This log snippet happened whilst I was gently increasing the throttle about to overtake another car. The throttle and boost readings rose steadily but the MAF signal dropped out resulting in a knock sum of 20 with a lean spot and excessive timing. The jolt was I think a lean misfire. I'll be changing the plugs soon, the engine seems very healthy, I think this is a MAF dropout.

Anyone else seen this?

I can't reproduce it at will, it happens only once every few months, but now I use the pocket logging I managed to catch an episode of it.

Even the standard cars do a little MAF over-read when you lift off gradually from light throttle as I think there is a reverse pulse going through the MAF. This makes them overfuel - in this case it went the other way.

 

My car does the same thing. I've had to drastically reduce the timing in the 120-160 load sites, particularly in upper RPMs, to avoid the knock. I have no idea why my car needs a map like this when so many others that I tune do not.

So I'm no help in figuring out WHY this happens, but I can back you up in saying that you're not the only one.

 

The both of you are joined by another..

I have a few data logs, displaying what your does. The interesting thing is I also have data logs that show the exact same thing (MAF dropout) but with no knock.

Like you, Pd1, I have been having to take some timing out here and there in the same load colums you described. I am deviating from the OEM maps in sometimes drastic timing changes when in the areas under load values of roughly 120 are unchanged. Maybe it is due to the reduced resolution of the maps because of my scaling changes.

I have been looking at all of the other values to see if something triggered the MAF to suddenly drop in value drastically.

 

I first noticed it during WOT throttle lift. If I lifted slightly at the 5k, 6k, 7k, range, it would knock like crazy. It wasn't until I took out a LOT of timing that the knocking subsided. It makes my 3D map look odd, but I haven't been able to figure out why this is needed, or how to correct it.

Most recently, I've noticed it in light accelleration in low gears. There's apparently more work to be done at the lower RPMs in this (I'm guessing) load area as well.

I've not paid attention to the MAF signal when this happens, but I will now since you say that may be related.

 

I previously had the lift off knock situation, but got rid of this by retarding the intermediate load areas. The MAF doesn't misread in these situations.

I have two MAF misread situations:

1. Over-read on gentle lift off from part throttle part boost - doesn't cause knock, happens on standard cars - triggered by bringing it to about 1 PSI boost and then gently lifting at about 3000-4000 RPM, often in a high gear.

2. The one I've posted in this thread where there is a much more pronounced jolting episode with a MAF under-read - goes lean, runs too much timing and shows knock sum. In the above log it looks like I was lifting off as boost was building, this is when it seems to happen like this - perhaps the MAF is getting a reverse pulse and/or the turbo is surging instantaneously?

 

John,

What DV are your running? Maybe too tight or sticking?

It sounds like it can possibly due to the DV being a bit tight, and not opening as it should, causing a brief stall/slowdown of the airflow. That would land you in much lower load cells and give you your lean spike and increased timing.


Eric

 

Standard dump valve. IX one is quite stiff though.

 

I also thought it was my WORKS upgraded IX DV, so I swapped it back to stock. However, this didn't impact the intermediate load knock.

 

Pd1, could it be a funky MAF?

As much as I despise "throwing parts at a problem", maybe try swapping in a different MAF to see if that cures it.

 

the rise in timing seems abrupt for so little a drop in load, that mixed with the sudden drop in afr,

I would consider leaning the rest of the accel since you are below 11, perhaps your cylinders are primed for knock.
I realize the 12 indicated on your log isn't bad, but maybe a drastic number is lurking between the readings.

 

Looks like you were part throttle boosting, lifted on the gas a smidge, BOV vented a little, still part throttling, MAF got some BOV air in the system and missed a few counts, engine bucked, knock sensor hit 20 (due to the bucking or engine mount hitting frame sharply), quickly back to 19, then to 0. The knock was just a short spike and the decrement counter in the ecu took it to 19 then to 0 as quick as it could.

It really isn't a big deal. A MAFT in the UICP solved my problems like that on my DSM, in the evo the metal evo 9 bov seems to blow off much smoother than the other mitsu BOV's I've had so I don't have that problem.

 

I think your description is right dan l, but I think the knock only went to 0 because I lifted out of the knock control range, although I agree the noise was probably very short and sharp. What worries me most (whether it was detonation or noise)is how lean it went - judging from the IPW and the narrowband O2 that goes to full lean whilst still under boost. The AFRMAP of 12 would be fine if the airflow reading was correct, but it appeared to go leaner than 14.7:1 based on the narrowband whilst on boost for that short moment. Thankfully it doesn't happen often, but it makes me want to convert it to speed density...

 

I get a choking type thing, this is what I do to replicate - works 50% of the time...:

* get on the throttle say 30% wait for a bit of boost to build (say 3-5 psi)
* get off throttle for 1 second
* get back on light thottle..

car will gurgle and feel like its accelerating more than the ammount of TPS im giving it... i have yet to properly log this event...

it started after I changed my frontpipe...

 

This situation is exactly the one I am dealing with the most because it happens when I am crusing around 60-65mph on the highway, and do a 50-75% roll on the throttle to overtake someone. Usually this is between pulls, while still data logging.

I tried last night to find some of the pulls, but like Pd1, I have taken out timing to keep this from happening. Finding the data logs is not an easy thing to do seeing as though I am at almost 900 data logs now.. (Yipes..)

 

I drove my 1g for 50,000 miles like that. I crushed my stock BOV and did the GUS mod so that it could hold 35psi of boost. Mine was bad however, the engine would actually shut off for 200ms or so then come back on. Its not a big deal and doesen't damage the motor.

Also the ecu should have two decay routines. One under high load and one under low load. The one under low load decays about 10x faster than the one under load. It looks like when you drop below a certain load threshold that the knock stops, but it is just decaying much faster. The proof is that you can still get knock at 100% load (actually 1.0grams/rev) when driving normally. However it decays much faster rate. You are correct though about at a certain point their being no knock control. It has a rpm threshold, load threshold, and TPS threshold if I remember correctly. Basically if your at idle knock won't cause your car to stall out, that is about all that one is for. Even under very light cruise knock control is usually still active.