NefMoto

Why is my rlsol not following rlmax?

(http://gm.mainframe.no/pics/rs4/rlsol.png)

(http://gm.mainframe.no/pics/rs4/kfmirl.png)

(http://gm.mainframe.no/pics/rs4/kfmiop.png)

Also, does anyone know milsol_w variable address for 551K?

Logged it some more and my mifa_w is at 82%.
Something is limiting requested torque.

I will add mrfa_w to the log and see what's going on there.

I think my problem is KFMIOP.
I need to rescale the axis.

According to this, with SY_TURBO=1 and FMIVL=1, the value read from KFMIOP based on rlmax_w and nmot_w becomes mimax_w.
(http://gm.mainframe.no/pics/rs4/frkfmiop.png)

If I now look at this:
(http://gm.mainframe.no/pics/rs4/mifaw.png)

I see that mrfa_w is multiplied by mimax_w.
So it seems that the only correct way to tune KFMIOP is to rescale the axis.
The purpose of KFMIOP seems not only to tell the ECU where requested torque is reached, but it is also the hard limit for driver requested torque...

Will try it and see :)
Of course the KFMIOP axis is also used by KFZWOP and KFZWOP2, but I guess nothing will blow up too much if I rescale the last 3 columns or so.

Yup, this fixed it.

Thank you for your sharing.
Good Job !

Please post the updated maps with logs.

I am still tuning timing and boost linearization.

But this is what I have now...
The log was done with a slightly lower KFMIRL (230, now 235) and I rescaled KFMIOP from 225 in last column to 232 since then.
Anyway, I am getting pretty close.

This is what I used to calculate my KFMIOP, it matches the existing (stock) values in the table quite well...

No comments? :P

Lots of comments.  Can you please describe your engine setup?  Can you also elaborate on what you did to convert your ECU to be RS4 compatible?
I know you did a chip conversion of some kind, but I only saw the desoldering thing.

My car is mostly stock. The precats are gutted and a stronger clutch has been fitted, that's about it.

(http://gm.mainframe.no/pics/rs4/rs4.jpg)

I guess the picture answers your question though :)

Another problem - my specified load is capped at about 218-219 from 3000 to 4000 rpm.
It follows ldrlts_w.

I don't really need more load for my car, but I want to understand if I am hitting a hard load limit here.
I attached the relevant section of the FR.
My KFLDHBN is plenty high.

I see that KFLDHBN and pu are both 8 bit values.
When they get multiplied, what happens to the result? Could I be hitting a FF limitation?

I'll try setting the KFLDHBN map to high values, and checking...

Ambient pressure was 1006 mbar, my KFLDHBN as is now is attached.
My logs are attached in a prior post....

I hit 2.461 request and stay there (DSLOFS was set to 0, but I've set it back now).

I honestly have NFI why I am hitting this limitation based on the FR.
I am guessing I have to start to disassemble this ECU. If only there was some information shared on that as well (descriptions and names of subroutines for example). I'm not scared of ASM, it just always takes so long to get into it...

Will FF KFLDHBN tomorrow, but I got a feeling it's not going to help.
Maybe I'm wrong and it will.

Ok, so rising KFLDHBN allows over 300 load into ldrlts_w, that is not a problem at all.
And now requested load can be adjusted at will.

However, plsol_w and pssol_w are still capped at 2461mbar. I can request 240 load or 215 load and plsol_w is still 2461.

So I guess I need to look at the boost calculation and see.
Because if I want to tune the car to produce in excess of the maximum measured boost pressure for some reason, then I need to be able to request FF boost.

Here is my latest log.

You can see how req boost is just capped, and then there is a sudden notch at cam changeover.

Thinking - if after cam changeover pressure is allowed higher than 2461mbar, then probably it has something to do with the selected camshaft position.

The culprit should be fpbrkds_w together with KFPBRK and KFPBRKNW.
I think that boost should be FF after rlsol_w is divided by fupsrl_w in FUEDK.

Then whatever PRG does does not matter, since it's an addition.

And finally, that gets DIVIDED by fpbrkds_w. And fpbrkds_w is set HIGHER with NWS active, based on KFPBRKNW.
Now when fpbrkds_w drops, requested boost gets lower as well.

My fnwue is 0.99604 3000 to 5000 rpm.
KFPBRKNW is 1.0564 in the same range. If I apply NWS correctly (0.99604*1.0564 + (1-0.99604)*1.0463) I get 1.0563.

The calculations there are 16 bit, so maximum requested value should be 65535/25.6/1.0563 = 2423.51 mbar.
I get 2461 - maybe that's due to PRG, maybe I'm just wrong.

But it seems to me that you can pretty much only request 2550 mbar if fpbrkds_w = 1.
And if that is true, then there is no point touching DSLOFS, since fpbrkds_w is never 1, but it's slightly larger.

Need to test my theory. I will probably also set KFPBRKNW and KFPBRK to the same value around the switchover point to prevent the PID going nuts.
I still want to PID control my boost, as I don't really need more than 2500 mbar.

It seems I was right.

rlsol is converted to pressure, and that value is capped. So FFFF.
Then pirg_w is added on top of that.

After that it is divided by fpbrkds_w.

I played with KFPBRK/NW and now I got nice boost response. 22 taper 20 without any boost oscillations...

Yup. I also flattened KFPBRK/NW

Not a fan of the huge discontinuity in req boost.

Not good for PIDs (as i point out in the s4wiki tuning article)

Flattenened as in you set it to 1.01581 in all cells?

yea.

Cool, Thanks.  I know the tables were mentioned in the Wiki, I just didnt know what should be done to them to make things work out. 

I set the last column of NW similar to the other one, so I don't get a rough changeover.
I would not say flattening it is the best way to go, but that's just my opinion.


Btw, it is always better to work out what needs to be done to them by yourself, looking at the FR.
You will learn many useful things in the process and you will also know what you can potentially break.

Before you flatten that table, look up what else it's output is used for.
Of course, if you don't speak german, it might be a bit complicated... oh well.

I agree.  I have been working on a bunch of other areas of my tune lately, and have been following threads like yours.  I didnt make the changes yet.  I like to know what the changes do, and what else can happen because of the changes.  I also do a lot of comparing of proven tunes.  I noticed these two maps were flat on most tunes I have saved, I just never knew why.  I have a lot of homework to do.

BTW,,,,Nice work

It is good you want to find out things for your own. That's the right attitude.
However, don't take "proven tunes" to heart too much. I learned that a long time ago.

Those, who speak in cryptic riddles on forums are usually not nearly as smart, as they want themselves seen to be.

Agreed with EVERYTHING in this post.

I second this.

I mostly compare the maps from Berttos's stage 3 file, Nefmoto's stage 3 file, and my base tune sent to me by Storm trooper.  I dont copy the changes, I try and find what they have changed and then research on my own why the changes were made.  I trust these tunes because most of the work done in them is documented in this forum.  I really like how almost every map  threee have modded, and I have questions about, there will be a 5 page thread about it with some really great input and the thread usually ends with a solid conclusion.  I am basically using their tunes as a resource as I make my own tune.

Right now I am focusing on load and am soaking up anything I can find related to it.   

had to bump this thread while searching for KFPBRK where PRJ actually has a question :o

Yeah like 8 years ago when I was just getting started with my B5 RS4 :)