The problem #
When I got the spectrum with consent from my last university, It came kind of functional.
By this I mean that it turned on, but it almost never actually booted to Windows (insane for its time). It showed a screen of the booting process, like booting all the analog modules (of which there are a lot) and making sure the transputers were booted.
Transputers are akin to the gpu’s we have today, and are like them a computer on a seperate board, made specifically to calculate things in parrallel.
When it was done with booting all the internal peripherals, it wanted to start booting Windows (in this case Windows NT), but this only worked when, you restarted immediately after everything booted, without disconnecting the power. what was observed was, that this was almost always the hard drive that stopped spinning, and halted the boot process.
Having a failing drive for something like this is terrible, because even when reaching out to Rohde & Schwarz (they did not reply 😒), I had no way of getting this software back for this machine.
So here comes the hail mary. I found out that on the EEVBlog, there was someone who kind of attempted the same feat, it was also the only information available online about it.
this is the project: EEVBlog (if it is not available anymore, copy the link into waybackmachine, you’re welcome)
In their instance, they seemed to be working with a slightly different machine, where the computational unit was a seperate plate on the backplane.
They have a nice image of the backplane: 
On this images we can see in the centre the backplane, this is surounded by aluminium locking rails, and connectors on the other side, with a digital communication bus. This bus then links onto the pc board, (on the left side of this image). you can see all the analog boards connected to the bus, and connected via gold plated copper coaxial cables (which are labeled amazingly, thanks R&S) What is also to note is the input filter box on the bottom right.