O2 sensor tuning with Headers

[JBZ]

I?m sorry guys as I know this subject has been beaten to death but I cannot get a definitive answer or results on how to tune closed loop with headers to not only improve commanded vs actual (WB) but more so reduce the amount of random misfires I am seeing. I have read multiple threads and commented on a few but would love to get more opinions or options to try out. I have really only found one thread mentioning random misfires due to headers.

The things I?m seeing are quite a few random misfires and also my wideband does not agree with my narrow bands and I?m trying to close the gap. When I go into open loop my wideband reads about .95 lambda when commanding stoich so I know I need to adjust my MAF table but when I do the CL corrections will just adjust back to what my O2s think are stoich anyways...right? and in CL my LTFTs are always within + or - 2% other then idle its about +5% (well except after I get gas LOL).

When im in open loop I may get 2 random misfires every 5 minutes or so but in closed loop i see from 2-34 on a regular basis. About every 30 seconds to a minute it registers misfires. I am not logging individual cylinders just total misfires seeing as how it only happens in CL. Actually the least amount of random misfires I see while in closed loop are when all setting are factory. also I was holding the gear and throttle (2nd at around 4500 rpm in CL...NO PE) the other day trying to log my switching speeds at higher RPM's and got a P0300 cell pop up and service stabilitrac and 146 misfires...I can only figure it was because of my o2's.

Following other threads these are the changes I have made:

1st attempt:
Closed Loop Proportional Base vs. Airflow Mode (multiply x .95)
Closed Loop Integrator Delay vs. Airflow Mode (multiply x 1.1)

2nd attempt:
Closed Loop Proportional Base vs. Airflow Mode (back to factory settings)
Closed Loop Proportional Gain vs. O2 Error (multiply x 1.1)
Closed Loop Integrator Delay vs. Airflow Mode (back to factory settings)

3rd attempt:
Closed Loop Proportional Base vs. Airflow Mode (multiply x .9)
Closed Loop Proportional Gain vs. O2 Error (multiply x 1.1)
Closed Loop Integrator Delay vs. Airflow Mode (multiply x 1.1)

4th attempt:
Closed Loop Proportional Base vs. Airflow Mode (multiply x 1.1)
Closed Loop Proportional Gain vs. O2 Error (back to factory settings)
Closed Loop Integrator Delay vs. Airflow Mode (multiply x .9)

5th attempt:
Closed Loop Proportional Base vs. Airflow Mode (multiply x .8)
Closed Loop Mode vs. Airflow (decreased to DOD settings for lack of better option or knowledge)
Closed Loop Integrator Delay vs. Airflow Mode (multiply x 1.4)

6th attempt: (current)
Closed Loop Proportional Base vs. Airflow Mode (multiply x .8)
Closed Loop Integrator Delay vs. Airflow Mode (multiply x 1.42)


I have also done a crank relearn for the random misfires

I absolutely understand what Proportional and integral controllers and settings are and maybe im chasing a ghost here but really don?t understand in this application how to adjust for the extended time it takes for the O2?s to respond to errors.

I have not messed with Closed Loop Mode vs. Airflow but have read that this could affect the gains.

the biggest thing I cannot seem to figure out is what settings are the actual transport delay settings ?or am I missing those settings in my tune. I really want to understand what are the actual tables that the PCM use for o2 sensor delay??? I just assumed it was the O2 Proportional Gain vs. O2 Error, but am not exactly positive.

I have seen others say they can narrow in their commanded vs actual while in CL by making the O2 changes and/or changing the switch voltages (which I have not tried yet) but then I have also read (and commented) on threads about switch points not changing actual switching in the E38 or 67?s.

Im kind of at a loss here and any help would be greatly appreciated. Thanks guys. log attached is at current settings.

[GHuggins]

I believe your making very small changes off of numbers I posted a long while back for numbers to start out with? Not all headers are made the same and every little thing about the headers will change how the settings effect things... I.E - different sized pipes, different O2 locations, different O2 install angles, O2's installed in one header pipe vs the collector or the better location - further down stream in a solid diameter exhaust pipe, and on and on... I know ARH typically installs them at an angle to help with codes and "seeing" the most possible exhaust mix - if that makes sense?

Tell ya what - give this a try - no guarantee... Multiply proportional airflow mode by .8 / integrator by 1.4 and then shift your airflow mode table to the left using the numbers shown in the replacement ECM thread... Again, this is just a test - I have been shifting this table, but never realized it did this to the O2's - I personally would like an "exact" step by step on how to best adjust it?

[JBZ]

Awesome thank you i will give it a shot and post up some results.

I did infact take the advise you had given and tried pretty much every set of p and i gains you suggested to others of couse. It seems like you have this figured out on your zl1 and i guess i assumed same engine and headers shouldnt be that much different (i know their not exactly the same) but after trying none have really produced the results i am looking for...which is mostly to reduce the erroneous random misfires and line up my cl and ol commanded fueling.

Its really to bad there is not a defined and well discussed formula for o2 placement change tuning from header installs as its probably one of the most common mods on all platforms. I would venture to say after messing with this for weeks now that probably most people running around with headers are probably in the same boat and have no idea that what their o2s are reading probably are not right. Now...is it really that big of a deal? No probably not, its still close enough that in most cases it works fine, but i really want my car to drive better than GM intended and personally feel that this is the first step in tuning. I have been making quite a few other chanes for months now, all for the better but if i can get this right i will more or less be starting over.

Thanks for your suggestions and keep um coming

[GHuggins]

That's actually one of our shop cars in my sig, which I need to eventually update BUT, yes I do try to adjust in O2's for all customers when possible... Just some more thoughts on yours - With your setup are you at least running .035" plug gaps? AND most importantly are your plug wires setup with heat wraps - headers tend to cook them and can cause actual misfires...

I'm hoping others will chime in on this - I know other people out there have a better understanding of these tables than I

[JBZ]

That's actually one of our shop cars in my sig, which I need to eventually update BUT, yes I do try to adjust in O2's for all customers when possible... Just some more thoughts on yours - With your setup are you at least running .035" plug gaps? AND most importantly are your plug wires setup with heat wraps - headers tend to cook them and can cause actual misfires...

I'm hoping others will chime in on this - I know other people out there have a better understanding of these tables than I

Im running .032" gap and no i dont have wrap on my plug wires. I really dont think its actually real misfires because when i go into open loop i get no where near as many...2 misfires every couple minutes at most. but there is absolutely no reason not to actually check out my wires and get the heat wraps on them if they show any signs of burning through.

[JBZ]

well I multiplied proportional airflow mode by .8 multiplied integrator by 1.4 (other then idle cell, kept it stock) and then shifted my airflow mode table to match my DOD for lack of a better formula, which lowered it by a little, see in the attached pictures. so these are the things I noticed in this mornings drive. first is the my rear o2s were more in line with my fronts...meaning, there has always been a significant lag between fronts and rears and although it is still there the gap has narrowed significantly. I did notice that my random misfires went down a bit but of course are still showing up. something to take note of, I don't get the random misfires at either idle or WOT...ever (for what ever that's worth), only at off idle and light cruising NOT under load. i will pull my wires off tonight (if i can get some time) and check them to make sure there are no issues.

I also know that somehow i will need to get my switching leaner and i assume that i can just move my switch points to lower then 450 mv (say 350 mv) because when i go into OL i am still commanding stoich or actually right under @ .99 lambda and the car is always rich by about 4 to 5 points running around .95/.94 lambda. the way i see it is...i need to adjust my MAF table and VVE tables to get this right, but don't want to do so until i get my O2s reading correctly because it will just add the unnecessary fuel right back in there. i would rather get the O2s switching right and have the PCM pulling fuel for a day or two, then take the averages and re-adjust my fueling then go into OL and verify that they are matching up correctly.

ive attached some screenshots of my current setting vs factory and a short log of this mornings drive with my channel config file so you guys can see what im seeing. also a screenshot of what i think i need to modify for getting switching lower can you guys confirm those are the tables that need to be changed. thanks for taking a look.

[GHuggins]

You might want to get it into PE a little sooner - at least a partial PE - You start to knock just a touch around your baro level... O2's look like they're pretty close at 50+mph then in spots below that, so I would say your getting pretty close... I didn't see a wideband to know what your actual air fuel is - so your saying you need to make your O2's leaner to be right? SO it's running or rather averaging too rich while in closed loop - or did I miss read that - it's just usually the other way around? You can try increasing the integrator more, just restart with the original value before increasing...

Yes, those are the correct O2 settings to change...

[JBZ]

You might want to get it into PE a little sooner - at least a partial PE - You start to knock just a touch around your baro level...

Probably not a bad idea, funny thing is this is the first time ive seen any kr in weeks. I thought i had it pretty much taken care of. I had my PE set to enable at 105 kpa i think, i can go back to factory 15 kpa but that seems low to me. Ill have to look at it a little closer, i didnt take notice and on my pad right now.

I didn't see a wideband to know what your actual air fuel is - so your saying you need to make your O2's leaner to be right? SO it's running or rather averaging too rich while in closed loop - or did I miss read that -

I do have a wideband just not wired in...i regret i got the cheap mpvi .

What i meant was the in closed loop my trims are good and fueling is good swinging around stoich. When i go into open loop it holds around .94/95 lambda. So something is a miss between open and closed loop.

All my tuning had been focused around closed loop and wot. Now im trying to get my open loop fueling to be as close as possible to my closed loop...which makes me think i need to change my switch points because the headers and relocated o2s are making it run falsely leaner then in open loop...i think??? Hope that makes sense

edit: I have been trying to re-create that KR event (who ever says that right?) this morning and cannot get any. I know every now and then (week or two) I get around 1 degree or less between 1400 - 2000 rpms. but it is rare. I would like to never see another KR even again but am far better then the 6 degrees of KR the tuner left me with.

[schpenxel]

Try like 85kpa for PE enable and see how that does

You're assuming that absolutely nothing else changes between open loop and closed loop.. which I'm not sure is true

[JBZ]

Try like 85kpa for PE enable and see how that does

You're assuming that absolutely nothing else changes between open loop and closed loop.. which I'm not sure is true

Ok dropped it down to 85 kpa and richened it up at 1500 to 2500 by .015 give or take. Do you guys just suppose that there is a delay from the time my commanded shows a lower then stoich value on the scanner to the time it actually adds fuel?

Do you mind explaining in a little more detail "absolutely nothing else changes between ol and cl"? Hope i dont sound lika a smart ass i really want to know what you mean.

[sevinn]

Probably not a bad idea, funny thing is this is the first time ive seen any kr in weeks. I thought i had it pretty much taken care of. I had my PE set to enable at 105 kpa i think, i can go back to factory 15 kpa but that seems low to me. Ill have to look at it a little closer, i didnt take notice and on my pad right now.



I do have a wideband just not wired in...i regret i got the cheap mpvi .

What i meant was the in closed loop my trims are good and fueling is good swinging around stoich. When i go into open loop it holds around .94/95 lambda. So something is a miss between open and closed loop.

All my tuning had been focused around closed loop and wot. Now im trying to get my open loop fueling to be as close as possible to my closed loop...which makes me think i need to change my switch points because the headers and relocated o2s are making it run falsely leaner then in open loop...i think??? Hope that makes sense

edit: I have been trying to re-create that KR event (who ever says that right?) this morning and cannot get any. I know every now and then (week or two) I get around 1 degree or less between 1400 - 2000 rpms. but it is rare. I would like to never see another KR even again but am far better then the 6 degrees of KR the tuner left me with.

Get the AEM 30-0333 and you can use the wideband in HPT, or if your wideband is one of the ones supported via serial now you can do the same without pro.

[schpenxel]

Ok dropped it down to 85 kpa and richened it up at 1500 to 2500 by .015 give or take. Do you guys just suppose that there is a delay from the time my commanded shows a lower then stoich value on the scanner to the time it actually adds fuel?

Do you mind explaining in a little more detail "absolutely nothing else changes between ol and cl"? Hope i dont sound lika a smart ass i really want to know what you mean.

Sorry haven't been back on here since you posted

I don't know 100% everything about what changes, but I've never had fueling stay exactly the same between closed/open loop. I remember being extremly frustarted in the beginning stages of tuning.. I would get everything just right on the wideband in open loop, go back to closed loop and fuel trims would be horrible. They did settle down after some time, but there were always additional corrections needed. I know evap purge turns back on in closed loop but is disabled in open loop, so that is one thing that will mess with fueling some. There may be some others, I'm not totally sure. I just saw that you were going on the assumption that nothing changes other than fuel corrections, and I don't think that's the case

[JBZ]

Sorry haven't been back on here since you posted

I don't know 100% everything about what changes, but I've never had fueling stay exactly the same between closed/open loop. I remember being extremly frustarted in the beginning stages of tuning.. I would get everything just right on the wideband in open loop, go back to closed loop and fuel trims would be horrible. They did settle down after some time, but there were always additional corrections needed. I know evap purge turns back on in closed loop but is disabled in open loop, so that is one thing that will mess with fueling some. There may be some others, I'm not totally sure. I just saw that you were going on the assumption that nothing changes other than fuel corrections, and I don't think that's the case

I honestly didn't know what changes there were other then corrections, and didn't think about evap purge.

OK so what is the general consensus then on OL vs CL? so you say that your trims were bad after getting OL corrected. then after that and your cruising around in CL and see pretty bad trims did you clean it up in CL or leave it the same and let the PCM do the corrections it does?

also, is there any way you guys have found to get open loop actual to be tighter in with closed loop (using trims of course). so for instance...my closed loop trims (LTFT + STFT histogram) after 45 minutes of driving (under all driving conditions that keep the trims learning) are all showing + or - 2. they seem to be pretty tight. then I use the scanner controls and put it into open loop and my AFR (I know you guys cant see it...sorry) while commanding stoich goes down to .95/94 at idle...off idle light cruise it goes even richer (for arguments sake .90-.85) then once I am actually up to speed I raises back up to .94 again. let off the gas and it stays rich. if im in light throttle getting up to speed and give it a little more peddle (have a quick tip in) then it goes lean for a second (1.03) then settles back down to .94 once im in a steady state again.

sorry all that to say I think I have a lot of work to do and need to just invest in my HPT upgrade or the AEM...which looks pretty cool.

[sevinn]

I honestly didn't know what changes there were other then corrections, and didn't think about evap purge.

OK so what is the general consensus then on OL vs CL? so you say that your trims were bad after getting OL corrected. then after that and your cruising around in CL and see pretty bad trims did you clean it up in CL or leave it the same and let the PCM do the corrections it does?

also, is there any way you guys have found to get open loop actual to be tighter in with closed loop (using trims of course). so for instance...my closed loop trims (LTFT + STFT histogram) after 45 minutes of driving (under all driving conditions that keep the trims learning) are all showing + or - 2. they seem to be pretty tight. then I use the scanner controls and put it into open loop and my AFR (I know you guys cant see it...sorry) while commanding stoich goes down to .95/94 at idle...off idle light cruise it goes even richer (for arguments sake .90-.85) then once I am actually up to speed I raises back up to .94 again. let off the gas and it stays rich. if im in light throttle getting up to speed and give it a little more peddle (have a quick tip in) then it goes lean for a second (1.03) then settles back down to .94 once im in a steady state again.

sorry all that to say I think I have a lot of work to do and need to just invest in my HPT upgrade or the AEM...which looks pretty cool.

One thing that helped me get my fuel trims back under control was to make a custom math that combined and averaged both banks (I found the advice from a member here on the forums) instead of using the built in one and tuned using that. I'd guess you have some afr correction in a table that was missed somewhere.

[JBZ]

One thing that helped me get my fuel trims back under control was to make a custom math that combined and averaged both banks (I found the advice from a member here on the forums) instead of using the built in one and tuned using that. I'd guess you have some afr correction in a table that was missed somewhere.

so not the typical LTFT or STFT or LTFT+STFT vs MAF htz scale? or have I completely misread your post?

please do tell, I like making new charts

[sevinn]

It is the same format chart except instead of using the built in math you make a custom one that just averages both banks instead of using the generic sensor. I'll attach a screenshot when I get to a computer.

[JBZ]

It is the same format chart except instead of using the built in math you make a custom one that just averages both banks instead of using the generic sensor. I'll attach a screenshot when I get to a computer.

cool thank you

[sevinn]

None of these may make any difference at all, but I found my numbers to clear up using them vs the built in maths. Your results may vary .

Seeing this makes me realize I need to delete some old maths


Disregard the fact that I don't have ltft in this particular channel config

[JBZ]

thanks again, I do have one question, in the last screenshot what is the filter function. it looks like its cut off so I cannot see the rest of it.

[sevinn]

thanks again, I do have one question, in the last screenshot what is the filter function. it looks like its cut off so I cannot see the rest of it.

[50090.155] < 50 AND [50090.155.avg(500)] < 50 AND [50090.155.avg(-500)] < 50 AND [50020.114.slope(500)] = 0 AND [50020.114.slope(-500)] = 0

This is my attempt to get steady state data in part throttle ranges. There are probably better pids to use but this is what I setup a while back and it seems to work fine for me at least.

In words, this is throttle position less than 50, is less than 50 for the next half second and was less than 50 for the last half second and the slope of vehicle mph is 0 for next half second and was 0 for the previous half second. These results have to be met before a value will be entered in the histogram. It makes tuning drives take a little longer but I feel better about the data.