Forum » General Tuning Discussion » Newbie trying to understand knock control and monitoring

Newbie trying to understand knock control and monitoring

General Tuning Discussion

Discuss all things tuning in this section. News, products, problems and results. 

= Resolved threads

Page 1

Hi All, recently had my supercharged 2zz-ge tuned with an ECUMaster ECU and though I'm happy with the map and how the car drives, I've been using it as an opportunity to self learn a bit and explore the software.

One area I worry was neglected during the initial tune was knock control. The 'engine noise' threshold was set at a flat 2.8V throughout the range and the recorded engine noise was spiking all over the place, triggering "knock events" like crazy.... this didn't actually impact the running of the car as the "knock actions" were disabled... which I felt was quite worrying. It seems the tune was fundamentally safe as it's survived some time on track and a fair few road miles already... but I'm still concerned.

I've been told from various sources that a supercharged 2zz is inherently noisy on the same frequency range that you'd expect knock on, so identifying it can be quite difficult but I find it hard to believe that disabling knock control completely is the right answer.

After a bit of time with HPA resources and general googling I decided to work on establishing an engine noise baseline and also tweaking the gain/integrator settings to try and dial out the spikeyness of the noise readings. I also did a few road pulls with a load of timing taken out to try and give myself confidence that the engine wasn't actually knocking, and despite pulling out 10 degrees or so the sound profiles/patterns remained similar - so fingers crossed it's not actually knocking.

The specifics that I did are as follows:

- Set the frequency to 7.27khz. It was originally set to 6.94khz and "the equation" brings me out at 7khz but I found that 7.27 has a slightly smoother reading.

- Gain set to minimum (it was already there)

- Integrator increased to 200uSec. I don't fully understand this setting, but moving it from 110uSec brought the peak sound readings into the 2.5-3V range which ECUmaster manual says is ideal

Following this, I used some logs from a track session to set a fairly relaxed engine noise baseline across the range and by the final track session of the day I triggered zero knock events. Finally I enabled knock control action with some defaults I took from a base map I found:

- 50% min TPS

- 2%/V fuel enrichment (capped at 5%)

- 2deg/V retard rate (capped at 5 deg)

- Restored after 120 revolutions

I'm fairly happy with the theory of all this, but I'm looking for some validation on my approach more than anything. I'm fortunate enough to have never experienced engine knock before so in practical terms I don't know what I'm looking for and I don't know if I'm potentially hiding knock through my settings.

Even if I went out with some det cans on, I really wouldn't have a clue what I'm listening out for. I know the proper answer is to go see a proper tuner, and that's being handled - but I'm still keen to learn and figure out this stuff for myself too in parallel.

I have some specific questions, hopefully somebody can help:

- When configuring an engine noise 'table', how much headroom should I be leaving? I assume it needs to be as close to the real noise as possible but as I was seeing slight variation between runs, I ended up over-gunning a bit just to stop the false positives.

- Do my noise readings look anywhere near normal? They rise/fall with supercharger RPM it would seem which makes sense - as it's very noisy! I get the odd spike on low/no TPS which you can see on the attachment - this seems to be the high lift cam disengaging as it's timed perfectly with that each time.

- Would you expect knock to manifest itself as a very obvious spike on top of the noise patterns that I'm seeing? ie, should it be really really obvious?!

- How do my knock action/control settings look? I have nothing to reference them against - I'd rather they were overly aggressive for now until I get more confidence.

- Have I missed/overlooked anything with my approach?

Thanks for reading if you made it this far, I appreciate any input.

Attached Files

hello Jon,

There's a bit to cover regarding knock but typically a tuner will tune to MBT, in cases where knock occurs prior to MBT they may pull some timing to save the motor and expect to see an ignition map -2 from the knock limit for safety.

With regards knock control set up; I'm not familiar with the ECU master functionality but output should be a a counter or index value based on the voltage, gain and other inputs not the raw voltage itself. I've attached a poor screenshot as an example of knock and it can clearly be seen, so on a calibrated set up it is obvious.

As noise is directly proportional to rpm we would expect the sensor voltage to increase with engine (supercharger) rpm so not much can be learned there. The same goes for reducing Using audible knock cans is incredibly valuable as this will validate your knock sensor calibration. Phormula do a simple system that's quite affordable and have examples of what knock sounds like on their website but from experience you will know when you hear it as it a completely different frequency/tone/volume to the background engine noise.

The process of validation might make you wince a but you will need to drive the engine into knock by advancing the ignition until the 'sound' is heard and re calibration the integrated knock controller until it is picked up. Your tuner should have already completed this process when calibrating the ignition map to MBT and temp/pressure compensation tables so its a normal part of engine calibration; however beware in activating closed loop knock control, just because the functionality is there it might not be robust and the tuner may have had a very good reason not to activate it.

Attached Files

Thanks Scotty. I've consumed as much HPA content as I can find on Knock and it's been really useful.

I don't think anybody can really answer my direct questions without being in front of the car on a dyno or whatever, but any validation I can get about my technique and approach in the meantime is about as much as I can hope for.

I have to trust the tuners and their original setup, the car performs brilliantly and if I'd not have got a USB cable out I'd be none the wiser about how the knock detection was setup (or not, in this case). I'd like to think there was a proper reason for not configuring it, but I can't get a straight answer out of them.

A recurring theme I am seeing though on the few 2zz configs I've seen from other cars is that nobody seems to be even attempting an accurate method for detecting/controlling knock. Different tuners too.