350Z knock detection

[Timboj]

Hi guys, I'm having another attempt at dialing in my knock detection. It's a noisy engine (VQ35DE) with bigger cams and a supercharger with E85 so I'm sure it's all false positives. I appreciate that unless someone is using an audio device to identify true positives it's all guess work. I just want to dial it in as close as possible for now.

I did an easy drive this morning and got this mess:

Do I just have knock gains set too high/threshold too low?

Log - https://drive.google.com/file/d/10qQPKX5mt3_RVgfkpuq-98BbNvhNFpMi/view?usp=sharing

Map - https://drive.google.com/file/d/16Vo9TOmD2_3sr5Q4E55pZqvmSy_KEs6A/view?usp=sharing

[Adamw]

I would say your frequency is no good, when you are cruising at very low load , with steady RPM & load like the example below your knock levels should be barely above 1 and very flat, there shouldn't be spikes like that.  

I would also take your window start to 30deg after ign so you are not listening to the piston slap at tdc.  

  

[Timboj]

Thanks Adam. 6kHz is apparently the correct frequency for the VQ35. I switched it to 12kHz in attempt to use the 2nd harmonic, but that might be counter-productive in my case?

I'll go back to 6kHz and test.

[DerekAE86]

4 minutes ago, Adamw said:

I would also take your window start to 30deg after ign so you are not listening to the piston slap at tdc.  

Interesting. I would have thought knock would actually happen before the piston reaches TDC?
So if you're IGN timing is 25BTDC and you start listening at +30deg (5ATDC) you'd miss the knock event?

[Timboj]

@Adamw While we're at it, would you recommend switching to a WB sensor where possible?

Curious about Derek's question too. I'd have thought the knock event would be detected before TDC?

[Adamw]

Knock typically initiates well after TDC, 10-50ATDC.  On a well optimised engine the peak combustion press will be around 15ATDC, on poorly optimised knock limited road car engine the peak cyl press will be more like 25ATDC, when knock does occur the biggest shock wave is often the first one right around the normal PCP but it can sometimes start off smaller and grow during the expansion phase so this is why you want to monitor for sometime after this point also.

 

  

[Timboj]

Always learning something here. I think I've been mislead by animations in terms of when knock occurs, or more to the point, is detectable.

 

Re the WB sensor, does it make sense to do so for piece of mind? Or does the frequency not change enough, regardless of mods (other than boring the cylinders)?

[essb00]

@Adamw The window start then varies widely as it is based on the ignition event. Wouldn't it be more useful if the knock window start could be selected in absolute degrees (e.g. -10°) instead?

[Adamw]

11 hours ago, essb00 said:

The window start then varies widely as it is based on the ignition event.

Yes, that is the point.  The location of the knock is closely related to the ignition event because you dont get any combustion pressure until the spark event.  You can see in the 3rd pic where the ignition has been advanced 4degs the knock initialisation has also occurred about 4deg earlier.

[DerekAE86]

34 minutes ago, Adamw said:

The location of the knock is closely related to the ignition event

How soon after TDC is a good point to start listening to avoid piston slap?

So for example if you wanted to start it a minimum of 2deg ATDC, and your maximum timing value is 32BTDC - would you set the window start as +34deg after ignition to make sure you're never listening at +/- 2 TDC?

[Adamw]

4 hours ago, DerekAE86 said:

How soon after TDC is a good point to start listening to avoid piston slap?

I assume it is right at TDC where the acceleration swaps direction but I dont know for sure, I've never really considered pin offset and compression load etc.  I generally aim for my window to start around 10ATDC ballpark at peak torque and it seems to serve me well in most cases.