NefMoto

Hi All,

It appears to me that the consensus among the experienced tuners on here is that approx -1.5deg KR is optimal in terms of making power, and that no timing pull could (potentially) be leaving some power on the table.  This has been discussed at length on here

One thing I don't think I've seen, and perhaps this is obvious: are we discussing the average knock retard across all the cylinders, or the maximum knock retard for an individual cylinder?
 
Similarly, what should I consider to be a "take your foot out of it right now" amount of knock retard? I think the worst I've seen is -6 to -8deg when I got an axis wrong or something.  


Assuming the car is running pump gas

Per cylinder

Golden rule for me - up to 6 deg

Zeros - safe but lost a power
3 good at all
6 fine
8 should drop zw a bit.
Its still safe and everything less than 12 are safe. But what if kr sensor faulty or  fuel quality or hot iat etc. Thats the reasons why you want extra space for kr
Its like tune fuel trims. O2 reg gives you pretty wide range of error but you want to keep trims near 0 just in case

Running any KR makes less power than running the correct timing without KR.
And if you have just one cylinder having more knock than others, there are maps to compensate on that single cylinder.

Having it run knock correction all the time means it is actually knocking all the time.
This pits pistons and destroys rings over time, especially with high boost levels.

For lower octane it is a much better idea to make use of the constant knock maps to actually mapswitch when it's detected.

On a stock motor I would not run ANY knock on any cylinder. How you end up with bent rods and destroyed pistons.

On built motor, max I would chance is -3 per cylinder not wkrm(avg) but really it should be trimmed out or modified so that there isnt any, in an event(high IAT w/e and not enough compensation) there is any knock it will only pull minimal amount not go massively into the - 's.

But what he^ said basically.

tune to *near* 0 KR under worst case conditions: heat, bad gas, etc.

Then, when it cools off and/or you get good gas, you're gtg.

To be safe, tune to 0 in the worst conditions as well.

ymmv.

Thank you everyone for your replies!


This is implemented on my file stock and I will look into making better use of it.

A couple follow up questions from reviewing my logs on the way to work:

1. If I stomp on the accelerator at say 2000rpm, by 2750rpm the engine is just peaking at 20 psi, 0.82 lambda, and I get a knock event(s) that pulls 3 deg from one cylinder.  zwist and zwout are 3.75 deg when the knock event happens.  The IAT is about 15 C and the coolant temp is about 80 C.  See attached images fig. 1 and fig. 2.

The way I see it, I have three choices: reduce boost, reduce the ignition timing, or add fuel.  Which parameter would you change and why? How much would you change it by?


2. My car has always had a large amount of knock activity between idle and the turbo spooling-- like -6 to -8 on multiple cylinders is not abnormal, though it's typically less than that.  I always assumed this was normal as the car has done this as long as I can remember (at least since I 1.8T swapped the car in like 2014). At that time, the only modification to anything was a single mass flywheel from (I believe) a Corrado G60-- I installed this at the same time I did the engine swap.  The condition of the engine mounts doesn't seem to change this behavior.  Is this normal?

You really are trying to get people to tune your car cell by cell through text?
No. This means your ignition map has too much timing in the the region where it is knocking for your fuel/CR combination.

It can be a hardware problem there , on that cylinder 9bad compression, valves , cams , or injector faulty).
If not, like PRJ said , timing is too high there or evenif your aft looks good, one of injector on that cylinder can make problems.

Zero lost power ?
Not really
dyno test different 1-6 retard and 0
zero retard more power
ecu any knock / ignition intervention less power

normaly 0- 0,5-1  random cylinder not all ..best
after 5-6 retard correction less boost and other ..

It's very simple why there is less power with knock.
Knock heats the cylinder and then for some time (few seconds at least) you have even worse knock threshold, so you must run retarded timing.

Even being off by ~1.5 degrees is better than having CR on more than one cylinder generally.

Yes, i've moved in this direction as well. 0 CF for normal operating conditions, near zero for absolute WORST case situations (temp, altitude, gas).

It's really easy to do this. Why try anything else?

I had a feeling you would say something like this.  I'm here to learn from you, PRJ.   I just pulled 1.5deg out in that cell and it stopped knocking there.  Maybe you would have said add fuel or reduce boost and I would have learned something.


I'm actually equally surprised and concerned to hear this.  Even when the car was bone stock it would knock like crazy in stop-and-go traffic. 


I'm working towards this.  It's a good thing I asked. 

That's actually not a good sign; it really shouldn't if you are using good gas

Question maybe a bit off topic, but does head port size affect knock(on 1,8T)? I heard what big port AEB/AGU heads has more problem with knock, comparing with small port head (AWT, APT, BAM....).Perhaps small ported head can run more timing? Maybe someone knows and can confirm or deny this myth?

My experience is exactly the opposite.

Interesting info, but sounds logical: bigger port, more flow, lower egt. Is the difference significant? I always used large port heads, so cant compare.So maybe larger cams also could help with more timing?

I implemented the PRJ/Elray/RS4 feed forward PID badly
The root cause of this problem was LDRPID was out to lunch.
I hope this helps someone. 


Best of luck,
Blair