"flat response" knock sensors

[Aedakin]

i know its been a while, but i am curious if anything came of this. I am planning to do the same thing. Planning to use the Flat Response sensors from my 06 4.3l truck, then use the information from the 06's PCM to try to dial things in.

I figure if i have a known good working setup, I have half a chance of making it work in my 02. I hadn't seen the pin locations for the ground side, only setting a ground at the sensor. I had run across the frequency in my 02's PCM.

I've replace the gen 3 sensors twice, then learned that most times factory sensors are needed. Those are a bit pricey and I tired of the run around. if i can fix it with Gen 4's, i'll be happy and plan to put the info out there out there for others.

[roughneck427]

i went thought and did an overlay of a 4.3 and a 5.3. I parsed out the knock sensor differences in Winols. alot of these definitely not in commercial tuning software. If needed for another OS i can quickly line it out

[cwalsh]

Looking all this over and multiple other post across the web as well but I always seem to be coming back here.

1 Has anyone gotten this work correctly?

2 I see the 2002 S-10 .bin get thrown around a lot but from research that shows to be a resonant KS. Why wouldn't an 06 Silverado 4.3/4L60E 12606807 OS that uses the flat response sensor be used? I get with out a dyno and proper 3rd party knock detection this is all guess work.

[NSFW]

I have a dyno tuning session in a couple hours. Going to rely primarily on a Plex Knock Monitor, but I also wired the sensors to the PCM and took my best guess at tuning for them. Fingers crossed...

Haven't seen any knock retard while driving, but the tune that's on the car now is very conservative and I've been driving it like a commuter, so they were reporting knock that'd mean I did something wrong.

[NSFW]

The knock sensors work... Probably too sensitive though.

I'll post my .bin when I get back home.

[NSFW]

A couple things to note...

I disabled per-cylinder knock detection just to play it safe. To use cylinder mode, you need to configure which sensor to use for which cylinder, and I didn't want to risk getting that wrong during the dyno session. It is the next logical step though, now that we know it basically works.

There is a table that tells the PCM what frequency to focus on when listening for knock: B6226, "Knock Sensor Frequency." There are online calculators that will give you the knock frequency for a given bore size. Multiply the frequency by 2. You don't put the actual frequency into the table - the notes in the table description tell you what number to put into the table to indicate a particular frequency.

There are two tables that tells the PCM how much background noise to expect, depending on RPM. That part was too sensitive at first, especially at higher RPM.

First, see table B6255 "Knock Sensor Multiplier RPM" in the big 7603 XDF. Think of it as representing background noise, so higher values = lower sensitivity. I think we got it about right for my car, but that may or may not be right for other cars.

And as I was writing this I realized that table B6246 "Knock sensor multiplier A (engine mode)" should probably have the same sort of shape to it. Based on the definitions in the XDF it seems that B6255 is used when RPM is increasing or decreasing (by an amount configured by B6258 and B6259), and B6246 is used when RPM is steady. So I took a random guess at what might be right... it'll take some testing to find out whether that guess is even close.

And maybe those two tables need to be tuned for every engine? I don't really know how much they're going to vary from one engine to the next. I will look into whether there's a parameter we can log to find out how much noise the PCM is seeing from the sensors. For the Plex Knock Monitor v3, you just free-rev to 7k and it shows you how much noise it's hearing, and anything above the noise threshold is considered knock. If we can log the sensor signal we can use the same approach with the PCM. But that's a big '"if" because I don't know whether there's a way to log it.

But anyway... This file is a decent place to start if anyone wants to try using flat-response knock sensors with a gen-3 PCM.

Wiring:

Bank 1 ground: J1, pin 49
Bank 1 signal: J1, pin 09

Bank 2 ground: J1, pin 50
Bank 2 signal: J1, pin 10

I'm not sure where the terms "ground" and "signal" came from, but I'm pretty sure polarity does not actually matter on these sensors. They're basically just passive microphones.

One of the earlier posts in this thread has a link to the .bin that I used for reference when I was setting up my .bin file for flat-response knock sensors.

[antus]

Good work thanks for shareing

[cwalsh]

Thank You NSFW for the effort on this and keeping an old thread alive!!!!!

What kind of car are you building?

Chris

[NSFW]

It's a 2002 Corvette, the LS6 was a total loss after the timing chain broke at a track day in 2020 and it's been a long journey to get it back in action. Spent a year making up my mind about what direction to go with the engine, waited over a year for parts because of the pandemic supply chain stuff, waited months for installation, spent months looking for the source of a misfire problem, blah blah blah.

Still need paint-protection film on the front bumper, add brake cooling air deflectors to the front suspension (had brake ducts before), and properly mount the PCM and TAC, which are on the floor in the passenger foot well. Then track days again.

It made 520whp on the dyno, and the tuner was stoked about it, but I was hoping for closer to 600... It made 690 on an engine dyno before Golen shipped it out, and I assumed that was on pump gas. But if that was race gas then it all makes sense. Going to try going on step colder on the plugs to see if it'll take more timing. We got about 25whp extra by going one step colder from what it came with, so there's hope. Not a high priority though - mostly I just want to drive it.

[cwalsh]

Sounds like a fun project!!!