Knock detected when going off throttle

[gips]

Hello,
i was wondering ,if anyone else is experiencing this phenomenon. Sometimes, mostly when the intake is still cold, but engine is already at operating temperature i am experiencing a large spike under the knock voltage, when going off throttle. This happens only, if the car was under boost before letting off. This rarely, almost never happen ,when the intake is already hot, like the car was drivenfor around 20+ min. I have not tested, while the car is cold for obvious reasons but i suspect it will be the same, if nto worse. So my question is, could this be a residual gasoline left on the chamber walls that autoignites after i let go of the throttle, obviously burning fast and creating a small ping. I can say its not a bad ping, because i almost can't hear it on the headphones. So my problem is not the ping itself, but due to the knock settings that i have put, the ECU retards the timing, so next pull is sometimes, a bit jerky Any ideas are appreciated.

P.S. Forgot to mention. Car is toyota SW20 with 3SGTE gen2, Wiseco pistons a bit of slap from them, when cold, around 9.3:1 compression, boost is ~1.2bar, using the stock narrow band sensor. Settings for sensor are 6.64khz, 1.06 Gain, 100 Integrator. AFR under boost is rich ~11-11.5:1

[pric]

I'm experiencing the same thing.
When I let off the throttle, i get a huge Spike, but I'm pretty sure the engine is not knocking.
The higher the MAP and RPM was before letting off trottle, the bigger the Spike is afterwards.
It's not temperature depended on my car.

Your theory with the remaining fuel sounds right.


I'd wish there was a function to disable knock controll when under pressure fuel cut is active, or under a certain MAP.
I temporarily "solved" the problem by reducing the restore rate, which isn't optimal.

Specs:
stock Wideband sensor(i think); Mercedes M104.941 (2,8l I6 24V, bore: 89,9mm); CR 10:1; completely stock motor; ~0,8-0,9bar; 100 octane fuel(EU); lambda under boost: 0,9-0,78. Gain: 0,267; Integrator: 200usec; 6,37kHz

[RHD]

this could just be the gearbox/ drive train / driveshaft etc, any slack on the doggs, gears or components going bang as they go from fully loaded into instant overrun, sending a shock noise back into the block... just a thought?

[pric]

Possible!

But for the problem, it doesn't matter where the noise comes from, knock is detected, when there is none...

Greetings

[RHD]

yes i agree that a time delay setting would be handy. Twin plate clutches cause a lot of havock with knock sensors too!

[gips]

Some update here, i seem to have been able to reduce the impact of this phenomenon by adjusting the knock sensor settings. I have induced knock under boost, to see what the voltage look like compared to the spike , when letting off. It seems the real knock, is a lot more louder under boost than the letoff spike. I have adjusted the knock settings, so that the letoff spike goes to around 3v and not more. At the same time even light knock under boost is at least 3.5v noisy, so i have put 3.2v as threshold. This has cured my problem. The settings that i currently use are.

frq: 6.64khz
gain: 0.55
Integrator: 100

Oh and something more I don't recommend to test it this way on stock pistons. Even a few knocks can destroy ringlands on them(personal experience), so testing by inducing knock on stock pistons is not recommended......

[Jadzwin]

If you want to test knock on engine you can generate it under low load. On heavy load it could destroy the engine very fast

[gips]

Yes i agree, however knock under low load and heavy load sound different, so i had no other choice, but to induce some knock under low boost like 0,5 bar to see how it spikes on the sensor. Its spike was clearly louder than the offthrottle one, so it was possible to make adjustments. And i have destroyed a few engines, so i know how much knock an engine can take. As i said stock pistons are fragile they can break a ringland on one pull, but after market can take a beating. After all if you want to get the most of an engine, you need to see it knocking under boost, to know its threshold . Also you know how stock ECU works, its constantly searching for the knock limit, because of the low/high octane tables, so its knocking from time to time, but it does not get destroyed. I mean knock does not destroy engine in a second, like some popular belief. It can probably, if you put 40 degrees of advance , but we all have heads on our shoulders