I've been fighting a weird issue. If I barely tough the throttle, the wideband O2 sensor shows my AFR shows going super rich to 10, the gauge limit. I can pretty much keep it there, with just the lightest of pressure, holding rpm, not accelerating.

( edit: I should have pointed out, with a heavier foot, I see 12-13 during actual acceleration events.)

After data logging today, I noticed a few events that may be correlated.

As the AFR goes to 10, The IAC follows (or does it lead?) straight in the toilet to near zero The BLMs go to 150-160 (which seems contradictory in nature) after being near 128 The other change noted was the BLM cell number goes from 9 (where it idles) to 10.

I admit to having no clue as to what a cell number indicates, how many there might be, what are their borders, how they can be changed?

Is idling in 9 is normal or strange, or am I seeing ghosts where there are none when it changes from 9 to 10.

Things go in opposite directions and stay there for no apparent reason that I understand. I'm sure this must be killing mileage. I don't follow how BLMs can be so lean when the AFR says pig rich at the same time. They can't both be right. Is the IAC following the problem or creating it, and what is the significance of changing cell numbers at the same point?

As always, in am indebted to those kind enough to respond.
Dave

 

any chance that you have a vac leak at idle, that somehow closes up when you hit the gas????

how long has this been going on?

 

Seems to have been there since at least spring. I haven''t been able to notice till I put a wideband sniffer into the car in June.


I saw the AFR go to 10 the first day it was in. I have been looking for correlations, but only recently looked at the IAC, BLM adn Cell # grouping. Really not sure which may be cause vs effect among them.

Thanks for the interest. I also had a friend suggest fuel prssure may be an issue. I know what my static pressure is (48lbs) but don't have a fuel pressure gauge hooked up to see running pressure.

 

If the TPS% doesn't go high enough for the ECM to leave idle mode, the IAC will head to zero to bring the idle speed back down.

Need to look at the NB O2 signal as that is what the ECM is using.

RBob.

 

RBob, please expand on this point a bit further. When I am viewing the NB O2 signal, what am I looking for? what do I need to modify?


For background, when I installed the wideband, I kept the narrowband in place and hooked up. There are simply 90* apart on the collector down pipe.
for idle, I have previously dropped both the O2 mean and thresholds voltages by 20mv but this is a while running issue, Dropped per the following thread per another thread at:

https://www.thirdgen.org/forums/diy-...-contants.html.

Thanks for your continued help aver the past few years.

Dave Buchholz
Rochester,NY

 

If the NB O2 stays below 450 mV the INT will increase. The BLM follows the INT.

Note that the 450 mV is a typical value, the O2 window values define the actual switch point for the INT fuel trim.

It doesn't take much of an exhaust leak to affect the O2 sensor reading.

RBob.

 

her's typical values that I'm seeing when it happens. ($8D)

O2 Mv at 954.7
Afr at 10.35
IAC at 9
BLM at 164
Some other tables of interest related to O2 closed loop

"Loop closed, Param Diff to make R?L window Fast O2" set at 0.502

"Loop closed. idle O2 rich lean, slow zero error lower " at 0.478
"Loop closed. idle O2 rich lean, threshold at 0.500
"Loop closed. idle O2 rich lean. Zero slow error upper at0.522

similar TABLES for MAP also found:
center of graphs at 50Kpa listed (Graph went from 10-90kpa)

lower 0.841
threshold 0.625
upper 0.669

 

What is your TPS reading at idle?
Where is your Integrator, what is it reading while BLM's are that high?
Did you check for leaky injectors?

 

Does the NB O2 signal just stay up around .9 volts with the BLM maxing out? Or does it cycle around .450 V?

 

To answer Both Your questions, I had an opportunity to datalog on the way home tonight.

TPS % at idle is an actual 0%, verified at several times in data log.
Noted during several "events" TPS% within a range from 7.04 to 13.3 %

BLM Integrator reaction is peculiar.
From morning data log at first Integrator followd BLM's up to 160, but as more events occured, Integrator was always less than previous event, until BLM int barely moved from 128-134 even though AFR was still down to ten.

From evening datalog, INT counts did not repeat this activity pattern, seeming more random with a range of 136-147 as events occurred.

Injectors have less than five thousand miles on them, not tested for leakage through the pintle into chamber, no sign of wetness around rail.

NB O2 volts during events, Taken from O2 sensor filtered Mv range form 906 to 954Mv. There is no cycling or drop off from that high Mv range until the throttle is either lifted or engaged further.

Another noticed point, The AFR tracing and IAC tracing are almost identical in nature, looking much like copies of each other.

Dave

 

The cell # indicates what BLM correction value will be used.
Each of the cells or "blocks" will store the fuel correction "value" to be used during that area of operation.
128 is no correction, higher = lean and is adding fuel, Lower = rich and is subtracting fuel.
The boundaries are setup like this:
There are two tables that allow you to change where the block values are, 1 table for the MAP values and one for the RPM values.
You should adjust your lower boundary (RPM) if your idle is higher than the setting so you are in Cell 4.
By stating you are in cell 9 indicates your idle is above the stock 900 RPM setpoint.
This allows you to have an "idle" fuel correction and then move to a higher cell for off-idle running correction.
Same for cruising MAP/RPM cells having a particular correction than the higher load areas.
The cells below the "low MAP" threshold are for coasting/overrun times.
Adjust the values if needed to suit your driving style or setup needs.
HTH

If BLM is that high and you are rich, search for an exhaust leak or a misfire as the first things.

 

Great Info JP. If this isn't a sticky somewhere. it should be!

I'll check for leaks or misfiring cylinders next.

I usually idle at or near 900rpm and 55Kpa. To reset the boundaries or cells, it looks like I should try to get my idle into cell four by raising the mid map to 55 and the lo rpm to 1000. This rpm would would still be below the threshold of where my "in the tank" events are happening, knocking my idle into a different box than where my events reside. correct?

Man, So much to learn. I had just responded to a junior member than tuning is not a one weekend event. It seems it's a second calling in life. Once again THANKS for all the immeasurable help.

Dave Buchholz
Rochester NY

 

The computer will keep cycling the Integrator, it will try to average stoich (128) by keeping the time spent rich equal to the time spent lean. If it is averaging away from stoich (128) it will add or subtract to the current BLM value, whatever that may currently be reading, moving the short term adjustment (INT) average back towards stoich (128). The ECM is still able to maintain stoich with your BLM's being pegged because the INT is not at it's limit, the ECM is still able to correct. BLM's will follow the INT, but if it doesn't (or can't) then there is a problem somewhere keeping your BLM's pegged, either with fuel distribution (too much, too little), a vacuum leak, a problem with spark (arcing wires, bad plugs), or a skewed sensor giving false information (bad O2)....

 

Thanks for the additional insight. Lots of things to check out with your added comments...

I'm most appreciative of your taking the time to respond.

Exhaust leak, bad plugs. (wires just replaced with expensive ceramic ended units due to past frying of normal silicone boots)
Will still look for arcing after dark to ease search
Double check fuel pressure that it hasn't changed.

NB O2 sensor is relatively new...5K miles on it.

 

.... anything that will effect either the actual mixture in the combustion chambers, or the burning of that mixture, will greatly impact your BLM reading. All the reading is, is a reading of either too much air, or too much fuel, being in the exhaust. The INT can average stoich but only to an extent. Since the INT is able to hover around the 136-147 mark, then your still getting O2 correction. Your NB is reading excessively lean so your ECM is throwing fuel at it, while your WB is reading rich because of that added fuel being thrown at it.

 

So it would seem to beg the question as to why the NB is reading lean.

 

As far as the O2 goes, the reading you want to be looking at is the unfiltered O2 volts. The slow filtered volts will just be an average of all the combined readings. If the unfiltered ones just hang up above .450 volts, then there is something wrong. The O2 volts should essentially form a sawtooth pattern when its plotted as the ECM cycles back and fourth around stoich.

If the O2 reading cycles, but its just pig rich, then its an exhaust leak, misfire, bad O2, or cam overlap. Cam overlap actually causes a sharp drop-off in aparent AFR to the O2s as the RPMs drop. Its fairly steep. The engine will be fine at cruising RPMs, but at idle it will blow black smoke due to all the air getting in and dragging the readings down.

 

I'm using Tunerpro RT. In the monitor section along with the filtered o2 mv, I found O2 crosscounts. I'm hoping this what you were referring to. It does show a slow sawtooth pattern, cycling from 0-255. It still cycles, although a little more quickly, during "event" times when the AFR dips to 10.The only time I see the cross counts locking up, is when Highway Mode becomes engaged and my BLM's lock at 128. Out of Highway Mode, the crosscounts pick up again.

This morning, I have changed the exhaust gasket (o2 sensor side), no burn seen, but plenty of smoked area at the 3/5 holes. Might be letting some air in. Checked and re-gapped plugs at 0.050.

I''ll return in day or so, after checking all these conditions out, some physically, some via datalogging, to discover and eliminate mentioned possible causes.

As a side note, this thread has been quite an education so far! Thanks to those who've taken interestand time to respond..

Dave Buchholz

 

I started wondering about O2 constants. After reading the following thread:
https://www.thirdgen.org/forums/diy-...ml#post5011995
It made me wonder if AIR if a nagginf issue.

The A.I.R. pump is off, and I remembering changing afew things way back when to compensate for it being gone. I'mwondering If I badly conpensated and still have an issue with the ECM thinking it's there.

IN $8D, does anyone recall ALL the parameters that need resetting to insure the 100mv factor from that circuit affecting the O2 sensor is correctly and permanently turned off? I think it would be a good idea for me to review that area of my programming.


I think I found the parameters that might need review, but would like confirmation.

"AIR, mph divert" make sure it way up there,like 200+ mph, so it never gets there.
"AIR min coolant temp enable" again set high that ot never gets reached, like 250F

"AIR rich/lean reduction volts" make it "0" so if ever engaged it doesn't do anything.

Did I miss something to check on?



Dave Buchholz

 

The BLMs shouldn't lock at 128 when in highway mode. They should lock at whatever value they are at at that time. This keeps the resultant AFR closer to the commanded AFR.

That is a lot of plug gap. With a CDI box I'll use a max of 0.045". With standard HEI I'll use 0.030 - 0.035".

AIR: Set the minimum CTS high and zero out the O2 voltage offset value. Then reduce all of the O2 Window values by the former value of the offset.

RBob.

 

RBob, I see anumber of O2 relaqted facors doing a search. I want to make sure that I am changing the correct "window" values Sounds like I should be seeing an upper & lower or min & max entries. Under "AIR" in my search, or am I missing the point to reset other min/max values.

 

Whenever AIR is in divert mode, the value at:

L849D ; 0.10 V DIFF R/L WHEN AIR DIVERTED

Is subtracted from these values:

L84A2: ; UPPER THRESHOLD FOR SLOW O2 R/L vs MAP
L84AB: ; LOWER THRESHOLD FOR SLOW O2 R/L vs MAP
L84B4: ; FAST O2 R/L THRESHOLD vs MAP

L8496: FCB 136 ; 600 mvdc, FAST O2 R/L THRES AT IDLE
L8497: FCB 136 ; 600 mvdc, UPPER ZERO ERROR O2 AT IDLE
L8498: FCB 128 ; 570 mvdc, LOWER ZERO ERROR O2 AT IDLE

Which is why the above values appear to be so high. Take 100 mV off each and they are much closer to stoich.

Also, don't forget about the coolant compensation of the O2 window values:

L84F6: ; R/L OFFSET vs COOLANT

RBob.

 

During the course of re-entering the parameters listed above by RBob, i discovered a previous error. I had the upper and lower O2 idle limits reversed, upper was the low number, lower was highest number. Been running quite song time that way.


Don't have enough datalog time on to say the other issues are resolved yet. One battle at a time

Dave

 

what did you end up using for your blm map/rpm cell values? just did a cam swap over the weekend and it seems to be almost identical to yours pulling 55-60 kPa at idle. i didnt adjust those tables at first with this cam, and once it hit a certain temp(blm learn enable im assuming) the idle bounced all over the place and stalled on decel. once those tables were adjusted a little closer to the kPa range and idle speed the surging quit as soon as i uploaded the bin.

 

I moved my cell values so that my typical idle would end up in box 5. After datalogging to see where I really was idling, I ended up raising mid Kpa to about 58 and low rpm to 975. In my case It moved the idle cell number from 9 down to five. Not sure if these will end up being final values yet.

Still haven't finished solving the sky/ toilet issue yet. Lately, being up to my butt in alligators is making it hard to keep my mind on draining the swamp.

 

OK this is really getting odd.

I've been looking for correlations in my datalogs to see what goes up or down when my AFR goes down to 10 under light foot pressure and virtually no throttle movement. I was useing the Monitor section in Tunerpro when I found:

The trace for the IAC and the AFR look like exact copies of each other, identical. For every dip , notch, peak valley in the IAC steps, there is an absolute correlating matching dip and peak trace in the AFR data. You'd swear it was the same tracing mislabeled!

So does it seem that the IAC is the controlling factor of my AFR issues.

When the IAC steps hit bottom and zeros out, the AFR is at at it's lower limit of 10. At IAC of 160, AFR is 16.1. At 200+ the AFR is maxed out at 20.

The O2 mv data seem to jump either high or low at those same areas in my datalogs.
So things LOOK related to movement of my IAC.

The question now is WHY, and how to fix it.

Any thoughts out there? Ma, I'm at a loss here!

 

maybe try the DFCO AE stall saver PW? i had to decrease mine to keep from flooding out, maybe you have to increase yours?

 

I'm trying to cure an overly rich problem to begin with, as was indicated in the opening thread. I don't follow, can't see how adding fuel to an overly rich condition cures it, nor does it explain the IAC and AFR trace as being copies of each other.

DFCO is for when you lift off the throttle as a fuel savings tool. This problem is when adding pressure to the gas pedal, not releasing, so I don't get how changing it would help.

 

oh snap lol. i read the last post wrong, sorry

 

I'm still looking for some additional insight as to why my IAC and AFR traces are identical in my dtalogs.

ANY takers on this? Something pin wrong? Glitch in program code?

 

that may be explainable. When IAC opens it adds fuel. check values for IAC-AE or AE-IAC. a while back I attempted to use it to help a bog condition(lean at tip in).

Did you repin the ECU connector? I see you did a TBI to TPI swap. is your .xdf corrupted. can you discard old def and use a new one.

I was taught to believe the WB is 99.5% accurrate

I presume you have a VAFPR. Post #1 you state "If I barely tough the throttle, the wideband O2 sensor shows my AFR shows going super rich to 10, the gauge limit."

If the VAC decreases with opening throttle could the VAFPR be adding excessive fuel? How is that handled in your def file(BPC?)?

Possible the engine is being flooded with fuel and the 02 sensor reads unburned 02 and reports lean? Not sure how your def file looks but are your values for VAC(MAP)/BPC reversed? Just asking. I know unlikely... VAFPR could cause havoc...

 

They shouldnt be. The fueling will effect the required IAC airflow, but they still should show some differences.

Can you post up the ADX file along with a datalog showing the problem?

 

First off, I should clarify that the identical traces seem to be the wideband AFR and the IAC MINIMUM LEARNED STEPS

I've also noted that the occurrence includes a jump in the injector pulse width from around 2.4usec to around 4.6usec. Essentially I'm seeing it double during events, so I'm definitely getting a big fuel dump at the injectors.

I'd be thrilled to have you guys look at a log for me! I 'll even BUY the beers! (edited for stupidity)

Since I've never attached a file in this forum, and am unsure of the process, cluing me in as to HOW would be a good thing.

 

You'll beer the beers?
sounds like something I'd be up for
Best thing is to just make the log a zip file and then attach it using the window when you post. It only takes certain file types and that is one of them.
I've lived through this same problem and am thinking what I did to solve it.
I do know it was a combination of throttle follower on the IAC and AE settings.
I had a very rich spike of idle no matter how much AE I took out.
Adjusting the IAC helped allot IIRC.
there's a post I did that I must find to explain it.
I'll dig it up.

 

I found the thread I was looking for that I did allot of bench testing on but it was for an RPM flare at idle and not coming down, not really related to your issue.
I did come across a note about "MIN TPS% for AE"
Increase this value to 5% and see if that help remove some fuel as you come off idle. Your IAC may be opening but the AE is just adding fuel possibly at the same time.
This has helped some with off idle surging as well.
Post up a log also and we'll get this figured out.
Jp

 

Maybe the ECM is switching from double fire to single fire?

RBob.

 

I'm on a differernt computer so can't post the Zip file at the moment but, here's something else I looked at.

Again looking for correlations, i figured it is all started with my right foot. So that aspect is tracked an translated as TPS movement. As mentioned I spotted a jump in Injector pulse width, which I now confirm as related directly to TPS movement of typically 8-10% from base idle level.

I then looked at my flags section while data was running, taking note of the specific "time" in the log that it occurred. At each event point, I would see a quick flash in red indicating "AE" was active. I began to look for AE controls for when, how much and how long. I also compared to a stock AUJP bin as well.

AE, min TPS for for AE ... raised from 2.7 to 4.0
AE, %TPS for 53B pump shot ..... raised from 3.13 to 4.0
AE, factor per BPW vs ref pulse, at number 1lowered from 3.320 to 2.75
AE, async factor vs Delta TPs... noted high numbers there at low TPS % relative to higher %tps. Since this is there my "events " are happening I lowered the first three entries from 1.0 down to 0.83

Willmake attempt to attach zip file today.


I have one data run in so far, with encouraging results noted. My AFR doesn't drop as quickly or as far seen it dip to 10.9 afew times, But that seems better than 10.0. SO I think I'm on the right track. Acceleration seems affected in a positive way.. seems more responsive.. maybe not getting flooded so much as with so rich of an AE event beforehand.

 

OK so here' my attempt at loading the file. It was created in Tunerpro RT version 4

Any and all responses to views, suggestions, ideas, thoughts, and cold beers are greatly appreciated, not only for the "AFR in the toilet "questions but general programming as well. If you spot something weird, I'd be happy to hear your input on it

Many thanks in advance,
Dave Buchholz, Rochester NY

 

after some glaring failures in programming that made things worse, I've back to the point in time of the attached Bin file.
Still working on the original issue of keeping the AFR value from going so rich. I'm still after the factors controlling how soon, how long, how much,to tame things.


A prior thread was for restart issues to help the lean open loop areas. I've made all the corrections for O2 tables as the AIR pump diversion is no longer there.

On the question of O2 mv values, I still haven't caught on to a simple point. If I want a richer open loop mixture Do I lower or raise the O2 values?

 

It won't matter if they are raised or lowered, they aren't used in open loop.

If the VE is correct then change the open loop AFR, versus MAP or versus coolant. Which depends upon when/how it needs to be changed.

RBob.

 

Thanks to JP_, I've finally got this under control. He sent a bin for his motor which was similar and I was able to do a side by side comparison of AE related parameters.
I was going the wrong way on most of them, and found an additional one with an apparent typo where instead of a value of 0.053 I had 0.503. So I was off by ten times.

If anyone needs details of what specifically cured the problem, let me know and I'll post the data.