Self-assembly: the computer soldered together on a kitchen table

A computer that wasn’t in the shops

Picture a kitchen table in the late 1980s: instead of dinner, on it lie a soldering iron, a reel of solder, a sheet with a schematic and a green fibreglass board studded with dozens of chips. This is not a factory workshop. This is home self-assembly of a Spectrum — and for thousands of people this is exactly what their first personal computer looked like.

Why, behind the Iron Curtain, the Spectrum was reproduced by hand we have already written about in the piece “The Spectrum behind the Iron Curtain”. Here we are concerned with the specific practice, not the general cause. The main thesis is simple: self-assembly became mass not out of enthusiasm alone but out of necessity. Ready-made machines were in short supply, while the Spectrum’s circuitry was fairly simple, could be reproduced from ordinary chips, and the schematics themselves circulated freely from hand to hand. A soldering iron on the kitchen table became the entry point into the world of the computer.

Why people soldered it themselves

There were two reasons, and they reinforced each other. The first was shortage: a ready-made personal computer on open sale was the exception rather than the rule, and an imported Sinclair original remained out of reach for most. The second was the accessibility of the design itself: the Spectrum turned out to be a machine that could realistically be built from a component base that already existed on the local market.

These two circumstances gave rise to the “do-it-yourself” culture. If a machine cannot be bought but can be soldered together, someone will solder it. And that “someone” was not a lone genius but a mass phenomenon: schoolchildren, students, engineers from radio factories, seasoned radio amateurs. For some it was a way to get a computer more cheaply, for others an engineering challenge, for others both at once.

A schematic from loose chips: why the Spectrum could be soldered at home

The key to self-assembly lay in one technical detail. In the genuine Spectrum the heart of the machine was the custom ULA chip — it took on video, sound and input/output (there is more on this in the breakdown of the architecture). The problem is that the ULA was effectively a “black box”: to make sense of it and reproduce it was the hardest part. So many clones took a detour — they abandoned the ULA altogether and reproduced its functions with ordinary logic chips.

It is exactly this that made the machine suitable for self-assembly. “Loose” chips — dozens of standard logic packages — could be obtained and soldered at home; a custom die from Britain could not. The best-known example of this approach for amateurs is the “Leningrad” design, developed in 1988 by Sergei Zonov. It got by with about 44 ordinary chips, without any programmable logic, while remaining compatible with the overwhelming majority of Spectrum programs. Its author made the schematic open, and it became one of the first genuinely “buildable” designs aimed precisely at an amateur with a soldering iron.

Design inheritance. A British engineer folded the Spectrum’s logic into a single die for the sake of cheapness on the assembly line. An amateur with a soldering iron made the reverse move — unfolded it back into loose ordinary chips so that the machine could be built at home at all. The factory-made Ukrainian clones — the Ikar, the Robik, the Orel — are the same principle, only carried onto a production line.

Schematics that passed from hand to hand

You can only solder what you have a schematic for. And schematics did circulate — in printed form, from person to person. The radio-amateur periodicals of the era were a familiar channel by which designs for home computers reached a wide audience: the Union-wide magazine “Radio”, for instance, at one time published the schematic of the “Radio-86RK” home computer and set the very model — “take a magazine and build it yourself”. For Spectrum-compatible machines the schematics spread by similar routes: sheets redrawn by hand, copies glued together into one large drawing, printouts that passed from hand to hand at the radio market.

Here an honest boundary matters. That radio-amateur magazines were the environment in which home-computer schematics lived is a documented fact. But which magazine and which issue first printed a specific Spectrum-compatible schematic, where Ukrainian builders got it and what they redrew by hand — this is exactly what needs verification against primary sources, not retelling.

The radio-construction kit: a set of parts in little bags

In time self-assembly became so widespread that a whole format was born — the radio-construction kit. This was no longer just a schematic but a set: a printed board, chips and small parts sorted into little bags, a printed schematic. In effect it was “build it yourself” in a box, from which a working computer was only a few evenings with a soldering iron away.

Where both the kits and the loose “rozsyp” were obtained we describe in detail in the piece “The radio market” — it was the market that was the infrastructure feeding self-assembly with parts, boards and flashed ROM chips. The specifics, though — what exactly went into a typical kit, how much it cost and who assembled it — we deliberately do not present as fact: these details are poorly documented and are set aside for the museum’s verification.

Debugging: where everything went wrong

Soldering a board is half the job. The other half, often the longer one, is debugging. A home-made machine rarely “came to life” at the first switch-on, and it was at this stage that the builder became, willy-nilly, an engineer.

The typical troubles were predictable. A dry joint or, conversely, a bead of solder bridging two tracks. A chip inserted the wrong way round. An error in the redrawn schematic itself, which the builder faithfully reproduced. A badly flashed or “drooping” ROM. Loose logic assembled without a ULA is, on top of that, sensitive to fine timing relationships: the machine might start up but “spray” colour or drop cassette loading. Diagnosis without an oscilloscope turned into detective work — and at the same time into the best school of digital circuitry one could imagine.

What we are checking. Printed-schematic self-assembly is documented worse than factory machines: it leaves behind no receipts or passports of its own. We are looking for the specifics of precisely the Ukrainian context — which magazines and issues with schematics were in circulation, what the radio-construction kits looked like, how much they cost, who assembled them and where. If you have kept schematics, printouts, photographs of home-made boards or your own recollections of building one — write to us. We will refine the material and, with your consent, add your materials to the collection.

Ukrainian builders and Ukrainian machines

It is easy to romanticise self-assembly as “Soviet engineering”, but its value for our museum lies in a different focus. Specific people soldered in specific cities, and it was fed by a specific milieu: radio markets, clubs, factory design bureaus where you got loose parts and advice. The do-it-yourself culture was local, everyday, Ukrainian — not an abstract “centre” but a kitchen table in Kharkiv, Cherkasy or Dnipro.

That is exactly why self-assembly and the factory clones are two sides of one story. The most widespread Ukrainian machines — the Kharkiv Ikar, the Cherkasy Robik, the Dnipro Orel — grew out of the same logic of “we’ll make it ourselves from what there is”. We consider them separately, in the clones catalogue, where each machine has its own city, factory and provenance, verified against sources. The years and print runs of individual machines we deliberately do not fix here: they differ from one reference work to another, so they live in the catalogue, where they can be refined.

Not to be confused. Self-assembly is a machine soldered by an amateur from a schematic at home. The Ikar, the Robik or the Orel are serial factory clones. And one more thing: the heart of a Spectrum-compatible machine was the Z80 processor (the Soviet analogue being the T34VM1 / KR1858VM1, from 1991), not the KR580VM80A — that is a clone of an entirely different processor, the Intel 8080.

In the museum’s collection

A home-made board, a soldering iron and a sheet with a schematic are not about one specific specimen but about a whole way of life with technology. Such a “typology exhibit” shows what is visible neither in a glossy Sinclair nor even in a factory clone: the trace of a human hand. An uneven row of chips, a soldered-on flying wire, a pencil mark on the schematic — all of these are traces of the evening when someone first switched on a freshly soldered machine and waited to see whether the familiar start-up screen would appear.

We show self-assembly precisely as a typology, not as an attributed specimen with a known author: the provenance of an individual amateur board is most often lost, and it is more honest to say so plainly than to invent a caption. If you recognise your own work, or the work of someone in your family, in such a board — that is exactly the story the museum is most short of.