Hi

As some tuning companies in my country offers "oem like" r32 turbo modifications I would like to ask if it is possible to "patch" r32 ecu to use stock n75, n249 and map sensor? That prepared ecu has functions of eg. s4 ecu (boost PID maps etc) and can be tuned in similar way? Does anyone from this forum provide such services?

I am far away from r32 turbo projects, but someday would like to give a try. For now I ask just out of curiosity.

I think Bische Performance does this aswell nowadays..

I agree. Wasn't really talking about completely factory-like functionality, rather some greatly simplified approach like you suggested. Patching stock code in can quickly snowball on you. However, I would probably still try to retain as much factory code as possible, only eliminating some redundant or time-consuming-to-patch-but-not-really-necessary features. It's not too hard, just tedious. Can probably do that in a couple of months, not a year. ME7 C167 code is largely portable, functions can usually be isolated as small code "chunks", which don't disturb each other or the stack and will properly work even when implanted into other ECU (provided, you alter all memory references ofc). E.g. I had copied one of the "double-adjustable cams" ZWGRU function to make my "WMI" ignition maps blend in smoothly with the factory ones.

I see the process like this:
1) isolate a list of "new" LDR functions (code chunks) that need to be patched in & their scheduling;
2) come up with a list of "new" memory vars needed;
3) cross-check which vars "intersect" with NA code, note which stock functions need patching;
(note: steps 1-3 can be greatly simplified by A2L parsing, but I would still go over those in IDA by hand)
4) modify assembly code of said "new" or "altered" functions to reference free RAM & ROM instead of originals;
5) patch everything in at appropriate call tables, interface with outputs;
6) check that everything works, de-bug and finally:
7) properly calibrate everything.

1-3 look relatively easy.
4-5 will take quite a bit of time, but are also doable, especially if automated.
It's 6 and 7 that scares me the most, because copying stuff over is easy, but getting it to actually function properly is a completely different thing.

I don't see this as an impossible task. Hard, tedious - yes. Impossible - no.
So I won't be really surprised if someone had already gotten through all the trouble, even though I find that highly unlikely.

But, I definitely agree, there are more ways than one to skin a cat.

I actually did look at it before writing so detailed reply. Like I said, I'm not arguing that it's difficult. There might be some catches (like not enough spare bitaddressable memory), but it can be done.

EDIT: (just to make an emphasis on it, if I didn't word it clear): I'm not talking about completely making everything factory-like. A lot of corners will have to be cut.

For the pressure before throttle plate "sensor" seems to put a table DPUPVDK which tells the intake pressure loss. Thought this could use to tell the throttle the (modelled) boost pressure?