How double sided PCBs were made

[falter]

I just got an original SWTPC keyboard - the first design featured in Don Lancaster's Popular Electronics article. Prior to getting this keyboard, a gentleman had sent me an original popular electronics construction guide for it, complete with the PCB artwork, which I'd never seen posted anywhere before. Anyway, I had scanned it and etched some PCBs on a trial basis, because I was planning to recreate it just for fun.

The way *I'm* doing it is by etching each side on 0.030" copper clad. I'm using it because it's genuine vintage PCB stock from the early 70s - has the right color, etc. And then I basically align the two sides and adhere with contact glue. I figured this was not something people did back in the day since you could buy 0.060" double sided easily enough. But I'm wondering now - I'm looking at the edge around this original SWTPC PCB, and I can see a distinct line in the middle, like they did what I did - etched one side, then the other separately, and then adhered them together after the fact. Is that possible? Or is it just an illusion from the way they cut the PCBs? Just wondering if I was accidentally doing something legitimately or not.

Pics here.

 

[Chuck(G)]

Etched both sides at the same time. Double-sided and multilayer boards were tricky before CAD. Registration was a nightmare--get one side a little off and your vias don't connect--or worse, they connect to the wrong trace. The usual commercial practice was to do the artwork on mylar sheet perhaps 4x size. Mostly dot-and-tape, with the occasional patch made with India ink. Each layer had a "target" pattern, and the various layers were stacked up on a light table for checking. Each sheet was shot by process camera and the image reduced to correct size. Photosensitive PCB stock was used and each reduced negative carefully registered. The result was exposed under UV light; the light hardened the coating on the clear areas of the negative. The board was developed, which washed away the unexposed coating and the result was cleaned, and etched. Drilling was done in stages, if a multilayer unit; vias were plated where necessary. Basically that was it--a multilayer PCB was quite expensive to produce.

Lots of talent required. For some odd reason, the best talent on doing the artwork seemed to be from East Asia (e.g. Taiwan, Singapore...)

Before the photoreduction process, there was Rubylith. There you cut into what amounts to a stencil using an Xacto knife. Early microprocessors (e.g. Intel 4004) were done this way. Massively labor-intensive.

You can understand why wire-wrap for prototype circuits was considered to be the "easy" way...

 

[Dwight Elvey]

Greater than 2 layer boards are stuck together. Two layer boards were usually done as a single sheet. Some were done as two separate boards stuck together. These either required plated through vias, ferrules or soldering on top of the board where top to bottom connection was needed.
To match typical edge connectors, they need to be 0.60 total thickness.
Dwight