Syvecs Forums

I tweaked my knock gain and gain tuning to get all 6 cylinders knock values to 3.0 on idle. I had to use the lowest gain possible of 0.111 along with a gain tuning of
cylinder 1 - 0.91
cylinder 2 - 0.89
cylinder 3 - 0.90
cylinder 4 - 0.91
cylinder 5 - 0.90
cylinder 6 - 0.91

I just wanted to confirm, is this normal behaviour? having to lower the gain to the lowest possible value in order to get the knock noise to 3.0 on idle? engine is a 2JZ-GTE.

Are you using the OEM specified knock sensors or something else?

Using bosch wideband knock sensors, I tried the settings above but even on WOT at 7,000 rpm it peaks around 6.5 which is too low I assume? going to increase the gain higher and ignore the 3.0 rule at idle. Do let me know if anything I am doing doesn't sound right.

Gpro wrote: ↑Sun Jun 21, 2020 6:23 pm Using bosch wideband knock sensors, I tried the settings above but even on WOT at 7,000 rpm it peaks around 6.5 which is too low I assume? going to increase the gain higher and ignore the 3.0 rule at idle. Do let me know if anything I am doing doesn't sound right.

That is your problem.
Seen this same problem on many calibrations that have been sent to me to work on/fix, typical of items made by Bosch though

p.s. I always get asked when setting up new customers cars, "should I change out the knock sensors to new/latest Bosch wide band type" simple reply is NO!

Not sure I would agree? You suggesting keeping the stock knock sensors, designed in the early 1990's?
I will play around with the settings some more, induce some real knock for testing and hopefully get it working perfectly

Gpro wrote: ↑Mon Jun 22, 2020 12:27 am Not sure I would agree? You suggesting keeping the stock knock sensors, designed in the early 1990's?
I will play around with the settings some more, induce some real knock for testing and hopefully get it working perfectly

http://www.riceracing.com.au/Knock.htm
Yes that is what I suggest, they actually work unlike Bosch in some instances, (I have seen your problem many times over on lots of different engine types on this ECU) but what would I know , be careful and good luck

Best thing (without telling you everything I know) is get a lab scope, test the sensor on an actual engine, you will find the frequency there is only one (not many like other ECU makers guess) and with the right sensor it acts exactly as per the guidance from Syvecs (read right amplitude for an actual knock event). There is a very real chance your engine will disassemble itself before you get to induce 'real knock' with a sensor that is not working as it should. And if you are going ghetto and trying to listen for knock, then I feel for your poor engine as it will be potentially molested in short time as there is no way a human being can react fast enough.

Thanks for the tips Rice Racing, managed to induce some knock at 0.3 bar of boost and sensor picked it up nicely.

Gpro wrote: ↑Tue Jun 23, 2020 12:18 am Thanks for the tips Rice Racing, managed to induce some knock at 0.3 bar of boost and sensor picked it up nicely.

No problem, what did you end up doing?

Another little pro tip, most engines will show 'knock' when you start them, this is a really quick initial zero stress way to see if you have the sensor set up right physically and in software, which yours seems to be doing now

That I did not know and did not expect, thanks once again. Looks like knock sensor is all dialled in now. Time to shift attention back to perfecting the fueling map, especially around spool up areas.