Have we been doing autove wrong?

[GMC383]

So I have been working on my autove tune with lc-1 and I just don't understand exactly it is supposed to work. We are supposed to be putting the ECU in OLSD and when we do we run off of B3506 and I get that but what I don't understand is that above 60kPa we start changing the commanded ratio which means the LC-1 that is calculating the BEN for 14.62 is now being fed bad data from the ECU. If you log commanded air fuel ratio (GM.AFR) you can see it moves around when it should be steady. Any input?

I have attached my tune files with corresponding log files.

[LastCall]

I'm on my phone, so I can't open the tunes or logs. The BEN will account for changing AFR based on your open loop table or PE modifier. You should be commanding a richer mixture above 75-80 kpa. And the BEN will account for that.

[joecar]

So I have been working on my autove tune with lc-1 and I just don't understand exactly it is supposed to work. We are supposed to be putting the ECU in OLSD and when we do we run off of B3506 and I get that but what I don't understand is that above 60kPa we start changing the commanded ratio which means the LC-1 that is calculating the BEN for 14.62 is now being fed bad data from the ECU. If you log commanded air fuel ratio (GM.AFR) you can see it moves around when it should be steady. Any input?

I have attached my tune files with corresponding log files.Yes, OLSD (you should see zero LTFT's, and a MAF DTC).


Yes, B3605.

The BEN is not calculating the BEN for 14.62... take another look at how it is defined (using LC-1 as example): {CALC.AFR_LC11.AFR}/{GM.AFR}

GM.AFR comes from whatever the commanded fueling is (i.e. from B3605 when PE is not actove, and from the richer of B3605 and B3618 when PE is active).


So yes, GM.AFR moves around, but then BEN pid uses this in its calculation.

[joecar]

I prefer to use GM.EQIVRATIO instead of GM.AFR because EQIVRATIO does not depend on the stoich AFR of the fuel you're running

( for example, tune file stoich is typically set to 14.63 or 14.68, wideband stoich is set to 14.7 for some and 14.5 for others, fuel ranges from 14.1 to 14.7 )


I use this BEN pid (using V2 wideband serial comms): {EXT.WO2LAM1}*{GM.EQIVRATIO}


both EQR and LAM are independent of fuel's stoich AFR (you still have to set B3601 to match the fuel you're running).

[joecar]

In your tune files, you have B3605 at EQR 1.1338 and B3618 at EQR 1.0039...

both of these are too lean for WOT...

set B3618 to EQR 1.18.

[GMC383]

I have two different tunes, one with the correct B3605 and one with 14.62 AFR all the way across. I understand what you are saying now. I didn't realize that it considers the commanded AFR. That would explain why I got the same results either way.

[GMC383]

so are you telling me to set B3618 to 1.18 even though I am still doing OLSD tuning?

[joecar]

so are you telling me to set B3618 to 1.18 even though I am still doing OLSD tuning?Yes, EQR 1.18 (which would be AFR 12.4)... to be sufficiently rich/safe at WOT load.

Or, you could set the high load columns of B3605 to EQR 1.18.

But, I'd do B3618... so when you go back to CL you will be safe.

[GMC383]

Ok I now understand how the BEN will take into account the commanded fuel. Maybe the autove tuning tutorial should be corrected.

[Tinbender59]

Ok I now understand how the BEN will take into account the commanded fuel. Maybe the autove tuning tutorial should be corrected.

lol YA think?? I have been fighting similar issues, I'm not an engineer or a programmer, so I HAVE to take this stuff at face value!!

[joecar]

Take note of something I said...

Yes, EQR 1.18 (which would be AFR 12.4)... to be sufficiently rich/safe at WOT load.

Or, you could set the high load columns of B3605 to EQR 1.18.

But, I'd do B3618... so when you go back to CL you will be safe.

This would only be true if your fuel's stoich AFR was 14.63...


this is why we should avoid taking AFR.

[joecar]

I know the AutoVE and other tutorials need to be updated... but there's never enough time.

[Tinbender59]

I know the AutoVE and other tutorials need to be updated... but there's never enough time.

:thumb_yello::thumb_yello::thumb_yello: BEEN THERE

[Tre-Cool]

Anyone have the original Document's & i can update for the newbies. :-)

[Blacky]

Anyone have the original Document's & i can update for the newbies. :-)


Shoot me an email
(paul@efilive.com)
And I'll send you back the word doc.

Regards
Paul

[Tre-Cool]

working on this document at the moment. i've added an additional 4-5 pages of info/instructions and im probably about half way done. It should be pretty thorough for everyone and give some simple explanations on tuning via ben's etc.

[joecar]

Should probably update the BEN's to include the V2 serial comms pids, and to use EQIVRATIO and/or LAMBDA instead of AFR...

and PE should be set properly rich.

I would like to add some stuff after you're done.

[Tre-Cool]

Should probably update the BEN's to include the V2 serial comms pids, and to use EQIVRATIO and/or LAMBDA instead of AFR...

and PE should be set properly rich.

I would like to add some stuff after you're done.

All over it like a fat kid in a candy store. I've added nearly 8 new pages of info/instructions all ready excluding the toc page.

20281

That's what i have so far.

[joecar]

TOC looks good :cheers:

[Tre-Cool]

I've added the how to create calculated pids, simply because it's probably not worth the efilive guys time to add them in the 7.5 software & well we know how bloody simple it to create them.

I have a whole bunch i use for different controllers, i.e e40, e38/e67, e83b.

I haven't looked into how we create them for V8 yet though.

[cindy@efilive]

We are scheduled to build beta software later this week. If you think your updates are ready for inclusion, please send it through.

Cheers
Cindy

[Tre-Cool]

I emailed Paul the updated document on the weekend. Hopefully he or someone else is just doing a final check.

If we can include some calc Pids. I'll edit mine and sent them through for inclusion.

[Mitco39]

Since this topic is here and I am doing one of my first VE tables.

Do any of you guys find the BEN amount to be to high one way or the other? IE if I go through on the dyno and hit every cell 50+ times and take this and multiply the VE table by the BEN amounts my VE table goes to a jagged mess. As a result what I started doing (after what must of been 6-7 hours of trying to do it the way that it should be done) I would look over the BEN table and add about half of the difference to areas of the table this allowed me to get +-5% but it was much more involved as I was no longer pinpointing specific cells but the about a 4X4 grid and making changes to that. I saw that Joecar posted up how he is calculating his BEN pid which I may try to see what happens.

When I was done I logged the MAF airflow vs the calculated VE airflow and they are pretty damn close, I just would have loved to do the BEN map in its entirely and get it in 1-2 shots not 20, lol.

This is on straight 91octane fuel.

I have only been a diesel guy until recently, but I am enjoying learning this whole new world :)

[joecar]

Since this topic is here and I am doing one of my first VE tables.

Do any of you guys find the BEN amount to be to high one way or the other? IE if I go through on the dyno and hit every cell 50+ times and take this and multiply the VE table by the BEN amounts my VE table goes to a jagged mess. As a result what I started doing (after what must of been 6-7 hours of trying to do it the way that it should be done) I would look over the BEN table and add about half of the difference to areas of the table this allowed me to get +-5% but it was much more involved as I was no longer pinpointing specific cells but the about a 4X4 grid and making changes to that. I saw that Joecar posted up how he is calculating his BEN pid which I may try to see what happens.

When I was done I logged the MAF airflow vs the calculated VE airflow and they are pretty damn close, I just would have loved to do the BEN map in its entirely and get it in 1-2 shots not 20, lol.

This is on straight 91octane fuel.

I have only been a diesel guy until recently, but I am enjoying learning this whole new world :)
Do you apply the low cell count filter/button...?

Do you apply the transient filter (i.e. to filter out any data where TP is moving more than say 5% in 100 ms)...?


Which BEN pid are you using...?

Does your wideband waveform look sensible (is the WB working correctly)...?

Post log files.

[Mitco39]

Do you apply the low cell count filter/button...?

Do you apply the transient filter (i.e. to filter out any data where TP is moving more than say 5% in 100 ms)...?


Which BEN pid are you using...?

Does your wideband waveform look sensible (is the WB working correctly)...?

Post log files.

I do have the filter applied and since I am on the dyno I will sit in one cell for atleast 50+ hits in a row.

I am using the BEN_LM21 - Base Efficiency Numerator 1 - LM2 LS1 Style.

My wideband is a LC2 and I checked the AFR that its own software displays agains what the AFR_LM21 displays and its within .25 of an afr. Not 100% perfect but very close. The wideband looks just fine. I dont have my logs from when I was calibrating my VE but here is a WOT pull on the dyno, you can see the MAF/VE table working in OL mode when it switches to PE mode.

20316

I can get some longer ones if you need to see anything else to possibly point me in the right direction

Thanks so much

[joecar]

I emailed Paul the updated document on the weekend. Hopefully he or someone else is just doing a final check.

If we can include some calc Pids. I'll edit mine and sent them through for inclusion.I'm reviewing it, and adding some stuff in...

I have not gone thru it in detail yet, but here are my notes:




Tunetool:
- make sure fueling units are EQR and not Lambda (to avoid confusion).
- make sure to set B3605, B3647, B3618 in EQR units, not Lambda.


OL STFT:
B4206 = disable
B3647: don't allow any cells to be stoich.

PE:
B3618 = EQR 1.175 (or even EQR 1.18 to be safe)
B3616 = 65% below 3200 rpm, 35% above (to make sure PE activates).


Single Wideband:
CALC.WO2BEN = EXT.WO2LAM1 * GM.EQIVRATIO

Dual Wideband (select leanest on the fly):
CALC.WO2BEN = (EXT.WO2LAM1 > EXT.WO2LAM2 ? EXT.WO2LAM1, EXT.WO2LAM2) * GM.EQIVRATIO



CALC.WO2BEN is what we've called this pid in the Calc.VET and Calc.MAFT threads.

Also, in Calc.VET and Calc.MAFT we've adopting a convention in terminology:
- use EQR when talking about commanded fueling.
- use Lambda when talking about wideband measurement.

[joecar]

I do have the filter applied and since I am on the dyno I will sit in one cell for atleast 50+ hits in a row.

I am using the BEN_LM21 - Base Efficiency Numerator 1 - LM2 LS1 Style.

My wideband is a LC2 and I checked the AFR that its own software displays agains what the AFR_LM21 displays and its within .25 of an afr. Not 100% perfect but very close. The wideband looks just fine. I dont have my logs from when I was calibrating my VE but here is a WOT pull on the dyno, you can see the MAF/VE table working in OL mode when it switches to PE mode.

20316

I can get some longer ones if you need to see anything else to possibly point me in the right direction

Thanks so muchYes, get a longer one.

[Mitco39]

Yes, get a longer one.


Yeah again this was the only log I had after I turned back on the trims and started to attach the timing, thats why you see it in CL during the non PE section of the log. Ill turn everything back off and fill out my VE map on the dyno and send the whole log, as soon as I get time to get back on the dyno.

I believe I have it pretty close but I would appreciate your advice.

Thanks so much

Mitch

[joecar]

Hi Mitch,

CALC.AFR_LM21 is different than GM.AFR by a factor of 1.2 or greater...

are you correcting the VE table (B0101)...?

if so, did you fail the MAF (do this by causing a MAF DTC to show up immediately, see P0101,2,3 in C6001, and see C2901,2,3)...?

also, you have to disable CL (B4205), LTFT's (B3701), OL STFT's (B4206 and/or avoid stoich in B3605/B3647) during correction.

[Mitco39]

Hi Mitch,

CALC.AFR_LM21 is different than GM.AFR by a factor of 1.2 or greater...

are you correcting the VE table (B0101)...?

if so, did you fail the MAF (do this by causing a MAF DTC to show up immediately, see P0101,2,3 in C6001, and see C2901,2,3)...?

also, you have to disable CL (B4205), LTFT's (B3701), OL STFT's (B4206 and/or avoid stoich in B3605/B3647) during correction.

Joecar,

Since reading your initial post this morning I have been searching around here and found the thread I am assuming you were referring to..

https://forum.efilive.com/showthread.php?15236-Calc-VET-correcting-MAF-and-calculating-VE-(in-single-log)

I started to work through it but it appears that efilive has probably change their format since this was first drawn up, and as a result I am been trying to incorporate the changes in that file into what I assume is the correct file (sae.generic.txt, as opposed to calc_pids.txt). Basically trying to work through it and set it up the way that was dictated in the above thread.

I did get the MAF to error out, and confirmed I had no trims working (I logged for STFT and the LTFTs and confirmed they stayed at 0. I switched over my VE units from %. Running out of time today to fight with it much more. Biggest thing I am a little unsure about is how the two txt files relate, as just changing the calc_pids file did not populate the PID list with those required calcs.

Thanks so much,

Mitch

[joecar]

Joecar,

Since reading your initial post this morning I have been searching around here and found the thread I am assuming you were referring to..

Calc-VET-correcting-MAF-and-calculating-VE-(in-single-log) (https://forum.efilive.com/showthread.php?15236-Calc-VET-correcting-MAF-and-calculating-VE-(in-single-log))

I started to work through it but it appears that efilive has probably change their format since this was first drawn up, and as a result I am been trying to incorporate the changes in that file into what I assume is the correct file (sae.generic.txt, as opposed to calc_pids.txt). Basically trying to work through it and set it up the way that was dictated in the above thread.

I did get the MAF to error out, and confirmed I had no trims working (I logged for STFT and the LTFTs and confirmed they stayed at 0. I switched over my VE units from %. Running out of time today to fight with it much more. Biggest thing I am a little unsure about is how the two txt files relate, as just changing the calc_pids file did not populate the PID list with those required calcs.

Thanks so much,

MitchOk, that is the Calc.VET procedure... please proceed as the Calc.VET thread directs you to, and please disregard what I said in post #29...

( I wrote post #29 thinking you were doing AutoVE (which requires MAF, CL, LTFT, STFT all disabled )


Calc.VET is designed to get your MAF corrected and a matching VE calculated from it... you can get it in a single pass, but you have to avoid large throttle delta changes, and you have to apply the transient filter...

on the engine dyno we can usually get MAF/VE done like this in one pass, usually getting hundred of hits per cell of the cells that are reachable, then we do another pass to verify.


Just be careful at high load that you don't run lean and don't over-advance ignition timing.

[joecar]

I'm wouldn't be very comfortable if the wideband is not reading spot on... you have to somehow check it's accuracy...

e.g. in CL the wideband should report Lambda 1.00 (or whatever stoich AFR it assumes).

[Mitco39]

I think we are both confusing ourselves with what is going on so let me start from the top.

To start I did aim at doing the AutoVE tables to do this I
-Disabled all the trims by flipping all the required switches (logged STFT and LTFTs to verify they were not changing)
-Disabled the MAF causing the P0101-3 codes by setting the frequencies so that it would fail immediately.
-Set PE mode to .80 lambda (not sure what the EQ would be for that as I just recently switched my PE tables over to EQ)
-Set PE so that it would kick on at 20-30% throttle basically always running PE while on the dyno
-Did my best to follow the LM2 instructions in importing the analog value into the scan tool. (I checked it agains the serial connection I have with a direct connection to the laptop to ensure it was close) However its not bang on so I do have to address that and figure out what is going on. I followed the instructions here to set this all up, tweaking it a bit to work with the newer version of efilive than when the document was made up - http://www.innovatemotorsports.com/resources/efilive_tutorial.php.
-From here I setup a BEN table (with the filters setup to remove throttle transitions and low ect values as per one of the other guides I found).

Now I set up the dyno and hit a good chunk of the cells 50+ times making sure to keep the throttle consistent and the load as constant as possible on the dyno. I would set the dyno to hold a certian speed and lay into the throttle to load it up keeping the rpm the same and shuffling through all the cells. From here I filled out a pretty good table and when I went to copy with labels and paste and multiply the ben table over it would come out looking all jagged, even though the BEN table was all +-10 throughout so it was close. I didnt get a chance to go back and get a log of this today as I had some cummins come into the shop with issues, tomorrow I hope to get back at it.

Anyways what I resorted to doing was adding 1-2% at a time in the regions of the VE table that needed it as per the general values in the BEN MAP via logs, smoothing as I went and keeping everything "neat" this got me to having a BEN table about +-.05 throughout the log. Semi satisfied with this (I now have about 15 hours to get to this point and want to move on) I started on the MAF (resetting the changes made above), changing cell B0120 to 800 rpm and forcing the truck to run just the maf. Made another map and logged the BEN value agains the maf frequency to see where in the maf table needed to be tweaked. After this was done I logged The Airflow from the MAF agains the calculated airflow from the VE table just as a check and they were bang on with the exception of higher up in the load which I attribute to not 100% getting the VE right.


This is where I stopped and after reading your posts started looking up what I could possibly be doing wrong to cause the BEN table to screw the VE table by so much. I found that if you use the VE% you must input the engine size into the scan tool. If this is correct then that could explain the BEN having such a dramatic effect on the VE since I am working with a 8.1L truck. I didnt get a chance to test this theory yet, but sounds plausible from I been reading on the few threads that I stumbled onto after searching on here a bit more.


=====

I would like to know if that is a valid method of tuning the VE and the MAF. Tomorrow morning I will spend the time to go through all the VE threads I can find until I can get this sorted in my head. I see there is a bunch more you can do with it and I will need to do some more reading on EQ vs Lambda Vs AFR and how it relates to importing a wideband and making sure all the units are correct.



I do very much appreciate your help and sorry to highjack the thread but it is somewhat along the same lines.

Mitch

[joecar]

...

I found that if you use the VE% you must input the engine size into the scan tool. If this is correct then that could explain the BEN having such a dramatic effect on the VE since I am working with a 8.1L truck. I didnt get a chance to test this theory yet, but sounds plausible from I been reading on the few threads that I stumbled onto after searching on here a bit more.
Mitch,

Yes, for VE% you have to enter the correct displacement in the tune and in the scantool.

I prefer to use g*K/kPa, gives me a more direct feel for cylinder airmass in grams.

But: this would not cause the VE table to turn out jagged... can you post some screenshots of this.


It depends on the roughness of the VE table...
large + or - spikes cannot physically exist (flatten them out)
ripples (row-wise or column-wise) can exist (resonance due to intake or exhaust runner lengths)
other roughness can be ok
un-hit cells will have to be eyeball-extrapolated/interpolated.




I would like to know if that is a valid method of tuning the VE and the MAF. Tomorrow morning I will spend the time to go through all the VE threads I can find until I can get this sorted in my head. I see there is a bunch more you can do with it and I will need to do some more reading on EQ vs Lambda Vs AFR and how it relates to importing a wideband and making sure all the units are correct.
Sounds like you have been doing it right (paying attention to VE without MAF and MAF without VE)...

and you have been loading up the engine by holding rpm and hit it with varying MAP... loaded engine produces better BEN.


For LS1B I like to use GM.EQIVRATIO instead of GM.AFR...
with serial wideband, this means my BEN is {EXT.WO2LAM1} * {GM.EQIVRATIO}
for analog wideaband, say LC-1, this would become {CALC.AFR_LC11} * {GM.EQIVRATIO} / 14.7




I do very much appreciate your help and sorry to highjack the thread but it is somewhat along the same lines.

MitchNo worries (how else are we to discuss and learn stuff) :cheers:

[Mitco39]

Ok so I got some time on the dyno yesterday and ran into similar issues. I will try and document exactly what I did to make it as clear as possible for you to see where I may have went wrong.

I started with properly calibrating the wideband input into the scan tool. I checked it against innovative's own software and a direct serial connection to my computer, I took a screenshot of this process.

20323

I had to tweak the offset from the desired equation which I assume is to voltage drop because I am using a fairly long lead from the O2 controller. Wither way you can see on this screenshot that the calulated afr in the scan tool matches the logworks displayed afr. I checked over as much of a range as I could and it seemed very good throughout, meaning the slope of the equation (IE 3) held up.

----

I then setup the Map and applied the following filter

20325

From here I loaded in the attached tune that I used while creating that log. I hit a majority of the cells 50+ times while keeping the engine in as steady state as possible (I dont know how guys do this without a dyno). From here I got the following BEN map which on its own looks ok. However on the right side of the screenshot you will see what those changes did to the VE table. However the loaded region of the map looks not to bad. Is it just that this method does not work well on a non loaded engine and you must interpolate and blend as much as possible? I didnt go as far as to start the process again and see what the BEN map looked like afterwords (although now that I am sitting here typing this I wish I would have).

20324

I think my mistake looking at this now could have been that I would take those high correction factors into the VE table and smoothed them which pulled up all the cells around it. Ill have to see because anything over a 8PSI MAP load appears to be ok. Let me know if this could be what is going on.


Again thanks so much! I am enjoying learning the gas side.

Mitch

[joecar]

I'll at your filss in more detail later today...

but try this: reduce your VE table (the one before you applied that map) by 15% and try again.

[joecar]

BTW: also set these:
B3609: 0.020 s.
B3647,B3649: change stoich cells to slightly different value, i.e. not EQR 1.00, but rather EQR 1.01.

[joecar]

Also, set logging mode to Stream-Fast (you want fastestsample rate).

[Mitco39]

I will make all those changes next week and post up another log with the results.

Again thanks so much, if you ever start tuning diesels let me know! haha.

[joecar]

No worries... :)

[Mitco39]

So I hit the same sort of crap I have been dealing with again. As per your instructions I pulled 15% from the VE table made the change to B3609, couldnt find 3647 or 3649 however. Now the truck is running so lean it does not register on the wideband. Its the same thing I was dealing with when changing my VE to begin with, All it seems to provide me with is where needs more air and where needs less, how much seems to turn into trial and error.

Just to check the O2 sensor (because it seemed odd) I flashed the Closed loop file into the truck and sure enough it was right around 1, bouncing slightly higher and lower as the sort term trims do their thing. So if its working there it has to be working in the other file.

So I guess why does this happen? BEN was showing roughly 0.85, but you change the VE by the 15% and it goes crazy the other way.

Again the only change to the tune from the last one I posted is 15% less VE and B3609. Thats it. I didnt get into it to much because it was sooooo lean.

[Mitco39]

Ill post the screen of the map just to save you opening the log. 20336

[Mitco39]

I went back to the original VE table that I first posted up, did a log changed it by the (roughly) 10-15% only the cells I hit and now I am sitting at 1.4-1.5 and its to lean to load up. I must be missing something.

[joecar]

...

So I guess why does this happen? BEN was showing roughly 0.85, but you change the VE by the 15% and it goes crazy the other way.

Again the only change to the tune from the last one I posted is 15% less VE and B3609. Thats it. I didnt get into it to much because it was sooooo lean.

This is very strange...

[joecar]

Ill post the screen of the map just to save you opening the log. 20336This shows VE is off quite a bit... something strange is going on here.

[joecar]

Try this:

in scantool:
- on PIDs tab, on MAP pid go rightclick->Metric.
- goto map properties, change MAP axis to match Metric (goto VE table in tunetool, in upper-left corner go rightclick->Copy-with-Labels, goto scantool map properties, goto Column tab, click Paste Labels).

and try again...


this is very puzzling... your BEN indicated 0.85 correction factor, so we removed 15%, and then it goes over-lean...

for sanity check, make sure you did -15% and not -15.

[joecar]

When I open your log file, I see strange values in CALC.AFR_LM21... and CALC.BEN_LM21 does not look anything close to the map you posted above.



In scantool go Help->About and see what build version it says.

[Mitco39]

Yeah I agree, this is what started this whole mess because I figured I had a good grasp on the general process around setting up the VE and MAF tables but it just would not work.

-I will try switching the MAP PIDs to metric this afternoon when I get back on the dyno to do a log.

-I can confirm that I removed 15% I double checked this and throughout the whole process I know I would have caught myself if I did something like that. I went and grabbed a screen of the exact VE table that was ran for that posted log.
20338

- Build version is 311

- I had to change that PID as it was not mapped out correctly in the sae.generic.txt file when I first started. On your log can you confirm that the voltage is atleast displaying the same? If the voltage is the same but your calc isnt wouldn't that point to a incorrect setup? You know way more about this than I do. I will repost the screen where it shows the modification I made to the pid calc, I know it was completely wrong when I first started fighting with this last week. Here is a screen of my voltage charted against the recorded AFR.

2033920340

[joecar]

Post your sae_generic.txt, that's probably why I see wrong values for CALC.AFR_LM21.

[joecar]

When I look at your log, EXT.AD1 and CALCAFR_LM21.V show the exact same values as each other...

I also see that this voltage stays at 4.35V most of the time and briefly dips down several times (to 1.0V, 2.83V, 2.88V, 2.52V, 2.53V).

[joecar]

My sae_generic.txt defines CALC.AFR_LM21 as V*10+0 so this is giving me wrong values when I view your log.

[Mitco39]

When I look at your log, EXT.AD1 and CALCAFR_LM21.V show the exact same values as each other...

I also see that this voltage stays at 4.35V most of the time and briefly dips down several times (to 1.0V, 2.83V, 2.88V, 2.52V, 2.53V).

The dips correspond to the afr going richer (atleast richer than 20afr) for a moment. Take a look at the log I posted previous to this one and you should see it all looks normal. With the 15% change this truck was running at 18+afr and thats why the log shows as such.

My calculation is

"{EXT.AD1}*3.00+7.05" this makes it match up perfectly with the serial connection and using the innovative program to monitor the AFR. I have very little doubt that the issue is the wideband or its importation mainly because I keep the logworks program open and on top of the log to continually confirm the afr posted in the scan tool exactly matches the afr illustrated in the innovative program.

Im just at a complete loss as to how it will go from .85BEn you change the VE 15% and now its a 1.35BEN on the next log. This is all other things being held constant.

Thanks so much

Mitch

[EspenT]

The dips correspond to the afr going richer (atleast richer than 20afr) for a moment. Take a look at the log I posted previous to this one and you should see it all looks normal. With the 15% change this truck was running at 18+afr and thats why the log shows as such.

My calculation is

"{EXT.AD1}*3.00+7.05" this makes it match up perfectly with the serial connection and using the innovative program to monitor the AFR. I have very little doubt that the issue is the wideband or its importation mainly because I keep the logworks program open and on top of the log to continually confirm the afr posted in the scan tool exactly matches the afr illustrated in the innovative program.

Im just at a complete loss as to how it will go from .85BEn you change the VE 15% and now its a 1.35BEN on the next log. This is all other things being held constant.

Thanks so much

Mitch

I'm pretty new to all this. But to me it seems like something in your setup does not react linear to your change. I have no clue what the PID is named, but it should be possible to se your injectors duty cycle. Is that in the logs? Does the duty cycle change as much as the AFR? If so, there is something making the PCM command the lean AFR. If the change in duty cycle only reflects the -15% in ve, then there's something with fuel system or injector setup.
Sorry if I'm all wrong or if you have already tried this. I primarily follow this thread to learn :)



Sent from my iPhone using Tapatalk

[joecar]

The dips correspond to the afr going richer (atleast richer than 20afr) for a moment. Take a look at the log I posted previous to this one and you should see it all looks normal. With the 15% change this truck was running at 18+afr and thats why the log shows as such. Hi Mitch,

In both your logs I see CALC.AFR_LM21 going above 20 (using your calc pid as you show below)...




My calculation is

"{EXT.AD1}*3.00+7.05" this makes it match up perfectly with the serial connection and using the innovative program to monitor the AFR. I have very little doubt that the issue is the wideband or its importation mainly because I keep the logworks program open and on top of the log to continually confirm the afr posted in the scan tool exactly matches the afr illustrated in the innovative program.


I tried this calculation, see attached pic, is AFR really going above 20...? Engine can't run at at AFR, it would misfire at load.

20346




Im just at a complete loss as to how it will go from .85BEn you change the VE 15% and now its a 1.35BEN on the next log. This is all other things being held constant.

Thanks so much

MitchYes, that does not make any sense, there has to be something else going on.

Are you sure you get an immediate MAF DTC...?

[Mitco39]

Hi Mitch,

In both your logs I see CALC.AFR_LM21 going above 20 (using your calc pid as you show below)...




I tried this calculation, see attached pic, is AFR really going above 20...? Engine can't run at at AFR, it would misfire at load.

20346


Yes, that does not make any sense, there has to be something else going on.

Are you sure you get an immediate MAF DTC...?

- I suspect she is going very lean, it has backfired and you can visually see the exhaust getting hot, so hot that I didnt want to load up the truck harder than the logs show. Of course I only have the single wideband so I guess it is possible that it is reading incorrectly but it is bang on when the truck is running CL.

- Yes sir I am. I even went one step further and set the MAF table to all 0's. If it was still running the maf table at all it should not run because I commanded the full HZ range to 0 flow.

[Mitco39]

I can attach another log showing both CL and OL and how the WB02 is responding in both cases, it seems to be doing what I would otherwise think.

[Mitco39]

I'm pretty new to all this. But to me it seems like something in your setup does not react linear to your change. I have no clue what the PID is named, but it should be possible to se your injectors duty cycle. Is that in the logs? Does the duty cycle change as much as the AFR? If so, there is something making the PCM command the lean AFR. If the change in duty cycle only reflects the -15% in ve, then there's something with fuel system or injector setup.
Sorry if I'm all wrong or if you have already tried this. I primarily follow this thread to learn :)



Sent from my iPhone using Tapatalk

I could try logging the pulsewidth and see if I can see anything that might be causing any issues, it is atleast somewhere else to look. I did confirm with the customer that the injectors and the engine (minus the cam) is completely stock.

[Mitco39]

I just stumbled onto something else that may help you pinpoint what is going on.

I have the truck in OL on the dyno holding the load exactly the same and the rpm is steady. With my current VE I have about a BEN of .85 when the truck is commanding 12AFR. Now if I go into the DVT and manually change the AFR to lets say 14.7 to force it the BEN goes up to 1.25. This is while the whole time sitting in the exact same VE cell in a steady state. From what I know wouldnt you expect the BEN to stay the same reguardless of the commanded AFR?

Could this be alluding to either injector tuning or maybe the injectors themselves?

This is what is causing my issue, the leaner truck is commanding the higher the BEN, the richer the lower the BEN. What could be causing this?

[EspenT]

I just stumbled onto something else that may help you pinpoint what is going on.

I have the truck in OL on the dyno holding the load exactly the same and the rpm is steady. With my current VE I have about a BEN of .85 when the truck is commanding 12AFR. Now if I go into the DVT and manually change the AFR to lets say 14.7 to force it the BEN goes up to 1.25. This is while the whole time sitting in the exact same VE cell in a steady state. From what I know wouldnt you expect the BEN to stay the same reguardless of the commanded AFR?

Could this be alluding to either injector tuning or maybe the injectors themselves?

This is what is causing my issue, the leaner truck is commanding the higher the BEN, the richer the lower the BEN. What could be causing this?

Hmm... If I'm not mistaken the BEN should be commanded AFR/Measured AFR.
So if you have a BEN of .85 on commanded 12, you have a measured AFR of around 10?

If you have a commanded AFR of 14.7 and BEN 1.25, then measured AFR = 18,37 ???

Seems like every change you do ends up being much larger than they should be.

If the injectors are stock, then I have no idea. If the injectors are to big, it could explain this behavior. As you would be running them in a non linear puls range. Again, it would be nice to know the duty cycle
http://fuelinjectorclinic.com/data-match-technology

Sent from my iPhone using Tapatalk

[Mitco39]

Hmm... If I'm not mistaken the BEN should be commanded AFR/Measured AFR.
So if you have a BEN of .85 on commanded 12, you have a measured AFR of around 10?

If you have a commanded AFR of 14.7 and BEN 1.25, then measured AFR = 18,37 ???

Seems like every change you do ends up being much larger than they should be.

If the injectors are stock, then I have no idea. If the injectors are to big, it could explain this behavior. As you would be running them in a non linear puls range. Again, it would be nice to know the duty cycle
http://fuelinjectorclinic.com/data-match-technology

Sent from my iPhone using Tapatalk

Ill grab you guys a log here once I am done with a customer. Ill manually change lamda while holding the truck exactly still on load and rpm, ill add injector pulsewidth data to the log for you to see what you think is up.

Thanks so much

Mitch

[Mitco39]

Here is a short log of me manually increasing the AFR, check out what the wide-band is doing with ever step up of the desired AFR.

Im at a loss. Stock injectors, stock injector mapping.

[joecar]

I just stumbled onto something else that may help you pinpoint what is going on.

I have the truck in OL on the dyno holding the load exactly the same and the rpm is steady. With my current VE I have about a BEN of .85 when the truck is commanding 12AFR. Now if I go into the DVT and manually change the AFR to lets say 14.7 to force it the BEN goes up to 1.25. This is while the whole time sitting in the exact same VE cell in a steady state. From what I know wouldnt you expect the BEN to stay the same reguardless of the commanded AFR?

Could this be alluding to either injector tuning or maybe the injectors themselves?

This is what is causing my issue, the leaner truck is commanding the higher the BEN, the richer the lower the BEN. What could be causing this?Mitch,

Can you get a log of this (showing GM.AFR, CALC.AFR_LM21, CALC.BEN_LM21)...

( start logging, and while logging is proceeding, goto DVT and force the AFR... then when done, deactivate DVT and stop the log )

Yes, because when the PCM commands some AFR, GM.AFR shows it (and so the BEN should stay the same, as long as the WB followed the AFR).

[statesman]

The wideband follows commanded AFR perfectly if you use the formula V+11. That's not a very wide wideband. :confused:

[Mitco39]

This is the wideband I'm using. Any ideas?

http://www.innovatemotorsports.com/products/lc2.php

This is the setup guide I used.
http://www.innovatemotorsports.com/resources/efilive_tutorial.php

Joe,

I'm pretty sure that's what I did in the log I posted just above. It's a short log showing exactly what i think you are asking about.

Thanks so much guys.

[statesman]

Try switching over to the brown wire and use "Analog out 2". Then do another AFR sweep like you did in your last log.

[Blacky]

Just jumping in here, sorry I haven't read the entire thread so what I say may have already been covered.

From the log file it looks like you're using an Innovate side band controller. Innovate uses a programmable stoichiometric value that is used by the wide band to determine AFR and hence (I presume) to determine the output AFR voltage. Are you sure the wide band's stoich setting is correct for the fuel you are using?

You need to use Innovate's Logworks software to set the stoich value.

Regards
Paul

[joecar]

This is the wideband I'm using. Any ideas?

http://www.innovatemotorsports.com/products/lc2.php

...This is the LC-2... in its default configuration/settings, it uses the same pids as the LC-1 (CALC.AFR_LC11, CALC.BEN_LC11).

[Mitco39]

Just jumping in here, sorry I haven't read the entire thread so what I say may have already been covered.

From the log file it looks like you're using an Innovate side band controller. Innovate uses a programmable stoichiometric value that is used by the wide band to determine AFR and hence (I presume) to determine the output AFR voltage. Are you sure the wide band's stoich setting is correct for the fuel you are using?

You need to use Innovate's Logworks software to set the stoich value.

Regards
Paul

Paul, I appreciate you taking the time.

I did check the settings, I have the firmware programmer that allows me to setup the controller outputs. I will take a closer look at that. It does kinda seem like statesman might be onto something, it appears like it could be functioning as a narrow band so ill go through all the settings to see. That said when using the Innovative program it is showing me the same values, which should rule out any sort of analog output mapping because I can check it against itself.

Judging by the comments here you guys think its something either with the controller or the inputs itself? Nothing in the tune could cause this type of issue, or nothing easy to overlook? Ill look over all my settings again.


Joe,

Are you saying that this link to setup the controller is incorrect? Im at a loss.
http://www.innovatemotorsports.com/resources/efilive_tutorial.php


Thanks guys.

Mitch

[joecar]

...
Joe,

Are you saying that this link to setup the controller is incorrect? Im at a loss.
http://www.innovatemotorsports.com/resources/efilive_tutorial.php


Thanks guys.

MitchIf you have an LC-2, I would set it back to default, and use the LC-1 calc pids...

if you have a serial comms cable you could use that to connect to V2 (needs null modem gender bender), it gives Lambda and AFR with no conversion to analog needed (i.e. you don't have to worry about which of the the 2 analog outputs you're connected to, and you don't have to worry about voltage offsets).

[joecar]

In its default configuration, the LC-2 (and the LC-1) assign the analog channel like this:



The default analog outputs are as follows:
Analog output one (yellow) is 0V = 7.35 AFR and 5V = 22.39 AFR.
Analog output two (brown) is 1.1V = 14 AFR and .1V = 15 AFR.

[statesman]

Judging by the comments here you guys think its something either with the controller or the inputs itself? Nothing in the tune could cause this type of issue, or nothing easy to overlook?

Yes, I would call this a wideband issue. In your last log... as the commanded AFR changes, the changes in reported injector pulsewidth would indicate that the tune is making the correct changes to your fueling.

I'm not familiar with Innovate products but from joecar's last post, it appears that the brown wire is for the simulated narrowband... so ignore my previous suggestion to switch to that wire.

[EspenT]

Paul, I appreciate you taking the time.

I did check the settings, I have the firmware programmer that allows me to setup the controller outputs. I will take a closer look at that. It does kinda seem like statesman might be onto something, it appears like it could be functioning as a narrow band so ill go through all the settings to see. That said when using the Innovative program it is showing me the same values, which should rule out any sort of analog output mapping because I can check it against itself.

Judging by the comments here you guys think its something either with the controller or the inputs itself? Nothing in the tune could cause this type of issue, or nothing easy to overlook? Ill look over all my settings again.


Joe,

Are you saying that this link to setup the controller is incorrect? Im at a loss.
http://www.innovatemotorsports.com/resources/efilive_tutorial.php


Thanks guys.

Mitch

I find this setup strange. You tell EFI live that 0v is AFR 10, and 5V is AFR 20. Then you configure the Wideband to behave differently. Where AFR 10 is actually 0.88v and AFR 20 is 4.2v.

I might be all wrong, because I do not know EFI live very good. But as I understand the sae_generic.txt the config for your wideband should look like this:

# Innovate WBO2
# ================================================
*CLC-00-1007
V 0.88 4.21 0.1 "{EXT.AD1}"
AFR 10.0 20.0 0.2 "{EXT.AD1}*3.00+9.70" #10.00 - 0.30 to compensate for cable resistance


Tried to run trough your log with this configuration, and you are leaner than commanded fuel the whole time.
Do you have simple wideband gauge so that you cold verify the AFR?

Sent from my iPhone using Tapatalk

[joecar]

The 10 and 20 you see in the CLC are the AFR chart/gauge display min/max limits (i.e. the scantool chart's vertical range for that pid, and the scantool gauge's sweep range)... the mapping of V to AFR is done purely the equation (regardless of the AFR min/max display limits)

[joecar]

Same with the 0.88 and 4.21, these are the V min/max chart/gauge display limits for V.

[EspenT]

The 10 and 20 you see in the CLC are the AFR chart/gauge display min/max limits (i.e. the scantool chart's vertical range for that pid, and the scantool gauge's sweep range)... the mapping of V to AFR is done purely the equation (regardless of the AFR min/max display limits)

Thanks for explaining that :) Then that equation he uses is in fact correct.
10AFR / 3.333V = 3.0AFR/1v

Since wideband starts at .88 not 0.
0.883 x 3 = 2.649
10 - 2.65 = 7.35

And we do still not understand anything about what's causing Mitch's problems. At least I learnt something.



Sent from my iPhone using Tapatalk

[Mitco39]

I'll be back on the Dyno Tuesday and I'll try a few of the suggestions listed here. The thing that makes me think it's not the wideband input settings is I have the innovative being input via serial to the laptop directly and logging through their logging program and I can confirm that the afr it displays is very close to that of the efilive scan tool. To me this means or should mean that my inputs are correct, if they were not these two different loggers would be different from one another.

I also confirmed that during closed loop my wb02 is displaying right around 14.7 and it's responding so quickly that I can see the stfts working as it bounces around 14.7.

I have a 5.3 waiting to go on the Dyno, I might just strap that down and go through the same process. This will rule out the truck or the wb02 depending on what my results are.

This truck does have about 200,000miles on it so it's possible there is something else at play mechanically. Switching trucks will rule this possibility out.

[Mitco39]

SO I think I finally got it figured.

Put the 5.3 on the dyno and same thing. It dawned on me that maybe the O2 needed to be calibrated again. Sure enough that was all it took. Put it in fresh air and calibrated it and now the desired follows the actual afr (within a small margin of error anyways). I think that was my issue all along.

I will know for sure once I get into the meat of tuning this thing! Didnt have much time left after I stumbled onto this.

Thanks guys so much! Remember to calibrate your WBO2s!


Mitch

[joecar]

SO I think I finally got it figured.

Put the 5.3 on the dyno and same thing. It dawned on me that maybe the O2 needed to be calibrated again. Sure enough that was all it took. Put it in fresh air and calibrated it and now the desired follows the actual afr (within a small margin of error anyways). I think that was my issue all along.

I will know for sure once I get into the meat of tuning this thing! Didnt have much time left after I stumbled onto this.

Thanks guys so much! Remember to calibrate your WBO2s!


MitchMitch,

Good job figuring it out :cheers:

( now at the top of my quickie notes I'll write: "has WBO2 been free-air calibrated recently" )

[Mitco39]

haha yeah to be honest I should have thought of it sooner, but I guess it was in my head that it was a new one and I thought I did calibrate it but maybe it needed a few heat cycles and then to be calibrated again, that or I didnt do it correctly the first time.

Thanks so much for not giving up on it with me, I still learned alot about the process from this mess. haha.

[joecar]

No worries... we all learn stuff from all the things that go wrong :) some of my best lessons.

[cindy@efilive]

The AutoVE Tuning Tutorial that TreCool updated has been added to the next software release and our website.

http://download.efilive.com/Tutorials/PDF/AutoVE%20Tuning%20Tutorial.pdf

Cheers
Cindy

[joecar]

All,

The updated AutoVE document has some very good additions and explanations added to it.

Thanks to Tre-Cool/David who took time to go thru it (and it does take some serious concentration to make it come out right).

[Tre-Cool]

No problems. I'm thinking it might be worth updating some of the e38/e67 style controllers and maybe making something up for the diesel e83 ones, but the e83's ecu are real simple nothing like tuning a petrol car they could probably be covered via forum post.

add fuel + boost = more power

[Chevy366]

Giggity, to you Tre-Cool.