ECU knock algorithm effective at 2500-3500 rpm??
#1
ECU knock algorithm effective at 2500-3500 rpm??
I've seen quite a few Evo IX logs both on the forum and via PMs, and no matter what the ignition timing value is, there seems to be very little knock at 2500-3500 rpm, usually no more than 3 knocks. This is true for ign timing values ranging from -2 deg to 5 deg, all with pump gas and fairly high boost values.
There's been some analysis of the knock measurement algorithm over at Aktive Matrix. Apparently, the ECU only listens for knock at a specific crank angle, something near 15 deg ATDC. I've been wondering for a while now whether this algorithm where knock is measured at a specific crank angle can work over the entire RPM range, and if not, what RPM range has Mitsubishi sacrificed. Maybe on Evo IX's the crank angle where measurement takes place is too far ATDC to effectively hear knock at 2500-3500 rpm?
Also, maybe the crank angle where measurement takes place is different for VIIIs and IXs? If the crank angle is closer to TDC on VIIIs, this could nicely explain why VIIIs are less tolerant of timing at low rpm, but more tolerant of timing at higher rpm.
There's been some analysis of the knock measurement algorithm over at Aktive Matrix. Apparently, the ECU only listens for knock at a specific crank angle, something near 15 deg ATDC. I've been wondering for a while now whether this algorithm where knock is measured at a specific crank angle can work over the entire RPM range, and if not, what RPM range has Mitsubishi sacrificed. Maybe on Evo IX's the crank angle where measurement takes place is too far ATDC to effectively hear knock at 2500-3500 rpm?
Also, maybe the crank angle where measurement takes place is different for VIIIs and IXs? If the crank angle is closer to TDC on VIIIs, this could nicely explain why VIIIs are less tolerant of timing at low rpm, but more tolerant of timing at higher rpm.
#2
I've seen big knock in the 2500-3500 RPM on my JDM IX, mostly on lift off, but I'm sure I've had it on full load as well before it was tuned out. It is a shame because I put some 95 RON (similar to 91) fuel in the other day and had a log of it at 22 PSI on my usual 99 RON map which I run about 4 PSI higher. Even so it just about kept the octane value at 100%, perhaps because it was running in my retarded lift off areas on full throttle at the lower boost. I often use the 2500-3500 RPM full throttle area in my daily drive because the performance is adequate for quieter passing of slow cars in the country areas in which I drive. Trouble is I changed the SD card I use on my Pocket PC that had the logs on it. I will keep a lookout.
I did think that the usual 1 or 2 coming onto boost was most likely due to the rate of change of the knock sensor noise level as it does seem fairly insensitive to ignition timing.
I'd be a bit worried about it not correcting knock well in the 2500-3500 RPM range. I did have a knocklink and det cans on the car previously, but I dropped and lost the knock sensor. I can hear significant lift off knock in the cabin, I hope I could hear severe detonation on full load too, but never have when listening as the ECU silences it all. The timing I use in the 2500-3500 RPM range is very similar to stock FQ360 which comes with a 3 year warranty. Doesn't mean there might not be an issue though!
I did think that the usual 1 or 2 coming onto boost was most likely due to the rate of change of the knock sensor noise level as it does seem fairly insensitive to ignition timing.
I'd be a bit worried about it not correcting knock well in the 2500-3500 RPM range. I did have a knocklink and det cans on the car previously, but I dropped and lost the knock sensor. I can hear significant lift off knock in the cabin, I hope I could hear severe detonation on full load too, but never have when listening as the ECU silences it all. The timing I use in the 2500-3500 RPM range is very similar to stock FQ360 which comes with a 3 year warranty. Doesn't mean there might not be an issue though!
Last edited by jcsbanks; Nov 21, 2007 at 02:31 AM.
#6
Evolved Member
iTrader: (9)
With my VIII, from 1500 - 3000 rpm is where it knocks consistantly, like every time. Interestingly, the highest knock sums occur from about 3"/hg to 5psi. The higher the boost, the more the sum tapers off, and this is not timing induced either. If I maintain the charge at that pressure in one of these episodes (5psi) the counts will continue on up the rpm but will disapear with a WOT application.
I've spent alot of time trying to pinpoint, i.e. xylene for fuel, EGR closed, EGR full open, of course timing, fuel, also tried downpipe wrapped in a glass blanket, catback removed.
I gave up on it letting it go as a harmonic noise at a specific knock filter value making it too sensitive at given background noise. Perhaps due to an inconsistancy between the preceived load on the stock knock filter tables and real load. Who knows
I've spent alot of time trying to pinpoint, i.e. xylene for fuel, EGR closed, EGR full open, of course timing, fuel, also tried downpipe wrapped in a glass blanket, catback removed.
I gave up on it letting it go as a harmonic noise at a specific knock filter value making it too sensitive at given background noise. Perhaps due to an inconsistancy between the preceived load on the stock knock filter tables and real load. Who knows
#9
Evolved Member
iTrader: (2)
Join Date: Mar 2003
Location: Santa Cruz
Posts: 2,218
Likes: 0
Received 0 Likes
on
0 Posts
interesting. This is where 90% of the time that I see my #$%#$ p0300 code thrown on my '03
conditions are always light throttle, light-ish load, coming on the throttle or coming off it.
tephra,
is there anyway given specific conditions TPS, load, open/closed loop that we can key off of to just ignore the knock event?
I can provide many logs that clearly identify when the p0300 is thrown, and when it's safe to ignore.
conditions are always light throttle, light-ish load, coming on the throttle or coming off it.
tephra,
is there anyway given specific conditions TPS, load, open/closed loop that we can key off of to just ignore the knock event?
I can provide many logs that clearly identify when the p0300 is thrown, and when it's safe to ignore.
#10
Evolved Member
iTrader: (23)
chrisw,
I had the same darn p300 code forever. recently changed turbo and in the process left off the metal underpanel and readjusted the intercooler pipe. As both were banging against other metal parts. The only other thing I have done is lower timing in low load/rpm.
...The p300 is missing so far...
I had the same darn p300 code forever. recently changed turbo and in the process left off the metal underpanel and readjusted the intercooler pipe. As both were banging against other metal parts. The only other thing I have done is lower timing in low load/rpm.
...The p300 is missing so far...
#11
Evolved Member
iTrader: (2)
Join Date: Mar 2003
Location: Santa Cruz
Posts: 2,218
Likes: 0
Received 0 Likes
on
0 Posts
problem is the buggy p0300 can also be thrown by harmonics coming from the transmission. I have an '03 and didn't find out about this until after my warrentee was up.
I suspect most of my problem are from a 95k mile old transmission which needs all the TSB work applied (and a rebuild in general)
I just recently rebuilt the motor, the p0300 problem is just as consistant as before.
I suspect most of my problem are from a 95k mile old transmission which needs all the TSB work applied (and a rebuild in general)
I just recently rebuilt the motor, the p0300 problem is just as consistant as before.
#13
Evolving Member
iTrader: (19)
Join Date: May 2006
Location: Ohio
Posts: 336
Likes: 0
Received 0 Likes
on
0 Posts
I have an 03 and the only knock I see is between 3-4K, mostly 3500rpms. We have ruled it out as false knock, as adjusting the timing made very little difference. I would see 3 counts with -2* at xxxxrpm and, and now with the ECU locked out, I'm still seeing 2-3 counts at the same rpm, with 2* of timing. Even lowering the boost made little, if no difference.
#14
Evolved Member
iTrader: (25)
I just got done with an experiment tuning session. I'm almost always getting 1-3 counts of knock at 2500rpm @ 120 load on WOT, so i kept an AFR of 12.5 there and adjusted the timing from 14 all the way down to 5, and the knock was always there. It's probably false but it sucks, cuz it's pulling timing as rpm goes up as the knock decays.
#15
Here's an example of someone lifting the head at low rpm with no knock:
https://www.evolutionm.net/forums/sh...88#post5046988
https://www.evolutionm.net/forums/sh...88#post5046988