Why am i still getting knock retard?! this is insane

[itsdaveonline]

So in summary I'm comfortable to where my tune is. It slightly on the rich side, but I keep on getting knock counts. It is not Burst knock but rather the computer perceives it as the real deal. IMO It's totally random.

Here is everything I've done to try to remedy this:

-replaced a knock sensor
-Running stock timing values
-Stepped down to TR6 plugs @ .040
-stepped from 7.400 to 7.425 pushrod to quiet down valvetrain (verified w/ pushrod measuring tool). Preload is in spec too
-tune is slightly rich for safety margins
-Virtual VE is within 1% calibration
-MAF is usually 1-2% within calibration typically on the richer side
-Visual on old TR5 sparkplugs show no detonation / preignition etc
-audibly speaking I don't hear any detonation or pinging (Rocks in a can sound)
-predictive co-efficients disabled
-replaced fuel filter
-seafoamed motor
-ran fuel system cleaner
-ALWAYS run 93, this is not an isolated bad gas incident. Ive tried different stations
-changed oil
-engine not consuming oil or leaking into cyl
-purchased and calibrated new wideband (AFR 500) verified the unit LCD is within .10 AFR of analog out

This is driving me fucking insane. This is stock ls2 motor with cam and headers. Is there ANYTHING in here which looks remotely suspect for knock. I'm scared shitless to disable the knock retard or lessen sensitivity What's your guys angle on getting this fixed.

FYI wideband AFR is ac pressure sensor. 9+[PID.7101]*1.400

[itsdaveonline]

This is a long log file. Here are approximate frames where you can find knock retard:

Frame 8160 2400 RPM
Frame 8679 4100 RPM
Frame 8761 3800 RPM
Frame 10374 4900 RPM
Frame 10624 4000 RPM
Frame 12780 4500 RPM
Frame 18781 4100 RPM (lots of knock)
Frame 21019 2200 RPM
Frame 31523 4900 RPM (1st and 2nd gear blast)

Here are some spark plug pics. These are iridum TR5's which were run for 5000 miles. In the log im running brand new TR6's. All of the plugs had a sooty tint, no melting, black specs, or glossy porcelain.

[GHuggins]

Under Spark-Retard-Knock Retard Attack there are two tables I'll modify very slightly. The ECT gain and Base Retard tables. You can subtract .25 from both of these without worry. Then under Knock Sensors, Initial Level - multiply by 1.1 - again nothing to worry about - most handhelds will double this table. Then you can increase the individual cylinders multiplier tables by 1.05 at a time in the cylinder airmass areas where your still seeing knock.

I see nothing in the logs indicating "true" knock. BUT the best way to make these adjustments is on a dyno to see if any power is actually gained from dialing them back and through adding the timing. With your mods, the headers will make the knock sensors more sensitive in a sense just because they'll pick up noise from them...

[itsdaveonline]

Under Spark-Retard-Knock Retard Attack there are two tables I'll modify very slightly. The ECT gain and Base Retard tables. You can subtract .25 from both of these without worry. Then under Knock Sensors, Initial Level - multiply by 1.1 - again nothing to worry about - most handhelds will double this table.

Awesome thanks, i will give this a shot (tuning down the knock sensors). So if there is genuine knock, will the PCM will still generally "catch it" with these settings your suggesting?

What are the other tale tale signs of knock / ping you look for in the logs? I will be keeping timing @ factory till I get on the dyno

[GHuggins]

It should still pick up regular knock with this. Most handhelds that you can buy over the counter modify some of the tables far more than this. In fact you might still see random knock even with these settings. If it's genuine knock it will typically show up at the same rpms and engine load circumstances... Not just all over the place, especially with stock timing tables...

[Higgs Boson]

Just run it out of gas, pour in 5 gallons of MS103 or the like, and check for knock. If you get any it's false. Repeatedly desensitize the knock sensors by 2 or 3% only in the affected areas until it goes away, make sure to test WOT also. After you run out of race gas, half tank of 93, run it out, then fill up all the way with a fresh tank of 93 and if you get knock it's real.

[sssnake]

Just run it out of gas, pour in 5 gallons of MS103 or the like, and check for knock. If you get any it's false. Repeatedly desensitize the knock sensors by 2 or 3% only in the affected areas until it goes away, make sure to test WOT also. After you run out of race gas, half tank of 93, run it out, then fill up all the way with a fresh tank of 93 and if you get knock it's real.

IMO the safest approach. Sorry you are still having issues.

[itsdaveonline]

I'll give the unleaded 100 octane a shot. I just hate the idea of driving 30 miles just to get some race fuel. It's a long trip from youngsville to cary.

I'll be even more disappointed if the kr goes away. Then I'll truly be at a loss. I have literally checked almost everything which would cause kr.

So last stupid question. By "desensitizing" are you just suggesting that I increase this table by 2-3% each time until kr is gone?
Spark > knock sensors > initial level ?? Not sure what you mean by affected areas

[Higgs Boson]

I'll give the unleaded 100 octane a shot. I just hate the idea of driving 30 miles just to get some race fuel. It's a long trip from youngsville to cary.

I'll be even more disappointed if the kr goes away. Then I'll truly be at a loss. I have literally checked almost everything which would cause kr.

So last stupid question. By "desensitizing" are you just suggesting that I increase this table by 2-3% each time until kr is gone?
Spark > knock sensors > initial level ?? Not sure what you mean by affected areas

depends on the OS, etc but which way to go with it is in the description. also not sure if it is initial level, post a screen shot of your Knock sensor tab.

[itsdaveonline]

Its a E67, here are all the retard and knock sensor items i can adjust:

[Higgs Boson]

adjust the Multipliers table according to the description (up or down).

only adjust it a little bit and only in the areas that correlate with where you are getting knock (load vs RPM or whatever).

[PDYoung]

I get knock when logging idle when i shut my door kind of hard, i guess the knock sensor is picking up me closing the door and i'm not slamming it shut. Just make sure it's not false knock

[itsdaveonline]

Okay i see the tables you want me to adjust but im not even clear on the description... It says:

"Multiplies knock sensor threshold when not in transient mode". I'm still unclear if I adjust these up or down by 2-3%. I would think reducing those values will reduce knock sensor sensitivity?

[GHuggins]

Increase them in the higher load areas or in the load areas your seeing knock. If you notice the oe all but ignores knock in lower engine load areas.

[GHuggins]

You can also go back to my original post for reference.

[itsdaveonline]

You can also go back to my original post for reference.

Doh! I didn't catch that you guys were talking the same table.

Now im on the same page though. Just wanted to make sure i knew exactly what to adjust before filling up with super expensive gas and making a 30 mile trip to the pump lol

Thanks again. I will report back with results of 100 octane and some knock sensor adjustments

[schpenxel]

Increasing=less sensitive. You're basically raising the "bar" (threshold) for how much noise the sensors have to see before it's considered knock. Or that's the way I understand it anyways

I had the same problem on my LS2 so I understand the frustration. I had a lot fewer instances of random KR when running SD only for whatever reason. I increased the multiplier tables on mine a little while running torco to try to get rid of all the false knock. If it were real I would have blown something up by now I believe.

[10_SS]

Or just pull 5 degrees of timing from that entire area and see if it goes away. If you want to really see then pull 10 degrees. Normally, if you have 5 deg or retard, pulling 3-4 deg of timing will make it disappear.

I see very specific amounts of knock occuring in a localized areas.. so if 5 deg doesn't do anything noticeable, pull 10. Or just pull 10 right away. If that doesn't work, then it's false knock... Then drive 30 miles to that gas station and work on your knock sensitivity.

If you still have cats on the car, try removing them if it's somewhat easy. I was left stranded one time due to clogged cats... the exhaust had 0 leaks so I had no clue... just didn't have power and knocked allot.. then eventually stalled out, would barely idle.

[schpenxel]

I had the HO table set to well under the stock LO table and still had KR all over the place, so.. pretty sure that was false (was running 93 + torco as well).

Replaced plugs, wires and knock sensors. No change.

[itsdaveonline]

Yeah im running stock timing in bulk of the areas and in some spots 2 degrees less timing than stock. Removing 10 deg more could be devastating lol. I actually did inspect the cats not long ago. They are not clogged and clean as a whistle.