The ecu's knock detection is quantitative, there is no qualitative measurment. It sucks. Majority of my knock is in closed loop, with stock engine parts.

I think the best way to tune for knock without a gas bench would consist of having the sensor hooked up to the amp with headphones somehow.

Also, never use "knock volts" as reliable data, it's not a linear output. Several years ago even competent tuners mistakenly tuned by this.

 

Until that headphone thing happens, using knock sum to tune is still the best bet is it not? It's not always 100% accurate but it works.

 

What are the original knock values for an 03 VIII?

 

^^^^^Every knock map or just the first one?

 

Those are not knock values....

 

Those are values that some of the clever folks here will eventually crack. Thinking the 12 filter maps are separated into 4 - 5 different catagories (relating somehow to CA, freq, noise level not during combustion, etc)

 

Many have suspected that these "knock filters" are nothing to do with knock control. Mattjin has found that some are related to fan control. Having now disassembled the knock control routines, I can confirm that these so called "knock filters" are nothing of the sort. They should be relabelled.

I suggest de-stickying this thread.

 

+1

 

Sorry for bring up a old thread, but would bad lifter tic cause the sensor to read as knock and pull timing?

 

It is possible, I believe I have heard this theorized before. In any case, you would probably see that picked up constantly when logging knock_base

 

Yes yes yes yes yes yes yes

 

+1

 

So lifter tic is making the ecu read as knock so its pulling timing causing real knock. how do i fix this short of getting new lifters?

 

I believe they are accidental fire detection maps for screwy rotation rates or accelerations, + and-.
And I wouldnt mess with them either.