G4+ fury on 2jz vvti, oem knock sensors, normal gain?

[Floppyz]

Halo everybody, got a rapid question: what about knock channel gain value?

I've set base gain at 1 and single cylnder gain at 0.8, but looks like oem sensors have very high gain because global knock value is quite loud and high evenif we checked for external machanical noise as well, angine is smooth and sounds good.

just only 0.6 bar boost and no pull adv as well...

[Floppyz]

Another big question about this car tuning:

we are useing a Storm G4+ for thia 2jz vvti, i've update firmware to latest 5.6.8.3671 but still can't see  'Auto Mode' in CLL... so can't tune al those features for a proper CLL behaivor... should be firmware the same for all g4+?

Please Adam help me!

[Adamw]

The toyota single wire resonant frequency sensors typically dont work with most aftermarket ECU's.  Their output voltage is very high and you will find even with low gain the signal saturates the input and it starts clipping the signal.  I have seen them work on some low power/stock engine but usually no.

Automode CLL is an advanced feature that was only available in the Xtreme and higher.

[Floppyz]

Tks Adam, so do you suggest to use motor sport Knock sensor like Bocsh one? this car runs with efr8374 and GSC S1 cams..

3 minutes ago, Adamw said:

Automode CLL is an advanced feature that was only available in the Xtreme and higher.

But i've used this in G4+ Evo9 p&p and G4+ Storm on another 2jz as well, look:

 

how is it possible?

 

[dx4picco]

P&P ecu's are based on the Xtrem platform so that is why it is available. storm only have stoich CLL

[Floppyz]

and what about Fury G4+?

[dx4picco]

its a Xtrem with integrated lambda controler so yes

[Adamw]

I said "Xtreme and higher".  That means Xtreme, Fury, Force & Thunder have the automode.

[0x33]

I generally run the OEM knock sensor for stock engine 2J's if its a mild setup small single at max. Only go to bosch 2 wideband knock sensor with big turbo build that is a bit more mechanical and heavy duty. 

These are my settings for most typical stock engine 2J's using factory knock sensor. I find it to be quite reliable, always detects knock when I can hear it etc.

You may need to adjust gain per sensor +/- 0.10

If you still have issues, then perhaps your knock sensors have been over torqued or you have interference on your wiring. 

[Floppyz]

On 4/24/2021 at 10:58 AM, 0x33 said:

You may need to adjust gain per sensor +/- 0.10

This is my last saturday test at 1bar boost (efr8374 and GSC S1 cams) with a simple knock target fix line @500.

Setup very close to your at least, but knock sensor input signal is near to saturation, so dangerous zone..

Anyway i've just ordered 2 Bosch new sensors

 

[0x33]

Must be something related to your setup. Check you haven't over torqued your knock sensors whilst installing. Or maybe engine isn't stock, you are brave to rev to 8k on standard internals lol. 

This is a log of my knock trace using settings above,  it is repeatable & similar on most stock 2J's I come across. 

T88 @ 1.6 Bar

[Floppyz]

Engine is rebuild yes...

What is oem sensor torque tighten value? I haven’t do that build but I can suggest owner to check this.

Wireing is made new from an US company , could it be a shield or ground issue? Also this high level of noise is repetitive even stationary high rev rpms...

[Floppyz]

I also notice that your windows setup is quite narrow than std setup: 0-30 start-length instead of 10-50.

What's your exeprience about that?

tks

bye

[BlakeR33]

3 hours ago, 0x33 said:

Must be something related to your setup. Check you haven't over torqued your knock sensors whilst installing. Or maybe engine isn't stock, you are brave to rev to 8k on standard internals lol. 

This is a log of my knock trace using settings above,  it is repeatable & similar on most stock 2J's I come across. 

T88 @ 1.6 Bar

I'm not a tuner but I was just wondering, I thought the whole point of tuning was to make sure you don't get knock? Is 265 counts of knock a high level? I'm only asking because on my Apexi PowerFC I use to get about 10 counts of knock if I pushed it really hard, but now with my Link I don't get any knock at all? 

 

Cheers. 

[0x33]

6 hours ago, Floppyz said:

Engine is rebuild yes...

What is oem sensor torque tighten value? I haven’t do that build but I can suggest owner to check this.

Wireing is made new from an US company , could it be a shield or ground issue? Also this high level of noise is repetitive even stationary high rev rpms...

44nm 

If you have a built engine then I would just use bosch knock sensor, I only use factory knock sensor for stock engine. 

I've seen before popular USA wiring companies using cheap wires that aren't shielded for 2JZ loom, so also possible you could be getting interference. You shouldnt be getting such high readings just revving on idle, unless there is mechanical problem with engine or weird wiring. 

I don't notice any major difference in window size between 0-30 to 10-50.

0-30 seems fine for me, always knock trace picks up any det as I listen on det cans, so I have learned to trust it.

[0x33]

4 hours ago, BlakeR33 said:

I'm not a tuner but I was just wondering, I thought the whole point of tuning was to make sure you don't get knock? Is 265 counts of knock a high level? I'm only asking because on my Apexi PowerFC I use to get about 10 counts of knock if I pushed it really hard, but now with my Link I don't get any knock at all? 

 

Cheers. 

Yes the goal is the same, no one wants a knocking engine. However, apexi knock and Link knock work differently so you cant compare numbers of a apexi versus what you get on a Link. 

I'm not getting 265 counts of knock. The 265 is a number to represent the maximum "noise" at a given point that I deem to be safe. Anything more than this number will be consdered detonation and ECU will pull timing. 

You should check out the help file in the Link software, its quite detailed. 

 

[Floppyz]

6 hours ago, 0x33 said:

44nm 

If you have a built engine then I would just use bosch knock sensor, I only use factory knock sensor for stock engine. 

I've seen before popular USA wiring companies using cheap wires that aren't shielded for 2JZ loom, so also possible you could be getting interference. You shouldnt be getting such high readings just revving on idle, unless there is mechanical problem with engine or weird wiring. 

I don't notice any major difference in window size between 0-30 to 10-50.

0-30 seems fine for me, always knock trace picks up any det as I listen on det cans, so I have learned to trust it.

the US company is Wiring spcialties. I'll check if wires are shield or not...

we've check several times for somthing machanical vibrations or possibile issue, but ther is nothing hitting... and engine is very smooth and silent to hear.

but let me undertend: you run 1.6bar with greddy T88 on stock internal??

[koracing]

I almost always use the OEM knock sensors on toyotas and it seems to vary from one to another.  Usually it's possible to get the sensor to read reliably with some value that's useful.  Some cars with rebuilt motors I do have to go to double the frequency (2nd order resonance) - like 13 or 14kHz and that often drops the nominal amplitude quite a bit, but it takes some qualification on dyno typically as well as later actually introducting a lot of timing to get some knock to ensure you have your thresholds properly set.  I wouldn't recommend introducing knock if you're on a motor that isn't built, and certainly not high boost or full rev range.

[Floppyz]

On your experience is it possible ti 'make knock' 2jz (rebuld engine 8.5:1 cr) for test knock system revving around  3/4000 rpms at least 1bar boost?

2jz looks like to eat timeing adv in middle charge and revs, isn't it?

[koracing]

Yeah it should be doable.  On dyno you would want to add timing and monitor power change up or down.  Something like 5 or more degrees past where you think it's safe and then observe that the knock goes well past the nominal values you recorded earlier.  I would keep the boost very low while you do this, but do before and after runs at "safe" timing values as well to ensure no placebo affects.

[0x33]

6 hours ago, Floppyz said:

On your experience is it possible ti 'make knock' 2jz (rebuld engine 8.5:1 cr) for test knock system revving around  3/4000 rpms at least 1bar boost?

2jz looks like to eat timeing adv in middle charge and revs, isn't it?

To get a nice ping, I usually do like 5-10psi at RPM where turbo comes onto boost.

If it's low compression engine you may need 30-40 degrees timing. I usually steady state the engine in that area until I get a good ping. From there I can establish if frequency and gain is correct for noise relative to the knock ping. 

Highly recomend that you use some form of det cans or alternative knock detection that you can listen to engine whilst you are testing this. Sometimes you can hear things that you cannot see easily on knock trace until knock system is fully setup and you can trust it. 

Also for your previous comment, use T88 1.6bar on stock engine. Very laggy and not much RPM headroom cause of stock engine, so bit of a meh setup. Not much stress, peak power is like 600whp and no torque relative to a newer turbo. 

[0x33]

Just to update this thread. 

Have a supra I'm doing currently on G4X SupraLink ECU. Stock engine, stock knock sensors with small single turbo. The noise I'm getting from knock sensor is much like your example Floppyz. 

This car is running wiring specialities loom, and the knock wiring looks unshielded. This loom seems to be reacting different to stock loom which is shielded. Shall see how the mapping session goes and will update if I notice anything else weird with this loom. 

[Adamw]

It is not wiring related, they are just highly sensitive knock sensors.  I have seen this on almost every 2J log that gets sent to me, the vast majority would be stock looms and I have seen the same effect with other ecu brands I have supported.  They output very high voltage when excited near their resonant frequency.  The factory ecu either has a low impedance input or a higher voltage measurement capability to work with these sensors but most aftermarket ecu’s wont.  If you have knock level global values getting near 1000 like floppy’s then that means the input side of the amplifier is maxed out and adjusting gains or frequency etc won’t fix it. 

[Floppyz]

Ok, that's clear now, as Adma says if input level is near to saturations there are no way to fix this limit, naither working on 2nd harmonic because distotions due saturation input could be an addiational input.

I'm tunenig a second Silvia with 2jz and this time I've make the wireing and I know that trigger sensors and knock sensors wires are shield, and also in this case oem knock level input is high...

Anyway ,as Koracing suggest,  could be intersting work on 2nd resonance, but could be a reliable risk due saturation distortions... aftermarket knock senosr are cost affordable so can we say that this way is the best solution?

Does enyone tried to put a resistor (470/1K ohm) in series for lower sensor gain ? Like a partitor or a balun?