If you have zero idea, then you won't even get a project off the ground. Have you ever done board design? Ever completed a project with an integrated microcontroller from start to finish (schematic, PCB design, fab, assembly, debug, etc)? There are too many variables for chipsets, board designs, I/O, etc. If you're just designing a single PCB, that will also carry a hefty price tag for the size and layers needed, 4 layer and at least baby-AT / micro-ATX size. Also, you're going to have to design your own schematics, unless you plan on doing a 1:1 copy of an exact replica board - which in and of itself is still very challenging to accomplish with even a small team of knowledgeable people working on it. You'll have to source parts that are no longer in production, or design all your own equivalents and have a microchip manufacturer make them. At that point, you might as well just buy a $20-$100 286/386/486/etc motherboard and call it a day since they are plentiful.

If you're thinking about a Sinclair ZX Spectrum, all the schematics are on the service manuals but it can't be remanufactured anymore. The main reason is the ULA, that was a custom chip made by Ferranti not available anymore. But the good news is that it has been manufactured (at least) three times with public schematics:

- Superfo Harlequin: A board 100% compatible (hardware, software and form wise) with the original ZX Spectrum. On this page (sorry, spanish) you'll find schematics and board files to get your own boards made. The main differences with the original spectrum are that the ULA has been replaced with TTL chips, two 32K RAM chips instead of 16+ chips on original Spectrum and everything works at +5V (the DC-DC converter was known to fail and blow the lower RAM). Also, it has composite video output instead of RF and optional RGB video output. You can put this board into a standard ZX Spectrum or ZX Spectrum + case.
- Harlequin 128k (on the same page): A expansion board / mod for the original Harlequin boards. It only provides 128k memory, not the other (AY) features present on 128k Spectrums.
- ZX-Uno: This is a FPGA based board that was intended to be a ZX Spectrum clone... but may emulate other computers ("cores"). The most developed core is the ZX Spectrum one (emulating every Sinclair and Timex model), but there are MSX, Nintendo Entertainment System, Sega Master System and even a PC-XT core. The storage device is a SD card, and it has PS/2 connector for your keyboard and composite video output.

I've got a Harlequin on a ZX Spectrum + case, and worked with everything I've got. IMHO is the best clone, and being manufactured with old TTL chips means that everything (except the SMD optional RGB output) can be soldered without any special skills. The ZX-Uno is a interesting toy because can emulate many 8 bit computers... but it's not really a ZX Spectrum clone.

On the non-freely (without schematics) modern clones (russians manufactured lots of varying compatible clones on the 90's), you can get a ZX Next or a ZX Vega, but I wouldn't recommend it. After all, they're not real clones (like the Harlequin)... and the ZX-Uno is cheaper than both of them.

I built a 8088 computer on a breadboard following a book (8088 Project) and it was quite a project. Can't imaging doing anything 16 or 32 bit. But I am sure someone has done it 😀

Zup wrote on 2020-05-04, 14:07:

If you're thinking about a Sinclair ZX Spectrum, all the schematics are on the service manuals but it can't be remanufactured anymore. The main reason is the ULA, that was a custom chip made by Ferranti not available anymore.

ULA, as the name suggests, is a PLD. There have been some attempts to RE it and I think some people did? And even if not, the logic can be re-constructed in modern PLD/FPGA based on it's behaviour. I mean pretty much all it does it steal CPU cycles to do video output. AFAIR the very early ZXs had RAM refresh circuit done on a bunch of 74 logic, only the later ones incorporated that too into the ULA. One could start with ZX81, same configuration but simpler system - and then build on that.

Anyway, the idea to just sit and "build a computer" seems ridiculous to me, because I know what's involved. However I wouldn't be what I am now if someone kept telling the kid-me "you can't do that". So while I won't help, I don't have that kind of free time, I wanted to tell the OP to not give up his dream. Just be prepared for 2 things: that you will fail a lot, and that this hobby - like any other - costs money. And one advice I can give is to start with a repair job or two of some 8-bit machines like C64 or Atari (or indeed a ZX). While repair can be a tedious task, the ability to figure out where the problem lies is a necessary skill to have - as you have to assume your first project will not work properly from the moment you power it on. Either due to design or construction errors.