I used a big turbo tune that works in lots ot other cars so ive finally confirmed the tune is not the problem unless the problem is within any of the 5 parameters i did change. Only maps i changed were for my hardware; krkte, tvub, frlfsdp, dslofs, and dslgrad. Ive taken like 300 logs trying to figure out what this problem could be from. The throttle actually cuts before any load request or boost request dips. With any of the mi variables i can log none of them show any change until the load request starts to dip which is after the throttle cut.

Btw if you cant even explain any specifics about what im doing wrong and only read what you want to read from my posts, why even post anything? You think i like wasting my time? Im actually trying my best, and im also only here to learn, not have some dickhead try to make me feel dumb for trying and asking for help when i need it.

Anybody can tell me to go understand the ecu better why not be more specific for my situation? Ive spent many hours reading the fr trying to get to the bottom of this. Im logging nearly all the diagnostic parameters including all the ones people in this thread mentioned and none of them indicate anything until after the throttle closes. If its not one of the 5 parameters that i changed in that file then its a hardware issue.

I have a genuine bosch 4bar sensor for a tdi i think and as far as i know i have the right numbers for it. 874 grad and 68 ofs with no 5120 iirc

Those posts about the pressure sensor scaling werent me.
I said probably 10 times that ive started with a stock file with the only changes being for injectors pressure sensor, ldrxn, and irl/iop numerous times. today in an effort to rule out a pressure sensor scaling issue i swapped in the stock pressure sensor and *drumroll* i got throttle cut at 6000rpm.
Maybe ill flash a stock file with only krkte, tvub, ldrxn and kfldhbn changed and give it a go. thatll be the closest thing to stock ill be able to flash without triggering overboost faults. my wastegate wont really let me make under 20psi so i have no choice but to raise some limits.

what sort of hardware issue can cause this? this is the 2nd motor with all different sensors but i suppose something could be happening in the harness but not enough to throw a code?

I have adjusted the boost pid and the axis for wastegate cracking pressure but it never made any difference in when or if the throttle cuts. Ive also adjusted to the timers for negative and positive boost deviation as well as ff"ing kldlul and got no change. Ill figure out how to find those parameters to log just to be sure. Im convinced the issue is from my ecu harness and i think its easier to swap harnesses than read and understand 1800 pages of translated german to figure out where the error is coming from. You also gotta understand that im only trying to tune MY cars and then leave them alone, i dont really have any need to become a full on me7 professional if i dont have to. id even buy a tune but im afraid the issue will still be there.
I guess i can get out the compound turbo car with essentially the same setup except stock-like boost onset and see if it happens in that car.

i think i fingered it out or at least i got good enough of results that i can settle if i have to.

It was the rev limit torque pid. Based on a predicted .3 seconds before hitting nmax the desired torque declines as a pre-rev limit. Im assuming the function exists so the driver will choose to shift gears because of the dysphoric lack of power the closer they get to redline, and/or because a gradual torque reduction limiter is smoother than a fuel cut, either way; I hate it.

The only map i changed so far was KLTNMXPR: "Time (seconds) for speed prediction of the NMAX control". I set all values to 0. Now my throttle cut doesn't happen until 6900rpm, but my requested torque dipped at 6700, which was caused by DNSIRES I assume?

DNSIRES is the rpm region before NMAX that the torque request dips as the pre-rpm limit. It was 300rpm stock. NMXPRLG is the proportional gain of the rev limit pid. KINMXRLG is the integral component.

As of right now I've lowered DNMAXH to 20rpm, lowered DNSIRES to 20rpm, set NMAX to 7200 for my stock head, lower the P and I components for the nmax pid, and I'll keep KLTNMXPR at 0. With these settings I think I'll have a pretty normal rev limiter. Once I get another log I'll report my results.

You'd be right if i didnt need to learn how to find new ram variables. I didnt know what minmx was and didnt have it available to log. Maybe you couldve mentioned my problem could be the rev limit and why that specific variable is so important? But honestly if you did that i still probably wouldnt have taken the time to dissect my ecu and find the address. I probably wouldve just did what i did sunday and looked through the nmaxmd section in the 518al damos and figured it out and fixed it in 5 minutes, which is what i did.
Sorry for not blindly following directions without understanding

I said that the problem comes when misol (soll-moment) drops, so the requested torque. That's the key.
It's a minimum of driver-requested, asr, getriebe and other sources that could possible request torque reduction
Log them, check FR for a source of bad guy, that's all. Be logical in your investigation, simply as that.