Hey everyone! I need some collective wisdom here, been having all sorts of trouble with this particular car.

This is not the same car I posted here about.
This one is basically stock, but is also converted to MT.
The main issue is: it just won't control the boost correctly. It overshoots to the point of throttle cut, then tries to recover and so on.

I uploaded the bone stock binary just to see how it behaves at stock boost levels, and guess what - control is still out of whack. Log and ecuxplot attached.

2 main questions:
a) What can be at fault? We tried multiple different N75s to no extent, pressure tested, etc. Everything seems to behave normal, actuator arms move, etc. The only guess that remains now - that actuator arms are actually out of adjustment. Do the symptoms line up with that assumption?
b) How to deal with it? I mean - if it's indeed actuator arms, re-adjusting them with engine installed is almost impossible. So, can this be dealt with by re-calibrating PID? I mean, I obviously can lower the DIMX across the board to follow my requested boost level, increase Q2 etc. Or should I rather re-adjust KFLDRL? Or is it "proper" to re-calibrate whole PID like it's outlined in the according thread?

Where to dig, what to do with this beast? Some input would be awesome! Thanks!

EDIT: noticed that my ecuxplot screenshot has two purple lines - and I can't find the button to edit attachments. Anyways, the one at 100% is wdkba, and the one below is n75.
EDIT2: Modified the title to note that the problem is mostly resolved. Wastegate actuators out of sync.

Well, that's where I started.
It still overshoots like crazy, to the point of throttle cut. My ecuxplot is still misbehaving, but in the second (TC) graph, the car is actually modulating boost by the throttle plate while n75 stays ~constant. I can fetch logs if that'll help with diagnosis. Basically, the similarity between behavior with stock and increased LDRXN had led me to the belief that it's rather a hardware issue.

Note, in this case, the tune already has DIMX lowered and Q2 increased. Only ~10-15% for both though, not much. Just tried to see if there will be any effect - and these slight changes didn't seem to matter much. Next step would be setting KFLDRAPP to zeros and seeing how it runs on just the WG cans, I guess.

Roger. Thanks!
Just thought it was quite weird to see that behavior with stock tune as well. But I'll just dig into PID then.

Actuator arms seem to move in sync when control line is pressurized. But, what actually happens inside the turbo, obviously remains unseen. I guess I'll just do the CWMDAPP=8/KFLDRAPP=0 test, to see what's happening without the N75. I'll update the thread with results.

Do your mafs match? Always replace in pairs. You can run the sensors out of any 2001+ wideband 1.8t as they are identical, just use the RS housings.

As for the boost, I didn't realize this was stock. Yes, something is not right. They don't behave this way stock.

You say that, but if you swap them side to side are the readings identical?

One reference line from drivers side turbo, single N75 control T'ed to WGs.

At this moment, I'm not 100% sure on what is the final cause. I have a strong feeling that this might be a heavily pinched boost control line to the "MAF2" turbo. Gonna re-test it.
"Fixing" it with software changes is out of question unfortunately, yeah.

Run with N75 electrically disconnected: