PorKen

Both, as I understand it.

Lower load levels, yes, by changing the MAF threshold number in the code above.
(It is richer only because I have additional code to make it so by changing an injector constant - I leave the base table at 14.7-ish. I use ~3000 for the RPM threshold as this makes better use of the first S4 torque peak.)


Note that MAF units, as well as RPM are on a logarithmic curve. The rough 3.6 or 4 multiplier is my own shorthand as I am useless (lazy) at figuring out logarithmic functions. I have a lookup sheet for RPM, but not MAF.



One beef I have with the stock O2-adapt threshold is there is no hysteresis. If you manage to cross the border, it is quite easy to fall back down below it, whereupon the O2 loop kicks back in hard at its previous bias level, which may be quite lean.

FredR

Not so sure I understand the fuss about adaption, maybe it is a legal requirement and if so fair enough but as someone who has only driven around in open loop models maybe I am missing something.

I am still of the view that there is something not quite right with my motor in terms of power development and quite possibly the problem lies inside my head. By running open loop all the time you can select what AFR you want when cruising along at low rpms/light load factors and crank it up when on it.

When pushing the go pedal I find that there is only a small array of cells that come into play- below 2k rpm is mostly irrelevant and if anyone thinks it is relevant then I wonder what you are doing in a car like this. MAF values from 200 to 300 seem to do the job on my motor and only in the highest rpm band do I see MAF values in the range 300 to 330. Of course this in air temps at 25C- if we ever had 15C maybe I would see bigger numbers. I do not use the WOT map [I was suspicious of the wide open throttle switch contact- it does seem to work] and I am happy with that arrangement. What I can say is that if I run my EZ with the stock maps I feel a noticeable degradation- I have not tried to run a stock LH map with my bigger injectors as I am not sure how well it translates.

That there may be better systems available on the market today I have no doubt but then I find the configuration in ST2 very user friendly. I dare say someone more experienced than myself may well get more out of it than I can but I have driven stock GTS's that did not feel anywhere near as lively as my 90 S4 motor.

I follow the AFR regime ST2 recommends but on top end I tend to run a little richer maxxing at around 12.5 as per the attached plot. I have noticed some anomalies at top end and perhaps I will try Ken's coding modification shortly. Similarly I have gone back to tuning since I disconnected the breather from returning to the inlet tract so maybe that would give me more advance potential coupled with the coding mod.

Regards

Fred

worf928

This is a very interesting re-start of this thread. I wish I could spare the time to give it my full attention. In general I agree with Tuomo's desire for more insight into what's really going on in the code. I tried to get that insight many years ago when I was working on the tune for D.R.'s twin screw. I had to resort to black-box testing.

On to this specific context:
I discovered early in the process that adaptation and the NB02 adjust will screw up a nicely-tuned fuel map. No ifs ands or buts.

The solution D.R. and I hit upon was to install a pressure-actuated switch downstream of the twin-screw that triggered the WOT signal. In my case I disengaged the pins for the WOT signal from the EZK and LH connectors and pinned-in the switch's output line to the LH (leaving the EZK with no WOT input since its single row WOT timing map was not useful.).

I assume the 'collective we' realize that when the WOT signal is triggered 02 adjust is discontinued.

This solution enabled me to get a nice fuel map with no 02 adjust for cells that were active at >~1PSI manifold pressure and allowed me tune those cells for power while at the same time preserving 02 adjust for 'lower' cells. In cases where a cell could be 'active' in both conditions, I tuned the cell for power and found that the 02 adjust was sufficient for near-14.7:1 operation when at lower manifold pressure. Typically, these 'in between' cells were not active very much of the time.

Edit: Obviously, this solution is of no use to NA engines.

jcorenman

Dave,

One of the issues with DR's setup, if I am recalling correctly, was that he was splitting the intake and running only half the airflow through a standard MAF. Do I have that correct?

What that does is to get the fueling in the right ballpark by measuring half the air and using an injector with twice the flow. That's great, except it also fools the LH about the load scale. So when you start getting on boost, the LH -- reading half the air-- thinks you are still cruising down the highway. Oops.

The pressure switch for WOT is a fine work-around, and I think Andrew does that also. The purpose of the WOT switch is to signal high-load to the LH, and (especially for a boosted car) intake or manifold pressure is arguably a better measure of load than throttle-position.

Another solution is to use a superMAF and measure all of the air, so that the LH knows what the real load is. That's what Tuomo is doing, but the concern remains-- you need to be confident that the LH is running open-loop when there is significant load. And in this case the LH-- along with the WOT switch-- is determining what high-load is, and shutting off the O2-adjust. Niklas understands how that all works, I am learning, and I am confident that we can sort out a nice solution.

The other thing to keep in mind is that adaptation was developed for production cars, to avoid having to tweak the mixture for each car. Folks who provide a pre-programmed chip set with each S/C kit have the same issues, and using adaptation is great. But for cars that get individually tuned, this pretty much becomes a non-issue. There are many thousands of Euro cars running around with no O2 sensor, no O2-adjust, and no adaptation-- just a pot to tweak every 10-20K miles. Tuomo wants adaptation to work for his application because it is cool and that's great, we'll make it work.

Cheers, Jim

victor25

Fred you may want to take a little ignition advance out of the timing on the top end. That's a whole bunch of knocks you have there.

I don't use the use the WOT map either. I tune the cars without it, as the switch usually doesn't work and if the do, who know when it'll fail. besides if you up in those numbers on the map, your foots probably on the floor anyway

Hilton

Those thousands of Euro cars also have no cats. As soon as the factory added cats, they added the O2 sensor (e.g. 1990 in the UK).

Cats are the reason I'm interested in this whole deal - I bought a set of cats to add to the x-pipe on my 89 partly to reduce the smelly exhaust in traffic, and partly for noise control (it has small resonators). I'm aware I can still get the benefit of cats without an O2 sensor, but I'm concerned about longevity too.

I found this interesting.

ptuomov

If you set "Oxygen sensor" to "Force cat mode" and "Oxygen sensor adaptation" to "Disabled" in the ST2 "Fuel parameters" tab, I think the following will happen: The LH will operate in closed loop mode at low load and rpm, pulling the AFR towards 14.7, and to 14.7 as long as it is within the allowed adjustment range. Yet, the low-frequency "adaptation" correction would be always zero. The car would be running near 14.7 when in closed loop mode and then automatically go to the richer open loop map when the LH decides the driver wants to drive hard. With these settings, the high-frequency O2 loop on average pulling off fuel wouldn't cause any unwanted corrections in high-load regions of the map.

Hilton

Interesting thought - it hadn't occurred to me I could save the EPROM with those settings. I usually only look at them at the start to make sure they're as required (i.e. off) for tuning, and when done, reenable and put back to "obey coding plug".

worf928

Yes. And that has virtually nothing to do with the context of my message which was 02 adjustment and 02 adaptation.

So... basically it's a great idea for an FI engine running LH 2.3.

And has nothing to do with how the air may or may not be measured.

It's not a work-around given the constraints. And of course, it is the constraints that make Tuomo chaff just like they did me.

I don't see how the addition of a superMAF by itself solves the problem. The only case I can think of where this would be true would be too serendipitous to plan on.

Unless there a 'new' superMAF that is also wired to the WOT plug of the engine harness along with a new version of the ST software that enables programmatic control of 'WOT' conditions.

The existing system that triggers WOT on 87+ when the pedal reaches the last 1/4" of travel is dangerous for 928 motors with FI. The FI 928s get to boost with little pedal travel. You need a power mixture once manifold pressure approaches atmospheric. Or pistons will become 'hole-y.'

Folks in the gallery need to understand that these assertions are specific to the context of a 928 motor running LH 2.3. Comparisons with modern cars are non-starters. New cars have mixture control orders of magnitude better than LH 2.3 on a 928. S4 motors are a little less sensitive than GT and GTS.

Next, folks in the gallery need to understand that in the context of FI engines WOT does NOT mean pedal to the floor. It means: the conditions under which stoichiometric operation is not desired, or not safe.

He needs a 'switch' other than the pedal reaching the stop. And he needs that 'switch' to kill 02-based trim dead.

He understands this. I understand this.

I understand this and have for a long time. What you do not know is that I developed the base chip set for D.R's system. All of his installed kits, with the exception of Andrew's, started out with my maps. The 16v and non-inter-cooled 32v maps were, of course, different in final form. But they started life as transformations of my inter-cooled 32v maps which D.R. then tuned.

I wanted 02 adaptation operational for all the right reasons. And not operational, at times, for all the right reasons. In particular I wanted 02 adaptation operating in the low-load cells because D.Rs system only measures half of the air and at low load the setup variances of the twin throttle bodies dominate the system. o2-based trim deals with that variance nicely.

OF COURSE, I would like to measure all of the air. But, that will not solve the 'o2 problem' when FI cars need a power mixture.

Again, my context was not individual tuning.

And, even if it had been, I then, like Tuomo does now, would have wanted adaptation to be active.

And how many are FI? It's not as simple as that.

I don't think that's why he wants it. But, he can decide to answer or not.

I like the can-do attitude. I wish you had been here in 2008.

ptuomov

Jim's 100% right -- this will be individually tuned car and for me it's just that I don't want to make the car worse in any way. For all practical purposes, the adaptation being off is not going to hurt me. But who said this is about practical purposes -- not me!

Now, for John Kuhn, it may be slightly more important if someone just buys a kit from him. It would be nice to ship a chip that has adaptation enabled. But even that is a theoretical possibility at this point, anyone in their right mind will buy one of these ST2 systems before starting to double or triple the stock power with stock computers. It's amazing what can be done with ST2 and a quarter century old computers.

Triggering the WOT map with a Hobbs (boost) switch in a FI car is a good solution. I can't see a downside for a car with a positive displacement supercharger, it's just better than the throttle position.

For a turbo car, it's not as clear cut. I definitely want the LH to go open loop when there's boost above certain threshold. However, I also probably want it to go open loop when I floor the throttle and the turbine hasn't spooled yet. So a more ideal solution for a turbo car might be to have either the actual WOT switch or the Hobbs boost sensor to ground the WOT pin in the LH, creating a simple OR logic.

Yet, it may be possible to do this with just the software. For a forced induction car, you can almost surely tell from simply the MAF signal whether the car should be in closed loop or open loop mode because of boost. For example, I can pretty much tell the boost level from the MAF signal, for a given RPM. In fact, the factory LH 2.3 logic is almost exactly what the dual "OR" WOT switch and Hobbs boost sensor would do: If the LH sees either some combination of high rpm and high load OR the WOT switch is triggered, the car goes to open loop. All that is needed for a software solution that will replicate the hardware solution is more control in ST2 for the parameters that govern the triggering of the open loop mode.

Coincidentally, exposing the parameters that govern the triggering of the open loop mode would also solve my adaptation problem. I would simply first figure out the map region where I would be comfortable tuning to 14.7. Then, I would set the LH to trigger open loop outside that region (or with the WOT switch grounded). Now, the adaptation will "learn" to change the fueling only when the components age, and not otherwise.

Thus, I am repeating my suggestion to expose in ST2 the parameters that govern the open-loop closed-loop switch in LH.

jcorenman

This is correct. Also, you don't need to select "Force cat mode" if the coding plug is configured for cats. Remember that overriding the coding-plug selection with the "force" option also changes which fuel map is in us.

Unless you are using the second map for something special (lean running, track, whatever) then it is a good idea to keep both maps the same and avoid confusion. (I get confused easily).

Everything on the fuel-param page is persistent, i.e. saved and used whether ST is connected and running, or not.

jcorenman

Actually it does. When you take a standard MAF and reduce the flow by half, the LH sees that as less engine load. Or conversely, as you start to get into boost, the LH (measuring half the airflow) thinks you are still cruising down the highway and leaves the O2-loop active when it shouldn't be.

The LH goes open-loop and stops using the O2-sensor when either the WOT switch closes, OR the engine load (measured by the MAF) exceeds a threshold. Measuring half the air "fools" the LH into not going open-loop when it should.

If the superMAF is still only measuring half the air then I agree-- nothing is different. But that wasn't my premise. I was just suggesting that measuring all of the air is helpful, which would require a superMAF in this application.

Backing up:

The basic issue is that the O2-sensor adjustment and adaptation, while useful and necessary in some contexts, need to be limited to the low-power parts of the fuel map. At high load the mixtures need to be richer, determined by the fuel map without the "help" from the O2-loop. I think there is no disagreement on that.

The LH has two mechanisms to accomplish this: The WOT switch is one, and engine load (MAF signal) is a second. Either will disable the O2-loop and force open-loop operation-- AND stop updating adaptation-- this is the OR-switch that Tuomo postulated. And both were originally designed for stock engines and presumably suitable for that task (we can debate that later).

In the context of boosted engines, you are absolutely correct that the WOT switch, alone, is useless as a measure of "load"-- you can cram a lot of high-pressure air through a relatively small throttle opening. So using a pressure switch instead of a throttle-position switch makes a lot of sense, everything else aside.

A much better measure of engine load of course is the MAF signal, that's its sole job and purpose in life. And above some load thresholds the LH does indeed disable the O2-loop and uses the map values directly. The question is where those thresholds are, and whether they are appropriate for boosted motors.

That's what Niklas and I are off in the corner discussing. We've actually been working on a bunch of stuff, to be discussed shortly. This is a new topic for me and will take a bit to sort out. But we will share, be patient.

Cheers, Jim

jcorenman

You are indeed repeating yourself...

First road test: Little dots are closed-loop with O2-adjust active, big dots are open-loop (O2-adjust = 0). This was just our wussy S4 but interesting nonetheless.

For the selected points, the throttle was rolled on to about 2/3 up until 230 load, then snapped wide open (Th=3) and then chopped (twisty county road).
The LH goes open-loop and stops messing with O2-adjust as load goes past 165, which is where I had set the thresholds.

Hilton

Interesting plot Jim, thanks for taking the time to do it.

So if I'm reading your post right, you've manually set the "go open loop" load threshold yourself at 165? How did you make that change?

Did you by any chance run a similar plot showing the stock load threshold where the ECU goes to open-loop?

I need to take a drive up the freeway north of Sydney to data-log for cruise-conditions and see where on the maps it sits on some of the steep freeway uphills in 4th and 5th. Looking through my logs I have handy from local roads, I'd be tempted to put the threshold around the 180 load area.

Sadly I'm in the office this week, so no "working from home" drives to enjoy my new wheel alignment.

jcorenman

Correct, as noted that was a first road-test of a Sharktuner test-version from Niklas that provides adjustment for the O2-loop thresholds. I set them lower for testing for separation from where the WOT switch normally operates.

No, but it wouldn't show much. For the stock configuration, the O2-loop is disabled strictly by the WOT switch below about 3000 rpm. For higher RPM's either increased load or the WOT switch will disable the O2-loop and force open-loop operation.

Our plan is to get this to Tuomo for testing in a real boosted environment, there is not much to look at for NA cars.

Cheers, Jim