During the tuning process using evoscan to logg and ECUFlash to make changes, the knocks that do get sum of 2-3 seem to always be at the end of the WOT, right at the point where throttle goes 100% to 0%. Why is this? Seems to me that at that point air should come to a stop but fuel still will be dumped. So it should run just a bit rich and not lean. It has to be phantom knock probably from vibration of the sudden loss of acceleration right?

 

I see this all the time. Mine does it often, i have also seen this on a few other Evo's as well. I'm not quite sure what is causing it. It's deffinatly phantom knock in my case and the cases of the other i have seen while tuning.

I would like to find somebody with a stock EVO to log this same condition. i'm thinking it has a lot to do with having a aftermarket downpipe/exhuast. It could also very well be the slaming forces of the engine rocking back after the run. Causing some noise that the knock sensor is picking up. It's kinda hard to say at this point.

the knock sensor usually gets more sensitive at high RPM's, this would cause the knock sensor to be more sensitive to phantom knock as well.

CJ

EDIT: is this on your ecplipse? If so, thats interesting.

 

Some people have suggested to reduce timing in the low load/high rpm areas to prevent "lift off detonation". If you search on that term I think something will come up.

Good queston.

 

i am familier with this theory. I however, have not bought into this theory.

CJ

 

Have you tried reducing timing to see if the theory works or not? Because there are people who have tried it and it works.

MB

 

sounds like a perfect MYTH BUSTERS" topic, I cant see that working, the high RPM not filtering out the delta frequencies of sudden drop of acceleration is just something we got to live by I guess, its not worth changing the knock sensor data. Plus if it was a real knock instead of retarding the timing, wouldn't it be better to up the accel enrichment map, I take it this map works both with positive acceleration and negative?

 

I experienced the 'lift-throttle' knock after I leaned out the A/F from stock in the upper rpm range. I seem to recall taking out some timing in the 120-180 load range which cured it.

 

I was asking about this same thing in my thread the other day...as in if anyone has been able to get rid of it entirely....

 

+1

 

y niot

 

Reduce timing in the load cells where your getting lift off det.. you might need to reduce your timing a little before the knock actually happens and in that case it will be on the right side of the logged load instead of on the left side.

 

It is detonation, you can hear it tinkle in the cabin if it is bad. You can hear it in det cans. You can see knock spikes up to 14 when it happens bad. You can see it on a knocklink. You can get rid of it by reducing boost, increasing octane, or mapping the lift off areas carefully. All properties of detonation - false knock doesn't respond to these measures.

When you close the throttle the ignition is still active and it ramps up very quickly - too quickly on occasion. However, there is still air and fuel in the inlet manifold. Some ECUs have programmable attack rates for this problem I gather.

Accel enrichment is only positive, and knock sum is based on differentiation of the knock signal. The ECU doesn't retard in response to these knock sums on lift off. This may be because it isn't a problem at stock boost, or because it does no/little damage (although I don't like to hear it in the cabin when you provoke it). The areas where it occurs do have knock control under increasing load though.

When I asked an Evo tuner about this he said it didn't happen on his tunes because he had the full load AFR and timing correct. However, I received a datalog from one of his customers because he was worried about the knock counts. The full load timing and fuelling were safe, all the big knock counts were on lift off. The only mapping method to fix whilst not losing performance is to have a saggy bit of ignition timing in the 3000-7000 RPM areas 120-180 loads. I'm only really there when feathering the throttle in a corner or gentle accel really anyway, and I didn't notice a loss of part throttle torque. With further improvements on breathing on my car the problem has gone away running stock JDM timing or even a bit more in these areas, but the full load stuff I can advance much more relative to where it started.

There are no myths to bust, I've spent ages working this problem and have it nailed.

 

But that would be considered a "block" timing approach which we know, from reports here, is bad. Just kidding


I wonder if aftermarket DV's may even make this worse by keeping the manifold pressurized abit following throttle plate closure - meanwhile load drops down into the 180's from airflow (and at 30* timing!) but the charge is sitting still at full boost.

 

are the cylinder pressures at lift off even high enough for the lift off detonation to do any damage?

 

No.....the throttle plate is closed or closing.... there is very little load on the engine at this point. You should worry a lot more about knock on the 2 step, which is a load/unload situation and can easily get tons of knock.

CJ