32bit lean burn

Freon · **Posted:** Sat Jun 21, 2008 10:25 pm

It doesn't quite make sense that you could make the signal from the rear O2 toggle at something besides 14.7 just by taking something electrical on it. I'm skeptical, but we can log rear O2 current easily enough to test it.

You've got me itching to try this myself now.

merchgod · **Joined:** Wed Mar 29, 2006 10:38 pm **Posts:** 4224

I think I've found a way to eliminate a/f correction and learning #3 for the 32-bit ECU. This is by modifying the final limits, which are hard coded values and are well outside the range of normal correction but setting the min/max to zero should eliminate any correction if I'm right. The actual limits appear to be variable. This won't make the next ECU update as it still needs to be tested by someone, however.

The standard SSM parameter 'A/F Correction #3' also has the incorrect conversion in romraider which is why it shows some ridiculous value (say -50%) -> this will be fixed in the next logger update.

williaty · **Joined:** Tue Jul 24, 2007 6:28 pm **Posts:** 214

Bill, #3 is the one that's from the rear O2 sensor correct? The NA crowd is having a bugger of a time trying to find ways to work with non-stock cats (or track pipes) without pissing the ECU off. I'll volunteer to test anything relating to this.

ride5000 · **Joined:** Thu Aug 03, 2006 10:40 am **Posts:** 564

themadscientist wrote:

ride5000 wrote:

nice mark. it looks like it might just work.

it's a bit easier for me with my 16 bitter.

I will see how it goes. The biggest problem is my WBO2 sensor died. I am just using the stock A/F sensor read through the Utec to see the changes in AFR.

funny you say that--my bosch sensor from my tuner pro just died a few weeks ago... i've got to get another one. i feel totally blind without it!

merchgod · **Joined:** Wed Mar 29, 2006 10:38 pm **Posts:** 4224

williaty wrote:

Bill, #3 is the one that's from the rear O2 sensor correct? The NA crowd is having a bugger of a time trying to find ways to work with non-stock cats (or track pipes) without pissing the ECU off. I'll volunteer to test anything relating to this.

yes #3 is the rear o2 (#3 because some Subarus have a 2nd front o2 sensor which would be #2). Send me a PM after I release the next ecu def update, sometime in July most likely.

vansickey · **Joined:** Fri Sep 14, 2007 6:13 pm **Posts:** 21

Freon wrote:

It doesn't quite make sense that you could make the signal from the rear O2 toggle at something besides 14.7 just by taking something electrical on it. I'm skeptical, but we can log rear O2 current easily enough to test it.

You've got me itching to try this myself now.

It's actually not that difficult, and if you understand how the sensor works, all we would be doing is shifting the voltages +... This however is just a cheap fix, and if you go all the way to 16:1 or so it may be impossible without a programmable O2. If for instance the ecu still needs to see a cycle of rich and lean, it may be difficult to tune, but if all it needs to see is say a voltage in the range of 350 - 550 mv, no problem. Of course all this would have to be trial and error testing for you guys with 32's.

merchgod wrote:

I think I've found a way to eliminate a/f correction and learning #3 for the 32-bit ECU. This is by modifying the final limits, which are hard coded values and are well outside the range of normal correction but setting the min/max to zero should eliminate any correction if I'm right. The actual limits appear to be variable. This won't make the next ECU update as it still needs to be tested by someone, however.

The standard SSM parameter 'A/F Correction #3' also has the incorrect conversion in romraider which is why it shows some ridiculous value (say -50%) -> this will be fixed in the next logger update.

is this similar or the same on a 16 bit?

Freon · **Posted:** Sun Jun 22, 2008 3:27 pm

vansickey wrote:

It's actually not that difficult, and if you understand how the sensor works, all we would be doing is shifting the voltages +...

Well, not exactly. It's not that simple. The sensor's chemical properties make it toggle output voltage as it the sensor detects oxygen. It toggles from about 0.1v to 0.9v, very quickly as it detects oxygen or not. It's just an on/off signal. Moving the voltage at which it toggle doesn't help, you need to change how it reads. You have to make it somehow toggle at 16.2:1 AFR. The sensor itself is incapable of detecting a change in this AFR range. I.e. from 16.0 to 16.4 AFR the sensor doesn't do anything. It reads "full lean" from 14.8:1 to infinity.

This may be moot if the affects can be removed via reflash.

Jon [in CT] · **Posted:** Sun Jun 22, 2008 10:52 pm

Freon wrote:

The sensor's chemical properties make it toggle output voltage as it the sensor detects oxygen. It toggles from about 0.1v to 0.9v, very quickly as it detects oxygen or not. It's just an on/off signal. Moving the voltage at which it toggle doesn't help, you need to change how it reads. You have to make it somehow toggle at 16.2:1 AFR. The sensor itself is incapable of detecting a change in this AFR range. I.e. from 16.0 to 16.4 AFR the sensor doesn't do anything. It reads "full lean" from 14.8:1 to infinity.

This may be moot if the affects can be removed via reflash.

I agree. There's no electrical device which can stop the sensor from switching between 0.1V and 0.9V as AFR moves across stoichiometric from lean to rich.

I agree, too, that the ability to eliminate any fuel trim associated with the rear sensor solves the problem.

ride5000 · **Joined:** Thu Aug 03, 2006 10:40 am **Posts:** 564

well, the o2 sensor itself is actually a current output device. that current works against a resistance to generate a voltage.

the big question would be whether or not the o2 sensor driver self-compensates/zeros any offset during power up, or possibly after an ecu reset. if not, level shifting the current output is actually not that difficult to do.

in any event, it's MUCH more elegant to be doing it in the code vs. discreet components, so i wholeheartedly endorse that approach!

vansickey · **Joined:** Fri Sep 14, 2007 6:13 pm **Posts:** 21

Jon [in CT] wrote:

Freon wrote:

The sensor's chemical properties make it toggle output voltage as it the sensor detects oxygen.

Not As simple as that, it's not just on off, it's progressive.

It toggles from about 0.1v to 0.9v, very quickly as it detects oxygen or not. It's just an on/off signal. Moving the voltage at which it toggle doesn't help, you need to change how it reads.

*Moving the voltage to the point that crosses the ecu threshold of rich to lean would help.

You have to make it somehow toggle at 16.2:1 AFR.

* I did mention a programmable one as well as others.

The sensor itself is incapable of detecting a change in this AFR range. I.e. from 16.0 to 16.4 AFR the sensor doesn't do anything. It reads "full lean" from 14.8:1 to infinity.

* The sensor DOES create votlage at much leaner than 14.8 and can be offset to facilitate the 500mv mark.

This may be moot if the affects can be removed via reflash.

* I agree, best solution. I was only offering a piece of info I've worked on with vehicles I can't reflash to essentially do the same thing.

agree. There's no electrical device which can stop the sensor from switching between 0.1V and 0.9V as AFR moves across stoichiometric from lean to rich.

* I am not talking about changing the sensor output, just changing what the ecu sees from the sensor.

I agree, too, that the ability to eliminate any fuel trim associated with the rear sensor solves the problem.

What about the o2 similators?

LGT-3-6 · **Joined:** Wed May 24, 2006 8:11 pm **Posts:** 187

I haven't checked in a while but I'm pretty sure my rear o2 maxes out rich already with one or more cats removed.

themadscientist · **Posted:** Thu Jun 26, 2008 9:52 pm

Well first tank of lean burn went from an average of 310 miles to the light to 335 miles. But that is with 6 ecu resets and 5 reflashes. Hopefully the CEL I have been fighting has been solved and this tank will be a one to one comparison to non lean burn mapping.

WolfPlayer · **Joined:** Mon May 01, 2006 2:15 pm **Posts:** 66

I'd be really interested in seeing what the EGTs are with a 16:1 lean burn and cruising up a long grade. For example, I live in an area where I have to travel over some mountains. I can have sustained 2.3-2.5v MAF while going uphill. This is still closed loop. I have to think that EGTs at 16:1 on a long uphill grade would be really high. But, who knows.

t

Jon [in CT] · **Posted:** Fri Jul 04, 2008 4:31 pm

WolfPlayer wrote:

I have to think that EGTs at 16:1 on a long uphill grade would be really high. But, who knows.

They'd be even higher at a stoichiometric AFR, assuming he's adjusted his spark advance properly to account for the slower combustion speed of the 16:1 AF mixture.

http://forums.nasioc.com/forums/showthr ... ost1874209

themadscientist · **Posted:** Sat Jul 05, 2008 1:24 pm

I am one of the few people that runs what most people consider high EGT under boost anyway. "Most people" consider anything over 1600F high. I just laugh at that as long 5th gear pulls I see over 1700F. Trust me the fastest cars don't stick to the 1600F limit.
In short I don't worry about EGT at cruise. In my experience with lean burn with the Utec EGT actually drops off the leaner you get from stoich.
I can not recommend anyone scaling the A/F sensor the way I have done it on just the stock ECU. I have no idea how it will effect your high load fueling since I am running a Utec in SD mode.

RomRaider

Forum rules

32bit lean burn

Who is online