k24a2 RBB in base rsx, stock kal - knocks

[djandy]

Hi, probably this is a noob question, but several days of googling, reading old posts and youtube tutorials did not give the answer.

So, base 2002 RSX, swapped 2005 TSX K24A2, used TSX RBB intake manifold only, everything else is from base RSX (exhaust, throttle body with AEM short ram intake).

Started with tsx-04-05 stock kal from KManager, made only basic changes (speed sensor, idle, etc). Datalogging shown lots of knocks, so I tried check what's going on.

I disabled fuel s.trim, started datalogging and went for a ride trying to fill lambda table.
Behold the results:

my K24A2 - 2 - no strim.kal

my K24A2 - 2 - no strim.kdl

my k24 - 2 - no strim actual lambda.png

my k24 - 2 - no strim - lambda delta.png

There are 315 knocks!
I marked the cells which correspond to knocks.

What confuses me is that the results seem far from lean, it is 10-20% more rich than controlled afr tables. And it still knocks.
But what is interesting - knocks are always at the cells which are more lean than neighbors. So, I assume that the problem can be solved with adding more fuel to these cells and around to leverage them with neighbors.

[djandy]

Applied the corrections, uploaded to kpro and run the same route. And..

my K24A2 - 2 - no strim - added fuel against knocks.kal

my K24A2 - 2 - no strim - added fuel against knocks.kdl

my k24 - 2 - no strim - added fuel against knocks - actual lambda.png

my k24 - 2 - no strim - added fuel against knocks - lambda delta.png

158 knocks!

The tables look more uniform now, but there are still places with knocks, and some cells unexpectedly became more lean than they were (???) so I'm making another round of corrections and will check the results.

And till that I have questions.

* Why is it knocking on rich AFR? The ignition and cam angles are stock, the motor is also completely stock. The gas is as much 91 as it could be.

* What are my options here, keep up adding fuel or it is time to retard the ignition in knocking cells?

* (minor) The rightmost WOT column is empty because I cannot get there - is that because of narrow base rsx TB + probably clogged AEM filter?

Thanks in advance!

[djandy]

After 11 iterations I ended up with the following cal:

my K24A2 - tuned - no knocks.kal

my K24A2 - 2 - tuned final.kdl

Almost no knocks

If anybody can check if it is a valid tune or give an advice on improvements - you are very welcome.


Used a part of a datalog above for a virtual dyno, and got a curious result:

k24 3gear - tuned 2 - virtual dyno.png

It looks like my vtec is only for funny sound, the power is built by low cam )
How could it be possible? )

[djandy]

Anyone? Please, still need a hint what am I doing wrong.
Currently I have mpg about 19, max speed 122mph and it seems that it could perform better.

And in the name of science - why my safest AFR is about 11.5 while target tables are at least 12.5?

[CalibratEd]

Anyone? Please, still need a hint what am I doing wrong.
Currently I have mpg about 19, max speed 122mph and it seems that it could perform better.

And in the name of science - why my safest AFR is about 11.5 while target tables are at least 12.5?

If you want 12.5 at WOT and you are getting 11.5 you need to reduce the fuel in the area.

- Ed

[djandy]

Thanks for the reply Ed, the thing is that it knocks as hell if AFR is near to 12.5, that was the reason I started adding fuel.
I assumed that ignition and cam angle tables are correct, and only fuel maps should be altered, so I ended up with adding fuel in every cell to prevent knocking.

However I've already got a tuned calibration for a very similar setup, and actually it has all ignition and cam angle maps different from stock tsx 04-05 ones. It runs a bit better than mine and it really has AFR ~12.5 in WOT almost without knocking (checked with virtual dino).
The owner of the calibration said that stock tsx 04-05 is terrible.

So it seems that my question is answered - the ignition tables had to be corrected too. I've also tried to set high cam angles to fixed 0 and 30 just to check how it affects the performance - it did not result in more knocking, only power curve changed a bit, so I kept everything as is.

If anybody interested, here is the kal I'm using now:

ice20-final.kal

[CalibratEd]

Yes, the first step when dealing with knock is to reduce the timing. I am glad you were able to figure it out.

- Ed