Dictys

The problem:

Only when hot and running the 02 cat maps with the 02 connected, the car stalls once started until a few revs are given. Normally this is with a little load such as A/C on. Once a few revs are given the car runs fine until next time the engine is switched off. However, in non 02 mode the car runs perfectly all the time. Nor does it matter what chips are in the car or wether it is a NBsim from my WBO or a signal from a brand new porsche 02 sensor the same problems occur. The Temp sensor has been replaced a few times the same issue occurs, new vacuum dampers and fuel regulators, ignition wiring good. The LH has been rebuilt and checked by John. The outside temp is between 30-45c and the coolant is at 86c.

So below are two maps one when the car is in non cat mode and the other is in cat mode (logging the funky idle and stall) with the 02 sensor controlling the a/f ratio. As you can see in cat mode the cat seems to go lean a lot near idle and the 02 adjustment is big to try and control it.

Anyone with any ideas?

jcorenman

Are these both stock maps, or have you done some tuning?

The first (I assume no-cat) map looks pretty sweet, but the second map (cat?) has map values in the idle area (700-990 rpm, 116-100 load) that are much more negative-- which would be lean without the closed-loop O2-adjust, as it shows and as you said. (Sharkplotter indicates what it would have been, without the adjustment).

So what happens if you copy the no-cat map into the cat map? The no-cat map looks like a better map, closed-loop or otherwise. (In ST, choose the no-cat map and then click the upper-left (gray) corner of the border to select all cells, Ctrl-C keys to copy, then choose the no-cat map and select the upper-left (white) map call and Ctrl-V keys to paste-- all with the ignition on, engine running or not).

If the map is lean in the idle area then it's going to be bitchy about idling. The O2-adjust correction will fix it once it settles down, but coming down to idle there is going to be a slight delay before O2-adjust kicks in, and you want the map to be pretty close. (If ST's O2-Adaptation is enabled then that will also compensate, but the goal is to have the map correct so that O2-adaptation doesn't have anything to do).

For those following along: ST's "O2-adjust" is the quick adjustment that the LH makes in response to the NBO2 sensor in closed-loop mode (engine warmed up, "cat" map selected). If the map is correct then O2-adjust will hunt either side of zero, roughly between +/- 4% a couple of times a second as the LH tries to keep the AFR at 14.7. And the AFR will hunt between roughly between 14.2 and 15.2:1. (The amount of hunting, and the frequency, depends on the age/condition of the NBO2 sensor). The O2-adjust correction can go from -20% to +20% fuel. The long-term average (over minutes) of O2-adjust is saved as "O2-adaptation", also +/- 20%, and is applied all the time to compensate for long-term effects, i.e. component tolerances on production motors and MAF aging.

If you want, send me the log files (and bin files) to support-at-sharkplotter-dot-com and I'll have a look and post back here.

Cheers, Jim

Dictys

Thanks Jim, I'll send them across to you.

The first one is the stock tuned non 02 map, then we used the adjust feature to create a updated map which we the copied that data into the cat map and then logged the second set of data where the 02 is enabled. Maybe we copied the wrong map over?

jcorenman

Thanks for sending the files, very helpful.

Here's your first set of data logs in Sharkplotter, showing the original map values (no-cat map, open-loop with no NBO2 adjustment).

At first look this seems pretty good, but there is one thing that is very curious: The whole plot looks a lot richer than I would expect. Normally a stock map will be close to stoich (14.7) around idle and in the light-load/lower RPM area, then a bit lean in the moderate-load/RPM area (middle of the plot), and then quite rich at higher load/RPM. This plot is rich everywhere by a couple of AFR units, which is very strange. (The red stuff along the top is fuel-cutoff decel).



OK, I then duplicated your adjustment with SP's "Adjust Fuel Map" function against the default target map, and got the same result.

SP's adjustments are correct, if you look at cell 990/132 as an example, it originally had an average AFR of 12.4, to get that to 14.4 (the target for that row) requires a 14% change in fuel, and each map count is 0.2%. So the -19 for that cell goes to -90, approx.



That all seems fine, so why the heck didn't the second set of logs come in closer?

What I think happened is that your original AFR's were not correct-- either the WBO2 isn't properly cal'ed, or there is a scale difference between the WBO2 controller and Sharktuner. SP took those numbers as gospel, told you how to adjust the map to compensate, and wound up with wrong map numbers based on the wrong AFR's.

The other possibility is that your WBO2 cal is OK, and the NBO2 sensor needs replacing. They don't last forever, some say 60K miles or so, and then they get very slow and flaky.

If you haven't done a recent free-air calibration on the WBO2, then start there. Innovate says to do an air-cal upon installation and then every three months or so, but it is a good idea to do a fresh air-cal before starting a series of tuning sessions.

The other possible issue is that different WBO2 controllers use a different analog-voltage scale. The Techedge controllers have a default analog output range of 0-5 volts representing 9.0 to 19.0 AFR, while the Innovate WBO2's use an AFR of 7.35 to 22.39. (Your closed-throttle AFR of around 19, which means ST is probably set for an AFR scale of 9-19. If you have a Techedge WBO2 then that is good, but check it anyway...).

So be sure that Sharktuner is set up correctly for whichever WBO2 you have-- Tools menu, "WBO2 Law". (And it is also possible to change the analog output for the WBO2 controller, so check that also if not sure).

Also, I would recommend smaller steps: Don't go directly from the original map with open-loop/no-cat to a new map with closed-loop/cat-map-- too many variables are changing at the same time.

If you have a NBO2 fitted, then start by warming things up and watch the fuel-monitor page for a while (idling is fine), in the normal closed-loop mode. What you expect to see is ST's AFR reading hunting up and down a few tenths but centered around 14.7. You will also see O2-adjust hunting up and down as the LH adjusts fuel for an AFR of 14.7, as read by the NBO2 sensor. It doesn't especially matter what O2-adjust is (within the +/- 20% limit), what you are looking for is properly closed-loop adjustment, and an average AFR of 14.7.

If you see this pattern then you know two things:
(1) If the AFR and O2-adjust are hunting up and down then you know the NBO2 sensor is functional, and
(2) if the AFR is hunting around an average of 14.7 or so then you know the WBO2 calibration is correct (at last at stoich).
Those are important things to be sure of.

Then do your data collection in open-loop mode with the no-cat map, as you did. Then adjust that map in SP and copy it back to ST, into BOTH maps-- just to avoid errors/confusion.

Then repeat your driving and logging, but do this still in open-loop/no-cat mode. This is your validity check, and it may take a couple of iterations to get the fuel map dialed in.

Then, finally, do a test in closed-loop mode and see if you get similar results. (And again, make sure that you copy any map changes into both maps).

Cheers,

Dictys

Dear Jim,

Thank you for your help.

The NB signal is coming from the Techedge NBsim output. I'll redo the calibration and check the above settings.

In closed loop the AFR reading is hunting up and down around 14.7, however will do some logging and checking as you suggest before we refer back to closed loop.

I'll report back once I have collected more data, checked the above and tuned the maps further.

thank you

jcorenman

OK, that all makes sense. The only problem with using NBsim is that the AFR, in closed-loop, will always hunt up and down around 14.7 whether that is correct or not-- because that's how the WBO2 controller is programmed. There is no independent check.

But that doesn't explain why things went wonky with the adjusted maps, in closed-loop mode. Even if the WBO2 were off calibration, the NBsim would have agreed with it and should have been happy.

Let me know what you find. The fun part about this stuff is the driving, and the excuse to put your foot into it (not that any is needed ).
(Of course we're looking for an idle issue but that doesn't matter).

Me, running the computer: "OK honey, drop into third, slow to 2000 and then take it to redline".
Sue, driving: "But isn't that the sheriff parked right up there?".
Me: "Oops, never mind".

Dictys

Jim, thanks again.

I am enjoying the testing out in the desert I found a slight wheel shake at just above 150...

Here is this mornings result (02 disabled, every thing checked as above and open loop all the way), however, when I try to optimise it some of the cells (around idle) go out of spec.

Tomorrow, I'll copy the updated map into the cat map and switch to closed loop and then do some further logging.

Here is this morning unadjusted map

jcorenman

I guess you don't have problems with the sheriff in the desert

That map looks pretty good, but still rich around idle...

Two important things to check: First, when running in the "no-cat" mode, make sure that ST's "CO Pot in non-cat mode" setting is disabled (lower-left corner of the fuel-parameter page). Cars with NBO2 sensors do not have a CO pot and this setting must be disabled, otherwise there is a false CO-pot reading which will make a large error in fueling (and AFR) in the idle and light-throttle areas of the map-- just what you are seeing. I don't recall for sure which way the error goes, but I think it makes the idle-area too rich.

Also, make sure that ST's "O2 adaptation" is disabled and reset (click the "reset" button after ticking the "disabled" box). This is on the fuel-parameter page, lower-center of the page. If this is not done, then the saved O2-adaptation (plus or minus) will be added to the the fuel. This effects the entire map, and is how the LH compensates for aging MAF's and other tolerances.

Cheers,

Hilton

Interesting - it never occurred to me to worry about the CO pot.

I have it on both my cars. Should I reset the CO pot to the stock "new MAF" value (around 320 ohms IIRC), and then worry about tuning?

Or should I use stock maps to adjust the idle mixture by calibrating the CO pot to the condition of the MAF, then tune from there?

FredR

Why are you in CAT mode at all? No 928's were supplied to the Middle East with O2 sensors unless yours was imported/most unusual.

You should have the CO pot fitted which initially I used to set afr at stoich on the base [stock ] map. You should also calibrate the WBO2 in free air each session unless it is fitted/operated continuously.

I also noted on your sharkplotter output that the dots are the same size at WOT. This suggests to me that your WOT throttle switch may not be working unless you have done this deliberately.

When sharktuning this is no problem as you can build the WOT map into the base settings.

Apologies if I am missing something here.

Regards

Fred

jcorenman

Either will work. If your MAF's are "new" (i.e. recently rebuilt/recalibrated) then I would set the CO-pot to 382 ohms and tune the maps from there. Then, as MAF's age and things get "leaner", you can make small adjustments to the CO-pot to compensate, using the WBO2 to re-set the idle mixture.

If the MAF's have some age then your second approach would be better: With stock maps loaded, set the CO-pot for idle mixture in the usual fashion, then tune with the pot at that value. The CO-pot is then set to compensate for the MAF aging, and you are not asking the maps to do that compensation. As the MAF ages further you can make further adjustments with the CO-pot, and when the MAF gets rebuilt/recalibrated then the CO-pot gets adjusted again-- and should wind up close to 382 ohms, if this science-stuff works

Edit: A clarification from JDS: The nominal or "zero" value for the CO-pot is 382 ohms, that's where it neither increases or decreases fueling. Increasing the CO-pot resistance adds fuel at idle and light-loads, progressively decreasing as load increases.

Cheers, Jim

Dictys

Fred,

The car is a ROW with cats (however they are removed at present), it is not middle east spec. It is a LHD built for the Japan market, then imported over here in the 1990's. As the car will be travelling with me to various places in Europe and the UK I need to have the ability to install the cats back so I run car closed loop most of the time.

What does the dot size mean? the top end stuff on that last map was mostly at cruise, i tried to do no hard acceleration when collecting data for that map.

FredR

Jim configured sharkplotter to use bigger dots when the throttle switch is open. Just a visual aid as it were to spotting when you are on full throttle [well 3/4 or so]. When I used the ST2 for the first time I noticed that my full htrottle contact was not working so it was missing the increment added to give a bit more zip. When sharktuning I just set the cells at full load to give the AFR I was looking for [13.2 or so].

Regards

Fred

FredR

I doubt you would get that load signal value unless your right foot was buried [LOL]!

Regards

Fred