jcorenman

Cool! Send me some logs if you don't mind sharing, it will be interesting to see the data from the "Half-MAF".

Leave it disabled. I've left O2-adaptation disabled/reset forever (well since I got my hands on a ST), it confuses the heck out of everything (including me) and everything seems to work fine without it.

FYI there are two ways to switch to open-loop for tuning: One is add a "test-switch" in series with the NBO2 signal wire, to disconnect the NBO2 signal (leave the two heater wires connected). The LH sees the open-circuit and goes to open-loop mode but continues to use the "CAT" map, works great.

The second is to select "Force no-cat" in ST's fuel-parameter page. In addition to forcing open-loop, this also causes the LH to use the no-cat map so be sure to copy the cat-map to the no-cat map. And, as mentioned above, when doing this it is also important to disable the CO pot in ST (unless one is fitted).

And in either case don't forget to disable and reset O2-adaptation.

One question that has been lingering since I started the SharkPlotter project is whether it is possible to tune fuel in closed-loop mode by using the O2-adjust numbers. In theory adding the O2-adjust correction to whatever the WBO2 says for AFR will give you what the AFR would have been in open-loop, without O2-adjust. It works, but my conclusion so far is that the resulting AFR has more noise than collecting data in open-loop mode. SP will do this, just collect log data in closed-loop mode.

John Speake

Hi Bill,
I avoid talking about "going into open loop" when adaptation is defeated. It could be confused with going to the non-cat maps.

When adaptation is defeated, the O2 loop ADAPTATION is stopped, but the main O2 loop operates normally all the time, trying to keep the idle/cruise mixture at stoich. Just the O2 adaptation is inhibited.

It doesn't matter if you permanently defeat the O2 adaptation, when you have SharkTuned, because you have effectively taken up all the errors in mapping that adaptation is trying to do automatically, while driving.

The main O2 loop is inhibited when in Autotune - set to mid range.

AO

I'm typing this for a second time now... accidently navigated away...

Thanks Jim and John as always for your input.

I now think I know where things went awry for me. It all started with the injector size selection. Because of my twin throttle body setup, I can't just select 42# and go from there. I need to select something that's roughly half of 42. The way I do this is to start around 20 and then select up or down. Because I had not selected "Force non-cat" the main O2 loop was enabled and thus tweaking my injector size to some extent.

On my last STing experience, my injector size ended up being 19lbs. I now realize that the O2 loop was adjusting things down to get me to 14.7. Because I had a less than ideal injector size (too wide of a pulse) all of my auto-tuned values in the map were skewed to the low end of the spectrum (approaching -127). And then when going back to normal operation, the O2 loop could only compensate so much before it ran out of adjustment range. This is probably why I was getting the "O2 too rich" check engine light that I neglected to mention earlier.

How's this for an approach?
1. Select "Force non-cat", disable O2 adaptation, reset values
2. Zero out the idle portion (or entire map) of the main non-cat map - I'm not sure about this step
3. Start car and select injector size that gets me closest to 14.7 at idle.
4. Begin auto tuning
5. Then data log and apply shark plotter for "fine tuning"
6. Copy final map to cat map
7. Switch back to "Obey coding plug"

Am I missing anything? Make sense?

jcorenman

Understood. In theory 21# would be the correct selection, you might start there (as a custom size).

ST's auto-tune would have disabled the O2-loop (effectively selecting "no-cat"), but if you hadn't disabled and reset "O2 adaptation" then I believe that would still be active and messing things up.

This all sounds right, smaller injector-size in ST means longer pulse and more fuel.

I think that will work. And absolutely, step 1 is disable/reset O2-adaptation. But the ST injector-size parameter needs to be chosen to keep the whole map in range, not just the idle part. So I would be inclined to first set ST injector-size to the best-guess (21#), then auto-tune idle, then auto-tune with increasing loads and see if the map stays in range and not too far from center. If you get a lot of big negative numbers in the map then increase ST's injector size parameter (and start over with auto-tune). If auto-tune misses some cells then just fill those in manually, to bridge values it did get. And be careful as you work up to higher-load cells, make sure that there is enough fuel and that ignition isn't too far advanced.

Starting with a zero map is fine, although the ST's default map is probably just as good a starting point. ST's auto-tune will quickly get you into the right ballpark in either case.

Then carry on with your step 5, it will probably take a couple of iterations with SP to get things right. Then exactly as you said, copy no-cat map to the cat map and select the normal cat mode.

As an alternative, consider adding a "test switch" in series with the NBO2 sensor signal wire. This is a quick way to force the LH into open-loop mode for testing/tuning. You don't have to change the coding-plug selection in ST, and it stays on the cat map-- but disables the O2-loop (you will see O2-adjust sitting at zero- I usually log that, just to know for sure if the LH was open or closed-loop). You still need to disable/reset O2-adaptation however.

AO

Thanks Jim. That sounds like a plan. I wonder if I can sneak out for the rest of the day? naw... too obvious... *cough* *cough*

AO

Quick update:

I don't have my plots with me here at work today, but yesterday I was able to get some shark tuning done on my drive to Columbus and back. 69 miles each way.

I was able to apply the shark plottler twice. I have now copied the map over to the cat-map and re-enabled the "obey coding plug" setting. I can see the O2 adjustment working and it looks pretty good. So far no goofy readings.

I did notice, however, that the car is pig rich on cold start up. Looks to be in 10's. I'm sure this is a function fo the Warmup enrichment map which I haven't touched yet. Will bump this down a bit.

John Speake

Hi Andrew
The warmup map does tend to be on the rich side, better somewhat rich than weak, less liable to stall out. But usually it is possible to reduce the cells values once temp gets above about 20C

dprantl

Here's an interesting thing I found out about the warmup map yesterday. When my blue cursor is in the 95 deg C cell in the warmup map, adjustment of that cell has no effect on the AFR of the car. I can set it to 255 and nothing changes. When the cursor is in the colder temps, it works fine (at first I thought it wasn't working at all, so when it was at 75 deg C I hit ++ very fast and the car almost died of richness ).

Is the LH supposed to ignore the warmup map after a certain coolant temp is reached?

Dan
'91 928GT S/C 475hp/460lb.ft

John Speake

By the time the temp gets there, the car will have switched to closed loop. The switching point is controlled by water temp (usually switch at about 80C) , but there's also a timer on there as well. Sometimes it can switch to closed loop at a lowish temp, like 20C.

dprantl

Right. But I was watching the O2 compensation and it didn't change either. I would have thought that even in closed-loop mode, 255 * 0.4% increase in fueling would make the compensation factor freak and go to -20% at least. Also forgot to mention my adaptation is disabled and reset, has been this way for a week now.

Dan
'91 928GT S/C 475hp/460lb.ft

John Speake

That's because when the LH goes closed loop it totally ignores any values in the warmup map when making the calculation of injector opening time.

Disabling the O2 adaptation seems to be the way to go, once everything is optimised with the ST.

dprantl

Ah I see, I didn't know that. Thanks for the info John.

Also, in that case I remember Ken suggesting recently to add fuel to the upper ranges of the warmup map in order to compensate for leaning out when very hot. I guess that will not work if your car is running on an O2 sensor.

Dan
'91 928GT S/C 475hp/460lb.ft

John Speake

True, once the loop is in control, your are stuck with stoich. The only time you can alter the mixture out of that range is when the LH goes open loop under rapidly increasing load values or large application of right foot for WOT :-)

PorKen

This is only for '85-'86, where the last cell in the WOT map is 120C.

The ST program would need a minor revision to make this work properly for '87-up, where the last cell is only 97C.

It works for WOT, which is always open loop.



BTW: I tried running WOT in closed loop, by disconnecting the WOT wire, and using the two analog outputs of a LC-1, one set at 14.5:1, the other set to 13:1. At WOT, the input to the LH switched to the richer output. It didn't work well, because the O2 loop was too slow to react. AFR's went up and down (starting from where the fueling was just before WOT), slowly getting richer.

FredR

General question for anyone using ST/sharkplotter and running stock- what AFR [average values as determined by sharkplotter] do you find produce max power? Particularly pertinent question for anyone who has tuned on a dyno [I do not have one here]. I am looking for 13.2/13.3 average on top end given the SD calculated

On the positive side, I had a run through the mountains with a couple of GT-3's and a twin turbo last week. They were impressed with the 928's straight line performance [the car is awesome in the bends]

I would also like to talk to anyone running boost on an automatic model off line.

Regards

Fred