NefMoto

ok while logging whats acceptable amounts of timing pull as my car seems to be loving more and more of the stuff. im getting -4.5 briefly for about a few 100rpm when the boost comes in strong then the odd spike at 2ish higher up on 99 ron fuel.

That's fine.

So could I go for more? I should probably look at getting a graph plotted to show you all really.

Logs would be good. Generally speaking, I shoot for CF's of 3 if running a consistent quality fuel. Anything over 5 I pull back timing. Sounds like you're in a good place.

Don't tune by CF's alone.  You are generally looking for the most advanced timing with minimal oscillation.

Rick

I shoot for as little correction as possible, ideally none.

Does everyone commenting in this thread run completely unmodified knock based corrections sensitivity maps?

I can't comment for everyone else, but I haven't changed knock hysteresis.

neither have i

I don't see the point in changing any knock tables

IMO, the goal is to find the knock border. I shoot for no KR under most conditions with a blip of activity occasionally. With 3-5* retard I will start at the onset of knock and pull 1.5* in that cell only and blend it out if necessary. This will usually bring me down to zero KR or at least move it up a good bit. Repeat if needed.

I don't touch anything KR related except disabling adaptation during tuning.

Edit: I should clarify, this only applies when street tuning a knock limited application

i speed up recovery a little bit

With this approach, since you're focusing on pretty much one cell at a time, do you do similar pulls in all of the gears?  How about something like 3/4 WOT pulls to get the other cells?  With literally zero corrections, do you worry that you're leaving some power on the table?

I don't tune FATS pulls. I try to get all driving conditions covered in my logs. Therefore, it really won't be one cell at a time, more like the whole map. When knocking occurs, I focus on the one cell that triggered the KR. I shouldn't really say one cell because the adjoining cells have to be considered as well and any blending out needs to be done also.

My overall timing is more advanced with zero KR than with more advanced desired timing and KR. There will be blips of KR occasionally because I'm riding the border, but never continuous knock.

I think you guys assume too much that all cylinders are equal.
They are not.
And if you don't have any timing pull, you are losing power, as some cylinders can easily have 3-4 deg more advanced timing.
Tested this on dyno as well, and everything is as expected.

I'll start by saying that I get what you're saying and I'm not arguing or disagreeing, but...

Typical results I see:
zwgru : 21*
dwkrz : 4.5*
zwout : 16.5*

zwgru : 19.5*
dwkrz : 0*
zwout : 19.5*

This is how I've found the best results.

So, my question to you is, are you doing something else? Modifying something in KR?

I don't get people's obsession with zwout, when zwout is the current output for a cylinder.
What you see in zwout just depends on blind luck, it's not really useful for anything. Yes, one or two cylinders will be retarded more, but the rest will be advanced, making more power in total. You can also adjust KR...

The fine balance is where about half of cylinders have some sort of correction. You don't want correction on all cylinders.
My experience is with dyno testing and what makes the most power.

Make some back to back runs and you will see what I mean. All these engines make peak torque and power with a little correction.
Your quoted numbers and observations would be correct for a single cylinder engine, but most engines have at least four cylinders.

Understood, I used zwout for simplicity. Same reason generalized dwkrz rather than being cylinder specific. I do shoot for some correction here and there, but overall near zero.

Regarding the Op's question, street tuning, how many degrees retard for dwkrz are you satisfied with?

Do you usually adjust KR, or do you usually leave it stock (on the 2.7t)?  About how many degrees correction (on half the cylinders like you mention) do you feel is safe but powerful ?

I aim for about 0.75-1.5 WKRM, as then the most knock prone cylinders get some correction, and the rest of the cylinders run more advanced ignition.
I've also found that this gave me the best result on the dyno.

It's pointless to have correction on all cylinders, that will just make less power. With knock control I sometimes adjust how much it retards on the first event.
That said, I have not done anything on my RS4 with KR, as the stock calibration seems quite good.

Thanks for the insight and dyno experience.

While I am green to most of this. Isn't it all just a wash?

If I have my tune currently pulling say, -6.5(average) on all cylinders but don't feel like that is "safe" and dial back 4-5 degrees to achieve -1.5(average) on all cylinders it's still achieving the same actual timing, right? The -6.5 would be causing ignition at say 10(hypothetical) degrees BTDC and the "dialed" back -1.5 would also be 10 degrees BTDC ignition.

When timing is pulled there is a delay before it's added back. Running optimal values is better all around.

Did you read this thread?

For one, it's better to prevent knock than to react to it.

KR will pull more timing to subside the knock than was necessary to avoid it.

As DD said, it takes time to recover from retard.

Also, the retard is adaptive, so dwkrz will be loaded with adapted values if learned in a particular load/speed range even in the absence of knock.

Also, in the case of something going wrong, you preserve the headroom in max retard.

Has anyone compared the KR hysteresis etc maps in the RS4 files to the S4?  I'm curious what differences there are, if any.  The knock sensor filtering is likely a little bit different, with the different pistons/compression ratio etc.

Are you talking about KRANH and KRALH?

These are for defining adaptation ranges.

You might find though dialling back 1 degree will get the average down to -1.5 .. Just because you gave a cf of 6 doesn't mean you need to pull 4 degrees to get back to 2..

I think there may still be some bad information about those maps in the s4wiki.

I have not gone back to correct them yet.