jcorenman

Victor, that was apples and oranges.

Go back and take a look at my comments on Jeff's plot: The "lean strips" are narrow, and deep-- they extend all the way to the highest load, with AFR's that run from 13.3 to 18.4 in the same cell. Pretty alarming, actually.

The narrowness of the "stripes" indicates a strong resonance. As discussed at length above, it could be air (e.g. MAF errors) or it could be fuel (e.g. changes in fuel pressure). Jeff did a series of runs where he accelerated through those "stripes" at constant throttle. I picked out some of those series of points, the MAF signal was a straight line (indicating that the MAF was telling the truth) and the injector PW was relatively constant. If the airflow is constant (he wasn't changing the throttle and the MAF didn't change) and the PW is constant, yet the AFR changes, then about all that's left is fuel pressure.

And yes, Jeff has changed a bunch of stuff and I am sure that is related.

What's interesting to me is the use of Sharktuner for diagnostics as well as tuning. The lean spots would show up on an AFR gauge, but I am not sure how you would relate it to anything without being able to sit down with the plots.

The plot I posted of our S4 was quite different. Yes, it shows similar variations in fueling with rpm, but broader and only extends down to MAF load of 90 or so... For a standard MAF that is pretty light load. That all seems consistent with some sort of intake resonance. In any event, being a light-load phenomenon it more a curiosity than a concern.

When I posted that I was still thinking that Jeff's "lean stripes" could be related to that, but his more recent plots disabused me of that notion.

ptuomov

Also:

large injectors and light load => short injector pulse width => less averaging of the fluctuating pressure in the fuel rail => large potential impact of fuel rail resonance on the fuel delivered.

In the opposite extreme, if injectors are small and load high, then the injectors operate close to a continuous injection system at near 100% duty cycle and there are rarely any fuel system resonance problems.

jcorenman

Actually Sharktuner's map pages have a 3D display option. What I am working on are some new tools for Sharkplotter, which I think will be fun.

Jeff's fuel maps are quite smooth, especially through the "stripey" regions, he's done a good job there. I'm not sure I understand the relevance of the EZK maps, but those looked OK also. I've been using them here for testing, I don't know if he wants to share them.

jcorenman

Not this space, some other space

victor25

OK, I hear what your saying. And I could see how a resonance could happen in the fuel rails, but not at so many levels. To happen at many rpm levels, and to happen all the way through the injector pulse time, seam to me to be almost untheoritical. I mean the injector pulse is a broad range of milliseconds from just off idle to WOT. Each would make a different pulse with through the fuel rail at every millisecond level.. yes? And the fuel would be moving through the rails at different levels also. So we have rpm, pulse length, pulse distance, fuel movement speed, and fuel pressure. There are so many variables that's like winning the lottery. I could see a pocket of resonance, but not multiple long bands. I'm no fluid dynamics engineer, but I'd take that bet

Lizard928

Jim,
The ign map has a huge effect on AFR.
Here is an experiment, have car running with say 5 deg adv with 14.7:1 AFR. Then ramp the ign adv up to 28 deg adv. It will require MUCH less fuel to achieve the 14.7:1 AFR.

Reason is that the mixture ignites, how much advance determines how much of the mixture is burnt before it leaves the cylinder. The extra little time under pressure in the combustion chamber makes a HUGE difference in the AFR giving a constant inj pw.

So if he had a cell that was retarded, yet a smooth AFR, it could result in a lean condition.

Speedtoys

I hope to have a ton of data tomorrow.

Plan:

ON the load dyno..run down each rpm column in the fuel map, centered on the RPM itself...while ignoring load in the first two rows, and first 3 columns initially.

Create logs of the dyno info, and on the sharktuner as well...and then..see what happens.

Once its all dialed in, per cell...
Run thru each column again, so that the paired Dyno and ST logs (tuned) can be used and compared by Speake/etc.


Question:
I...Dont see much over 140 in the MAF absolute numbers at high RPM..what AF's should I be tuning for (non cat mode, no adjustment of course) on the dyno for the fuel map cells?

At "some point" the LH will put itself into the open loop mode..so..what do I do?

ptuomov

I am not an engineer either, but I would take that bet. The almost vertical stripes are exactly what you'd expect, slightly tilting with the load because load changes the injector pulse width and effect being generally stronger at low loads because of less (time) averaging over a shorter pulse. I don't think there's any magic here. It can always be some other resonance that we haven't thought of yet, but I would absolutely take an even odds bet that it's fuel system resonance that's causing the stripes.

jcorenman

Well, from 5 degrees advance to 28 then sure. But it's not really lean, it just looks that way to a WBO2. At 5 deg advance the mixture will still be burning going out the tailpipe, with lots of not-yet-burned O2 passing the WBO2 sensor. The sensor sees the excess O2 and says "lean". But the fuel is there, in the correct proportions, it just hasn't burned yet. (And not doing anything to generate power, either).

It's sort of like reading lean on ignition misfires, the fuel is still correct but it looks lean on the WBO2 because the O2 didn't get burned.

But what we've got here is advance going from 21 to 20 degrees from 1900 to 2400, while AFR's go from 14 at 2300 to 18 at 2400. It's not an ignition issue.

ptuomov

There's a very remote possibility that somehow the electronics are misfiring some fraction of the time at those specific rpms and those misfires are incorrectly registering as lean readings on the WBO2 sensor.

The way to eliminate that possibility is to use an exhaust gas analyzer on a dyno.

As a short cut, you could also hook up the WBO2 after the cat. The cat burns some of the unburned fuel and oxygen and misfiring is the problem then my logic says that you should get a less extreme of a measured lean stripe on the WBO2 reading after the cat.

Hilton

Jeff's car is an auto? Could it be the EZ-K timing retard when the box is shifting?

SeanR

I keep meaning to ask how we shut that part off but always forget. Thanks for the reminder.

Speedtoys

Send porken your EZK BIN file, have him do what I asked him to do.

Speedtoys

Jim/Colin/Tuomo:

https://rennlist.com/forums/10037664-post82.html

Thoughts on where to tune the AF map..since I only use the top half of it under real load?

Lizard928

I have figured out how to get some 3D mapping ability with the STII using a couple of programs.
So this means do some adj, then transfer the map, smooth, transfer test and repeat.

However one of the reasons that I think this is needed becomes very apparent when you actually upload them. These graphs were taken from someone that emailed me the files to look over.
These came off a running car......