This one?
1. My own preproduction prototype Model M also has some PN etched on the bottom of the PCB and a different one written on a manually applied skewed sticker on the top side. And my prototype was built in 199-something in Greenock and has a warranty sticker from some retailer 
2. Product numbers are just that - identificators of production units.
3. Soldered down, permanently programmed microcontroller doesn't exactly smell like a dev-board to me.
4. A keyboard can't be older than any of its parts, in particular the 1986 buckling spring assembly.
5. Did IBM even do any Model M R&D in Greenock?
6. Your nerd-penis officially grows by 2" for this square clock, but I'm afraid that's all.
It doesn't have to be a dev ver, square clock early test field model maybe, like an early alpha :O? Is there a official IBM PN register?
How did they it end up with this type of clock if only it was avaliable for R&D and test depts? has it appeared in model F? its at least certainly rare, in fact I haven't seen one ever appart from the model of the OP and mine. It shouldn't be there.
The only part surely to have been made in the UK is the steel plate, the PCB could've sent from abroad.
The 1986 steel plate could've been swaped afterwards for repairs.
Look, the OP's keyboard is a prototype version of the M Industrial and we are 100%
positive of this fact. IBM never used surgical tape or UV-EPROMs in production M's, and Ferrari is a known M code name.
That said, what you need to understand is how rapid prototyping worked in the 1980's and how IBM manufacturing worked. We've now found a total of three examples with the square crystal, with closely spaced serial numbers - but the other two had it adapted to the production controller layout.
When you did rapid prototyping in the 80's, it was not rapid. Not unless you hacked up a production part. Otherwise, you had to make a LOT of parts that could be used for prototype and production. IBM tended to favor the 'build a whole bunch' approach generally. It would not surprise me to see
similar PCBs in production parts - but they will
always an entirely different uC and certain other identifying marks. Like the FRUs and part numbers. The alternative was modifying a production part.
They'd also use a lot of parts right off the line, including QC rejects, intended-for-new, or whatever was laying around on the shelf. If they had a part stamped 1986 in a prototype, that wasn't indicative of anything. Could be the part was a mis-stamp, could be a 1986 part, could be they tossed one plate and grabbed another off the line. Or disassembled another.
Literally these things were largely built out of whatever was laying around or available on short notice.
Mandolin, your keyboard is a production part. Period. The back label went missing or wasn't applied - it happens - but it's still a production part. Period. Not a prototype, not all that interesting aside from the sloppiness on the controller. That's not 'manual application' that's 'the machine wasn't calibrated correctly.' IBM would not reject for crappy silk screening and it is also not skewed - C3 always looks like that. In fact, I can tell you exactly what model and why the sticker's missing.
It's a 1988-1989 Model M, International Market, manufactured in the United Kingdom. All stickers for these keyboards fall off as they're low quality and frequently applied incorrectly. (They were applied so that they were half in the recess and half out.) There was also no lamination applied. The adhesive ages and they fall off or just get worn off.
Controllers have compound part numbers. It's a 1392993-1351526. An
uncommon controller to find I will grant you, but it is also a production part. Foreign language, though I couldn't tell you which off the top of my head. All of the foreign controllers have unique part numbers because of the IBM way. Since the EEPROM load is different they require a different uC line be used, which means they are given a distinct PN. This is true for almost all of the non-US non-ANSI Model M's.
As to the rest of it? Welcome to standard practice at the IBM UK/Ireland plant. The PN stamp isn't hand applied, just halfassed. The markings aren't prototyping - those are the QC pass marks. (In fact, all IBM UK plant M's have pen or marker initials on the label for QC pass.) It's a unique thing to the UK facility, but that's how they built every last one of them. Gods only know why. There's also all kinds of crazy part shenanigans and substitutions, half the chassis are missing the blue date stamp on the lower half (if not more), and so on.
