tuning knock fast attack rate and recovery

Just trying to understand tuning these to try and help the factory knock system on my 95 lt1 work better. Its rebuilt, with a c503 cam, headers, roller lifters, etc..

i was getting some extensive knock and seemed to be maybe some false knock so im trying to tune the sensor to help with that from some stuff i have been reading, but i could be completely off track.


After some trial and error i figured out raising the recovery rate causes it to recover from the knock quicker, i cut the recover rate numbers in half. Now it will pull the timing, and start to add it back in nearly instantly, which may really be a little too fast, but its kind of nice so you can tell exactly where its happening.

I also tried adjusting the knock fast attack, first by lowering 30% and then by raising 30%, it seemed to take away the knock i was getting when i raised it by 30% so im thinking that was a step in the right direction.

Though it was also cooling off outside, so only more testing will tell.

Can anyone tell me if im going in the right direction? thanks as always guys

pull like 4 degrees of timing where it's knocking and see if it goes away, that'll tell you pretty quickly if it's real knock or false knock. if it's false knock, just make your timing table more conservative in that area, and disable knock retard for that range. you only need knock detection in regions where your timing is aggressive enough to actually cause knock.

raising knock fast attack by 30% wont remove knock, generally ends up causing more knock retard when knock events happen.

one thing that's harsh is the stock 'maximum timing retard' settings are like 9 degrees at WOT which is rediculous. my thoughts are, if pulling 5 degrees of timing doesn't make the knock go away, then it's not real knock. i always dial that back.

for example i had my exhaust hit my floor pan in a hard corner, which freaked out the knock detection and pulled max timing; 9 degrees in the stock configuration. it felt like i lost two cylinders when that happened. if i was limit cornering, i might have ended up in the ditch.

It takes more timing reduction to stop knock once it starts than it takes to prevent it in the first place. Where 2-3 degrees less advance in a table can prevent knock, it can take 6-9 degrees reduction to stop it after it's begun.

Increasing the ramp up after KS activity is a good way to increase the average advance delivered when burst knock is detected. But the better approach when time allows is to work out the reasons why knock is detected and, if possible, tune them out. I've made the analogy before that running against an engine's advance limit is like carrying an unstable explosive. All it takes is a little bump to set it off and recovering after the event takes more work than preventing it. While viewing a log file, remember that a log represents dynamic conditions and you need to try and capture what's happening over time. When you see real knock occurring it's often the result of a previous condition. If you have a nice spark curve with strong acceleration and suddenly knock develops, back up in the log and search the O2 readings for a slight turn toward lean which might have increased chamber temps slightly. Watch for a slight increase in the rate of spark advance or one cell that the ecm traversed that had an advance entry just a tad higher than the others. If you're doing full throttle runs or WOT maneuvers you might have to make changes a thousand RPM or more prior to knock beginning.

the knock seemed to be real and happening between 3000-3500 . I was able to get it all out by lowering my timing to 29. It's 27 in the tables but two comes from somewhere. Every once in a while if it gets hot enough out it woild still try to pull a degree or two( max at3) .

At this point I have it set so that it wont pull any timing .

I wonder about what you said about the other possible occurrences. My wideband never runs lean as far as I've seen but I still need to log it.

Also this may pojnt to something....when I was running on the dyno the last time I was able to get the most powerful with a 12.5 wot a/f instead of running 13-13.5
It made a consistent 10 hp difference.

Quote:

---

Originally Posted by zoomo

. . . Every once in a while if it gets hot enough out it woild still try to pull a degree or two( max at3) .

---

With $8d, it's possible to get knock retared degrees without any reported knocks, Other GM OBDI masks are similar. This happens because a reported knock (slow knock) is only reported after 255 "fast knocks" have been detected. And it's these fast knocks counted "under the covers so to speak" that are converted to knock retard degrees, not the "slow" ones that are reported as 1 knock for every 255 fast knocks.