Turning off knock sensor under certain conditions?
This would be the best way, you could then read the plugs and see whats happening to an extent to see if its real or phantom.
It would definitely be worthwhile to have this mod, especially with high mileage, noisier lifters, aftermarket cams, springs, etc, that may throw things off a bit.
DSMLink had this ability and I had noisy lifters in my DSM (over 100,000 miles). I made sure it was phantom knock first, but once determined, I simply disabled the knock sensor in low RPM/load ranges. The car drove 100% better during daily driving.
Since most cars with phantom knock issues seems to have them at low load areas, I think a patch like this would be a great addition. Of course, before using this new feature, you should be absolutely sure that it is phantom knock first, either by checking the plugs as Lucas mentioned, or via timing adjustments, higher octane gas, etc, etc.
Eric
DSMLink had this ability and I had noisy lifters in my DSM (over 100,000 miles). I made sure it was phantom knock first, but once determined, I simply disabled the knock sensor in low RPM/load ranges. The car drove 100% better during daily driving.
Since most cars with phantom knock issues seems to have them at low load areas, I think a patch like this would be a great addition. Of course, before using this new feature, you should be absolutely sure that it is phantom knock first, either by checking the plugs as Lucas mentioned, or via timing adjustments, higher octane gas, etc, etc.
Eric
what about a user definable rpm dependent scaling factor for the knock sum, you could essentially tune the knock sum this way.
example,
if between 1000-2500 rpm the ecu sees say 14 counts, have a table where you could apply a scaling factor of say 0.5 and export that value (knock count of 7) to what the ecu sees for timing reduction purposes.
maybe this could be incorporated into a patch
just a thought
example,
if between 1000-2500 rpm the ecu sees say 14 counts, have a table where you could apply a scaling factor of say 0.5 and export that value (knock count of 7) to what the ecu sees for timing reduction purposes.
maybe this could be incorporated into a patch
just a thought
Good thought on that one!
I can see a lot of advantages by doing that on the stock ECU.
I can see a lot of advantages by doing that on the stock ECU.
Looking at my logs...most of the time, it just throws ridiculously high number of knocks, retards timing but the OCT number remains at 255. At this point in time, I just want to turn the damn sensor off during low load. I will take the 0.01% chance that it isn't phantom knock
last bit in ecu perephery 0 - if changed to 0 it turns the whole knock control off - at least this is what I ve understood on the aktivematrix.com discussions - but I am too scared to try it 
For 96530006 - perephery 0: stock value is 56da - if you change it to 565a - knock sensor is disabled.
definition of those value in Ecuflash for 96530006:
<table name="ECU_PEREPHERY0" category="Misc" address="faa" type="1D" level="4" scaling="Hex16"/>
<table name="ECU_PEREPHERY1" category="Misc" address="fba" type="1D" level="4" scaling="Hex16"/>
<table name="ECU_PEREPHERY2" category="Misc" address="fca" type="1D" level="4" scaling="Hex16"/>
<table name="ECU_PEREPHERY3" category="Misc" address="fda" type="1D" level="4" scaling="Hex16"/>

For 96530006 - perephery 0: stock value is 56da - if you change it to 565a - knock sensor is disabled.
definition of those value in Ecuflash for 96530006:
<table name="ECU_PEREPHERY0" category="Misc" address="faa" type="1D" level="4" scaling="Hex16"/>
<table name="ECU_PEREPHERY1" category="Misc" address="fba" type="1D" level="4" scaling="Hex16"/>
<table name="ECU_PEREPHERY2" category="Misc" address="fca" type="1D" level="4" scaling="Hex16"/>
<table name="ECU_PEREPHERY3" category="Misc" address="fda" type="1D" level="4" scaling="Hex16"/>
Last edited by evo828; Jul 22, 2008 at 12:02 PM.
I was going to turn off the knock sensor completely when I made an interesting observation. My knock counts dropped by a fair amount after I flashed in the same map with absolutely nothing changed in order to reset the fuel trims. Now I'm wondering if its the ECU that's acting up after being flashed so many times...I probably have around 300 flashes on mine currently
Are you having problem with part throttle low load knocks too?
Are you having problem with part throttle low load knocks too?
I have over 3000 flashes on my stock computer and dont have random issues yet. It might be something else that was causing the problem. There were no changes at all though in the tune?
Account Disabled
iTrader: (38)
Joined: Aug 2005
Posts: 9,319
Likes: 1
From: Virginia Beach, Virginia
sometimes a flash will go awry during the process and simply reflashing will clear it up. I've had customers turn the key off during etc...pretty scary but I've never had one not recover.
In my DSM, I've disabled knock timing pull below what an Evo will represent as 100% load. You will NOT hurt a 4g63 doing this due to knock.
It would be nice to dial in a certain amount of timing pull per knock at the higher loads. Like 1 degree per count, or 1/7 degree per count. Or if you want to turn it off, 0 degrees per count.
The ideal situation would be to make your own knock filter map, but it seems this is much to complicated for the time being.
It would be nice to dial in a certain amount of timing pull per knock at the higher loads. Like 1 degree per count, or 1/7 degree per count. Or if you want to turn it off, 0 degrees per count.
The ideal situation would be to make your own knock filter map, but it seems this is much to complicated for the time being.
Maybe its other issues than the stock ECU but I am not sure.
I am still running in my motor but I have noticed with the run in map that it picks up some knock on 1 or 2 part throttle pulls...
Not sure if its going to be a problem, but I am aware that Forged motors seem to pick up a lot of Phantom knock and this will mess around with the timing... limiting the true potential of the motor.
Just thought that I should be prepared if it does become a problem when I'm tunning...
.
In my DSM, I've disabled knock timing pull below what an Evo will represent as 100% load. You will NOT hurt a 4g63 doing this due to knock.
It would be nice to dial in a certain amount of timing pull per knock at the higher loads. Like 1 degree per count, or 1/7 degree per count. Or if you want to turn it off, 0 degrees per count.
The ideal situation would be to make your own knock filter map, but it seems this is much to complicated for the time being.
It would be nice to dial in a certain amount of timing pull per knock at the higher loads. Like 1 degree per count, or 1/7 degree per count. Or if you want to turn it off, 0 degrees per count.
The ideal situation would be to make your own knock filter map, but it seems this is much to complicated for the time being.
Would be nice to be able to set that up to pull X amount of timing per X amount of knock as danl stated.








