

Hi,
do you have the adress for my A2WC522N Map? My WBO is installed and I see an AFR about 14,2 -14,3 at idle. The AF#3 correction limit is set to zero.
I'ld change to SD for next time maybe it would help for some useful logs.

Greets
Marel

td-d · **Posted:** Tue Aug 14, 2012 1:49 pm

dj-phoenixx wrote:

Hi,
do you have the adress for my A2WC522N Map? My WBO is installed and I see an AFR about 14,2 -14,3 at idle. The AF#3 correction limit is set to zero.
I'ld change to SD for next time maybe it would help for some useful logs.

Greets
Marel

Code:

ROM:0002C91C Closed_loop_fueling_target_comp_RearO2_min:.float -0.5
ROM:0002C920 Closed_loop_fueling_target_comp_RearO2_max:.float 0.5

dj-phoenixx · **Posted:** Tue Aug 14, 2012 3:34 pm

Hi,
thank you very much

is it correct?

<scaling name="AF3 Lambda" units="Lambda Additive" toexpr="x" frexpr="x" format="%.4f" min="0" max="255" inc="1" storagetype="float" endian="big"/>

<table name="Closed_loop_fueling_target_comp_RearO2" category="Experimental" address="2C91C" type="2D" level="1" scaling="AF3 Lambda">
<table name="X" type="Static X Axis" elements="2">
<data>High</data>
<data>Low</data>
</table>
</table>

I can see only an -0,5 as an high map and an +0,5 as an low map.

td-d · **Posted:** Tue Aug 14, 2012 3:56 pm

I have found in the the earlier roms, the values are flipped around - min first, max second. So just change:

<data>High</data>
<data>Low</data>

to

<data>Low</data>
<data>High</data>

RomRaider

Rear O2 sensor closed loop fueling target limits

Who is online