Getting rid of FALSE knock for good?

[Blackdevil77]

My car keeps getting false knock. It is certain it is false, that is for damn sure. We de-senstized the knock sensors and it goes away for a bit. Than slowly comes back started at about .1 to .3, than eventually getting to a little over a degree. Why does it keep coming back? Do they knock sensors learn and adapt?

This is for a G8 GT with a cam btw. I can not post the tune, it isn't my work, its my tuners and he worked very hard on it. I'm just wondering if I am missing something with these knock sensors.

[ayousef]

take one degree off where its false knocking, test, if its still there its false knock, youd probably need to edit the decay rate for starters, and there are more than just a few tables that can fix this issue, yet Im no pro either.

Also, take a look at knock sensor level initial and minimum.

[Blackdevil77]

take one degree off where its false knocking, test, if its still there its false knock, youd probably need to edit the decay rate for starters, and there are more than just a few tables that can fix this issue, yet Im no pro either.

Also, take a look at knock sensor level initial and minimum.

yup I have done that test and the knock was still there. Through all the testing we have done, we have taken up to 4 degree's of timing out in a whole area for one spot with around .1 to .3 degree's of knock, and it changed nothing. On the dyno, still gained power when bumping the timing. We de-sensitized it and everything was fine. After a day or two, it slowly creeps back like the knock sensors are getting more sensitive again.

[99LS1SS]

I lowered the sensitivity of my knock sensors and it helped some. I'm not sure if the L76 has the same knock sensors as the LS3 but, if it does, they are very sensitive. I chased false knock for a while until I lowered the sensitivity.

[Blackdevil77]

I lowered the sensitivity of my knock sensors and it helped some. I'm not sure if the L76 has the same knock sensors as the LS3 but, if it does, they are very sensitive. I chased false knock for a while until I lowered the sensitivity.

It does. My tuner confirmed this to me and told me the same thing

same here. I got my sensors very desensitized now. And the false knock is gone for now. But eventually will probably come back just like before. It's like the sensors learn or something.

[madvette08]

The computers do have the ability to learn knock. the l76 and ls3 have the same knock sensors. And i would be careful on how you desensitize the knock sensors.

Ian

[Blackdevil77]

The computers do have the ability to learn knock. the l76 and ls3 have the same knock sensors. And i would be careful on how you desensitize the knock sensors.

Ian

Any tips? Is there a way to keep them from learning and becoming more sensitve? It's getting ridiuclous to keep having to check to see if the false knock is coming back

[madvette08]

There is a way to turn knock learn off, i have never tired to mess around with it. It is in spark control > retard > knock learn.

Ian

[Blackdevil77]

ok I see that. Just not sure what to do with it lol. Thanks at least I know what needs playing with

[tillman speed]

+1, would love to know how to disable this.

Have doubled and halved the decay rate as well as the attack. Have slightly altered the maximum knock retard (while still allowing the lower right cells to pull plenty if neccessary). Have tried adding as much as 30% to the minimum and maximum knock levels and it helped, but it comes back as well.

Would love to know how to disable knock learn, so that I can bring the minimum and maximum levels back to stock or slighly above.

Also details on the best way to safely lower knock sensor sensitivity in the case where ACTUAL FALSE knock is present.

[LSX378]

The knock sensors and the related signal processing on Gen IV's are very good, and flexible. For the average non tuned engine GM's knock tuning allows them to run safely at the edge and no doubt maximise mileage.

The knock multiplier tables are the best area to focus on if you think there is false knock there. Your logs will help map the specific cells against the specific cylinders that are on the edge and you can de sensitize against per cylinder air flow rates and rpm's. Further adjustments can be made to cover transient situations like a rapid hit on the throttle or richer less knock prone mixtures.

But! Real knock needs to be monitored closely to ensure you are not opening pandora's box. Even if you cant hear knock in the driver's seat, doesnt mean there is not intermittent light rattling or timing induced piston noise happening.

[Blackdevil77]

Now with this false knock, it doesn't seem possible to pick up any slight, REAL knock, as the valve train noise is louder than the combustion. With the false knock, its showing knock retard before it would even be able to make REAL knock. De-sensitizing it, It could be making real knock but you will not know because of how much you have to lower the sensor level to not pick up false knock.

Those cylinder multipliers under knock sensors for each cylinder, does increasing or decreasing those values de-senstize the sensor?

[BoostedYards86]

reading. more info please.

[LSX378]

Now with this false knock, it doesn't seem possible to pick up any slight, REAL knock, as the valve train noise is louder than the combustion. With the false knock, its showing knock retard before it would even be able to make REAL knock. De-sensitizing it, It could be making real knock but you will not know because of how much you have to lower the sensor level to not pick up false knock.

Those cylinder multipliers under knock sensors for each cylinder, does increasing or decreasing those values de-senstize the sensor?

Increasing the values de-sensitizes the sensors. Have seen GM use values as high as 16 and they need a little work for forged pistons too.

[Blackdevil77]

Increasing the values de-sensitizes the sensors. Have seen GM use values as high as 16 and they need a little work for forged pistons too.

Does it really? Wow, been doing it backwards lol. That explains a lot!