Both tephra and I see knock in the same load/RPM range as you mentioned. For me, its mostly a 1-2 count kind of thing with the very occasional 4-5 count. I see it for a wide range of timing values, and it happens both for gasoline and E85. Based on these observations and the fact that it happens in a very narrow RPM/load range, I decided it was phantom knock.

 

jcs, I recall something mrfred posted, about another timing map overlay.
This one added timing during warmup. Could it be the cause of your issue?

 

mrfred, was/is yours just in warm up too?

 

It happens with a warm or cold engine. merlin.oz is right, there are two different timing overlay maps. One has retard values, and the other has advance values. The advance values can be pretty significant, but the engine has to be cold to use the values, and for my ROM there is no advance in the cells we are discussing. Screenshots are here:

https://www.evolutionm.net/forums/ec...ing-trims.html

 

I dropped the load for knock control to 100 at 2000 RPM which extends the control range down to my spool up area. It isn't picking up any knock (or false knock) there, but nice to have it listening given how much earlier we can spool than stock when we make light mods to these engines.

I nudged up the 2500 RPM multiplier and adders just a click each (not enough to stop it correcting proper knock, just gives a bit of space between knock base and knock ADC in typical use). So far, so good, will need to test for about a week to see that it is sorted as it is very occasional.

 

Mr Fred, have you checked to see if it always adds timing for the addition map when cold, or is it only for certain conditions like maybe accel enrich?

 

My 2500 RPM issue seems to be fixed, even when trying to provoke it, but it can still register a sporadic 1 or 2 which I'm quite happy with as I have mapped it to do this over the entire rev range so that I can get maximum safe timing as often as possible.

I hardly ever hear my engine knock as I always aim to have knock sums <=3, but as I have a little spot at 7000 RPM that is lively since I changed fuel type and had a slightly higher VE on SD than the engine needs, I had a spike to 7 knock sum today, 99 RON, 21 PSI, 33C charge temperature, 12 degrees timing, about mid 11s AFR, so nothing wild, other pulls managed a knock sum of 2 or 4 in the same area. It was interesting that I could hear this in the cabin quite distinctly even without det cans, as I've had so little knock over the last three years with this car and engine to be able to test it. I think I only ever had one other episode that wasn't lift off knock, this was triggered by changing the boost controller setting to a much higher boost level mid-pull, and the knock sum was something like 12. It isn't terribly hard to hear even with an aftermarket exhaust as long as you don't have music or the radio on, but you need to be attuned to the sound of knock.

 

Sorry for reincarnation,

bit 0 is set if engine RPM is over 6500 (@20AC = 0xD0)
bit 4 is set if engine RPM is over 3800 (@1E3E = 0x7A)

setting of bits is actually at loc_17FC2 in sub_17F62

 

Thanks for clearing that up

 

So just to be sure I am understanding:


13 is used below 3800 rpm
20 is used betwenn 3800 and 6500 rpm
18 is used above 6500 rpm?

 

I wanted to do a check to see if any issues have come about since modifying these tables.
Thanks.

 

This thread needs to be bumped. It has been a great help to me and I agree with your advice here.

I've been tuning with a friend (Evo VI w/ VII ecu on 98640014) and the car had been showing consistent 1-2 knock counts around 3500rpm @ 1bar with base map tuning. The car is running 91 RON on meth with about 4* at peak torque...nothing crazy for a car running meth and the boost is pretty low because we just started tuning and started at wg pressure. The motor is forged and runs a Full Race Pro Stock manifold on a IX turbo among other mods.

The funny thing is when boost was upped to around 1.1 bar (about 16psi) we noticed the knock increased as well:


At 18/19psi it got even worse...the knock sensor would go nuts 5-10 counts and start pulling timing.


No matter how much we decreased timing it wouldn't go away at any of these boost levels past 1 bar. EGTs and plugs looked fine, nothing could be audible heard in the cabin and the car isn't loud....yet the logs showed counts always starting at about 3500 rpm.

At first I though an engine damper might have been the culprit but that proved wrong. Like someone pointed out, I began to think it was related to the noise and harmonics associated with that specific RPM. I read through this thread and changed the low multiplier from 13 to 18 (about a 40% increase) and mid from 18 to 20.

Reults: Knock is down to about 1-2 in pretty much every log so far. The car runs so much better with no more timing being pulled. We're still playing with boost settings so the load here is a little lower for comparison but identical boost settings to that in the first image. I'm very pleased with the results. It's very subtle but notice how the timing advance graph is a little smoother with absolutely no changes to the maps themselves.



I was very adamant about modifying the values unless 99% sure it wasn't real knock in the first place...but based on the logs only this region of the rpm caused problems and for no apparent reason. I really want to thank you for this find...it's a great tuning asset if you're experiencing knock issues that can't be resolved. From my results thus far I don't think it'll be necessary to change the Knock Sensitivity Load Threshold table.

 

Probably still worth checking with some det cans to be absolutely certain I think.

 

Hmmm... I've been starting to wonder if I'm running in to the same kind of issue with my car. I get a knock at 3500 rpm pretty much any time I pass it whether I'm running hard or not.

At first I thought it was a peak boost issue causing the ping, but I took my boost down a bit, kicked the afr down a bit (low 10s), and am only running 1 degree of timing advance, but I was still getting 3-4 counts of knock every time.

I've been doing some cruising logs lately and I noticed that every time I pass the car over 3000-3500 rpm whether accelerating or decelarating, it registers knock, even if I'm in vacuum at the time with no throttle input! So I'm wondering if it might be just the harmonics of the engine at that rpm. Has anyone else had an issue like this?

 

If anyone has tip-in knock and has used the old Lean Spool information to activate Lean Spool via , BoostEnhancement (anti-lag) please look in https://www.evolutionm.net/forums/ec...5-10-a-25.html and start at post 973.