Speedtoys

To answer this post and Colin's:

I set the flappy hardware back to OEM operation to "see and test". RPM and load values are OEM as well.

I can easily plug the vacuum source to the flappy to be full time open in the AM, no problems. I have 2hrs of my normal mountain plus flat land cruise to do tomorrow to see what this does.


When I hit them, i'm accelerating up into them, and I can set the cruise control on them as well. *stumble stumble stumble*

Speedtoys

I put it back to be more methodical on what im doing...and..we're here.

Lizard928

Ahhh ok good to know that it got wired back up.
I pulled the plate right out of my car.

jcorenman

The trouble with this theory is that the lean-stripe at 800 is actually around 900, and 1600 is either not there or very weak. The 2400 stripe is quite sharp and sure looks like resonance, as does 3600-- but that last stripe is actually at 3400, not 3600.

I still like the resonance theory, but I am not sure how to fit it with that data.

OK, so it is not a transient phenomenon. As Tuomo said, if it is intake-related then it should change if the flappy is open, or closed--- you can change that in ST (if the flappy is connected). Or disconnect the solenoid and either connect the vacuum full-time, or not at all, for testing.

A big problem here is too many changes and not enough data.

But even a stock setup shows some signs of resonance... here is a plot from our S4 from a couple of years ago with lots of data points. You can see a weak resonance every 300-400 RPM or so. I make resonance lines at 1330, around 1550, 1815, weakly at 2100, 2300, 2660, 3050 and a broad area around 3200 (picking RPM's off the SP plot with the cursor). In this case it only happens at light loads (flappy not active?) and isn't nearly as sharp. And doesn't always show, some days it is not there at all. I've seen this effect on other cars as well, including our GT-- which interestingly went away completely when we got rid of the MAF (with no other intake changes).

ptuomov

If it's the fuel system resonance, then the resonance rpm should change predictably with the injector pulse width. The closing of the injectors causes the pulse, and the time from the injector closing to the next opening depends on, among other things, on the rpm and load. That's why fuel system resonance rpms aren't exact multiples of some number and at higher loads the resonance rpm would be higher (consistent with the slight angle in the data).

I think the first thing to do is to log the injector pulse width, MAF voltage (or whatever is the best to compute the air mass), and the AFR form the WBO2 censor. Then focus on the area where the problem is clearest, 2400 rpm, and try to keep the car in a steady state. Cruise control on different gears is a great idea. If the measured air flow and injector pulse width at 2400 rpm are both approximately the average of those at 2350 and 2450 rpm, yet the WBO2 shows a lean spot, then I think it has to be fuel pressure fluctuations in the fuel rail caused by resonance.

hacker-pschorr

+1

Chatting with Victor at SITM I suggested this if he ever experiences any odd tuning issues (such as this).

IMO everyone with a modified air intake should be using one just to be safe. Keeping the air more uniform is always a good thing.

This isn't a "band-aid" fix like some have mentioned in the past. Manufactures are using them now. My 2003 Yukon has the same straightener in front of the MAF from the factory.

Lizard928

Which more so confirms my theory of turbulence in the MAF body,
Though this will show up with MAF signal, and inj pw.

Speedtoys

Locating one that will work with victors IC would be a challenge.

What options exist that match our MAF size?


Or...Jim...live with it for a few bits, then go MAP?

ptuomov

This is $0.02 worth of internet engineering, but I believe that the MAF is responsible only a small fraction of tuning problems that it's been blamed for over the years. It's kind of like the modern version of the saying "90% of carburetor problems are electrical."

For example, the 928 throttle body combines the pulses from eight different cylinders. To me, this means that the pulses must cancel to a great degree. I don't understand how the combined pulses could confuse the MAF, regardless of the cams or whatnot. There's some disconnect between my understanding and what I read on the web about this, and either one is wrong.

The honeycomb ahead of the MAF is not going to hurt, it might even help a little bit. Replacing the wire mesh with a honey comb can work as well as the wire mesh but with less flow resistance. However, I would be very surprised if adding a honey comb in addition to the two stock screens would make a measurable difference to this problem, or any clearly identifiable problem with the 928.

With both screens on feeding the stock 928 intake manifold, I've come across one and only one situation in which the MAF signal may give an incorrect reading. That one situation is when the throttle plate is slammed shut. This creates a pressure wave that cools differently from typical flow, and it may confuse our hot-wire style MAF. Throttle being slammed shut is a very special condition which I believe one can tune around such that the cells visited in this failure mode give sensible results. This is the exception to the "working MAF works perfectly" rule.

Contrast this with the fuel system. The engine batch fires all the injectors at the same time. As the injectors shut, this causes a pulse wave in the fuel system. If fuel pressure is high, the pulse wave is stronger, so this should be a bigger deal with boosted cars that have the fuel-pressure regulator referenced to the manifold pressure. If the injectors are large, the impact of the pulse wave can be large. This is because large injectors are open a very short time at low loads, and the pulse can cover the whole injector opening time. For longer pulses, the injector opening averages over a broader section of the wave. At high loads, the red columns fade away in both Jim's and Jeff's graphs, consistent with the longer injector opening time averaging the fuel pressure over a broader section of the wave.

I am of course just speculating here, what do I know about pulses in fluid -- nothing. Still, it's a better theory than what anyone else has come up with.

jcorenman

This doesn't make sense. Aren't we talking about pressure waves and resonances in the intake plumbing, i.e. organ-pipes? I would think that everything has to do with valve timing, and nothing to do with fuel injection.

MAF signal [absolute] is what you want to log, that is the LH load. And that is easy to do, and the data may already be in the logs-- all you need to do is accelerate past 2400 (e.g. from 2300-2500) with constant throttle, lift and slow to 2300, rinse/lather/repeat. With constant throttle the load vs RPM should trace a (relatively) straight line on the plot. If it is going lean because the MAF aiflow is falsely dropping at 2400 RPM, then that will show.

The question is why. I am not so sure that I buy this turbulence theory, in Jeff's case. For a stock intake, yes: you've got an air filter on one side of the MAF, but that's pretty transparent to airflow. And the transition into the MAF is far from ideal, with sharp corners and airflow coming from two sides-- so yes, turbulence makes sense.

But Jeff has a single long pipe from the SC, and an intercooler just upstream of the MAF-- I would think that's acting as a pretty good air-straightener/diffuser setup. And I don't recall this being an issue with other similar SC intake setups.

But I think taking a careful look at the log data will help answer that question.

jcorenman

The other theory is fuel pressure, artifacts related to the dead-ended rails and single damper. That's a large change from the factory setup (and common practice) with flow-through rails and dampers at each end.

I am not saying it is good or bad, and I have no idea if that is related to this problem-- just that when something is different then it needs looking at. The problem is that measuring dynamic fuel pressure is not easy, but if the logs rule out MAF load and Tinj changes, then that might be the next thing to look at.

ptuomov

What I was writing about is resonance in the fuel rail.

That's what I was suggesting in the earlier post:

"Contrast this with the fuel system. The engine batch fires all the injectors at the same time. As the injectors shut, this causes a pulse wave in the fuel system. If fuel pressure is high, the pulse wave is stronger, so this should be a bigger deal with boosted cars that have the fuel-pressure regulator referenced to the manifold pressure. If the injectors are large, the impact of the pulse wave can be large. This is because large injectors are open a very short time at low loads, and the pulse can cover the whole injector opening time. For longer pulses, the injector opening averages over a broader section of the wave. At high loads, the red columns fade away in both Jim's and Jeff's graphs, consistent with the longer injector opening time averaging the fuel pressure over a broader section of the wave."

That is about the resonance in the fuel rail, not in the intake manifold.

These are not just decorations:

jcorenman

Sorry, you are right-- I missed that change in topic.
We're in heated agreement then, I think. Let's see what Jeff's data shows.

Speedtoys

Ya..I will have 2hrs of driving this afternoon..a normal day.

1/2 up 4000' of curvy mtn road, 1/2 flat highway where I can set cruise at the bad points/etc as directed.

Lizard928

The MAF is very close to the throttle plate, so at partial loads where the plate is only partially open. There will could potentially be greater turbulence in that area.
Also Murf does have honeycomb in his 90 which leads into the MAF......
With my turbo car prior to converting to VEMS with MAP sensing instead. I would get one spot where about half the time it would create a judder at one set RPM. I narrowed it down to the MAF signal going wonky (right as I maxed it out ).
A good portion of the factory MAF's these days come from the factory with the honeycomb, and their wires are much more protected than ours.

Jeff maybe look for a piece from here http://www.hongzan.net/html/Products/Honeycomb_Core/