So after I already described in another tread - I have 1.8TFSI engine in A3 8P car, that I stroked to 2.0 displacement.
So it is time to make some changes in the software!

I was planning to use PCMFlash tool that I bought in the past to tune my previous MED17 1.4TFSI A3.
I'm very happy with this tool, because I'm just hobbyist not a professional tuner and I can't afford to spend thousands for a tool that I will use few times per year.
With PCMFlash you can buy a module when you need it, and modules are on a cheap side compared to other tools.

BUT.. it seems my current ECU (8p7907115h) is not present in Audi flashdaten database, so it is not present in virtual read database of PCMflash either.

However while I was thinking to reflash with ODIS-E K or N version of the file, which seems to use the same hardware part number, I decide to try to read original file with old MPPS V16 clone that I have in one drawer.
And unexpectedly it happens - MPPS successfully read a file from my original H ECU, and it was stated that program will fix checksum of the file also.
So far so good.

I made some changes in the file that I have now, and decide to write it with old MPPS. It was stating, that checksum was not ok (I already know this), and that it will take quite some time to correct it. I give it a go, and it seems to be working very well.. but until reaching 99% then I get checksum error message and not starting car with cooling fans on.
Ok! this was not nice.. I know that in this cases maybe I'll need original .sgo file in order to recover with ODIS-E, and I already know there is no such file
However I decide to disconnect cooling fans and to write back original file with MPPS - and unexpectedly it happens - everything vent ok to the end, and car is alive again.

At this point I was thinking MPPS is good for the job, but can't fix checksums correctly.
Ok I asked a friend for a help - send my file to him, in order to correct checksums in WinOLS, so I can just write corrected file to the ECU.
Unfortunately MPPS is stating that corrected file have bad checksum, and ask to allow to program to correct it while writing. Of course I know it is not good idea, so I aborted this session.

Good news is that I managed to recover, bad news is that I don't have idea what to do now.
I know that if I flash another stock version (K or N file) with ODIS-E and if it is not working, than I will be f**ked badly
What I know for sure - I don't want to open my ECU.

Any suggestions for possible solution are highly admitted.

Cheers,
Chris

The only reason that I'm anti ECU opening is that on newer VAG, they decide to put ECUs directly outside (on older it was in water tight boxes - much nicer solution).

Anyway there is a good news.
I decide to check again by correcting checksums before writing. I found a old version of OLS, checksumed my file again - and it was fine. MPPS doesn't want to correct checksum on its own, and write it correctly to the ECU.
I'm really amazed that this old cable made the job for me.
Taking into account how fast my friend send corrected file to me first time, I think he didn't click checkbox for RSA correction in WinOLS (it is quite slow).

So the car WAS very nice to drive, with 170-174g/s between 4800-5700rpm.
Max boost was 1bar, tapering down to 0.60 to redline.
Fueling was spot on on my request, 0 CRs.
Power delivery was very linear and smooth.

Unfortunately my joy didn't last long enough, to try some improvement on tune.

Last night when I was heading to home, I left the car to decelerate from highway speed to the stoplight on 4-th gear, around 1300rpm I felt some hesitation so I disengaged to neutral, and engine stops. I put 3-rd gear and restart it.
I stop on the stoplight waiting for the green signal and everything seems fine.
As soon as green signal comes and I was taking off, there was a awful beeeeeeep and red oil can comes to DIS with message to stop the engine and check oil level
I checked it and was few millimetres above min mark, definitely not the reason for this message. I take VCDS, and there was fault for low oil pressure sensor
Because my home was less than a kilometre away I decide to restart and to continue. Unfortunately after maybe 300 meters the signal comes again together with check coolant level.
I stopped the engine, and this time I look at all temperature readings in VCDS.
What was strange - t of the engine 120C, while t of the Radiator outlet 25C. So it seems like water pump failure. But this will not explain Oil pressure warning.
Unfortunately what I know for EA888, It is more likely balance shaft failure, and because water pump is driven by intake side balance shaft, so it leads also to no coolant circulation.

I hope in next 2-3 weeks to find time and place to take off front engine cover in order to confirm or deny my suspicions.

Yesterday I had very strange behavior.

Previous night I wrote a file with corrected turbine model taken from a stock file of TT model that I sourced turbocharger from and few other changes.
Cheksums corrected in WinOLS 4.51 (VM found on the net), written by MPPS like before, car starts, get it for a ride and it was w000wwwww... instead of lagging during spool of turbine I heard compressor surge at 2500rpm.
So far so good.
I leaved the car in front of the house for next day and last night when I get the car it won't start! Cooling fans starts blowing, ESP faults etc. (usual symptoms of bad programming).
Disconnected cooling fan fuse, connected VCDS and ECU was reachable but it looks strange - no component description in right upper windows, Measuring Blocks grayed out just like on UDS cars.
And when I drove the car immediately after writing it previous night it looks ok, even I logged with VCDS few custom made measuring blocks. 

How it is possible with no modification between last two evening to get to this situation?

I wrote previous file again trough OBD with MPPS (Checksums corrected in WinOLS 2.24 last year) and everything was fine again. This morning I tried the car again and it starts without problem.

I'm really confused what gets wrong, and what should I correct in this half bad file.

Cheers,
Chris

So you think old MPPS v16 was in the root of this strange behavior and not checksum correction of newer ols version.

Of course I prefer PCMflash, unfortunately last time when I tried to find virtual read file form my ECU there wasn't such.
I can't find also SGO file in order to restore with ODIS-E if somethings goes in bad direction.

This weekend I'll try again virtual read with PCMflash if there is a file already on there server.

I don't think 8P7907115H is hardware number.
I think 8P0907115Q is a hardware number for the same ECU and few others and Bosch hardware number is 0 261 S04 513

And in VCDS it is shown like 8P7 907 115 H  0030 so I think this is software number.
Software version should be 399984.

Yes I read treads that you mentioned for precontrol hack for MED17 but I think there was a reason that Bosch engineers decide to take this path of calculating precontrol element.
Because I can easily find calibrations used exactly for this turbo made by same guys from Bosch, VAG or whatever - why not to using their experience? At least they heavily tested their solution.

I'm totally convinced by your idea of using map based pre-control if I decide to use some aftermarket turbocharger.

As for the logger - you got me! I really don't even know there is such thing like proper logger for those ECUs.

P.S.
Your software number is 8P7907115H and your software version is 0030.
ALL other numbers are irrelevant on VAG.
These two numbers uniquely identify your software version.

Also, you will never find an SGO for your ECU because it is UDS. SGO was used up to TP2.0.

I attached your stock file that you can use with PCMFlash etc.

Ah, sorry.
I thought you had A4 1.8TFSI for some reason the entire time. I don't know why. The file even has 8P in part number.

The unfortunate thing is, I don't support your ECU in my logger. Only MED9 is supported for TP2.0.
That said, I have a Passat EDC17CP14 TP2.0 ECU here, so I can try to add support for MED17 over TP2.0 as well. It's just very different from MED9.