Back in the early 80's I was designing Ku band frequency multipliers for the terrestrial end of a spread-spectrum satcom system. We laid out the Ku band multipliers using Rubylith and an X-Acto knife at 10X scale which were then shrunk to 1X photographically. A typical layout covered an entire 8 foot long table while we were working on it.
I finally managed to wangle a copy of Pcad and a 8MHz IBM PC-AT for our next design. It didn't get any better than that......
A friend taught me to use paint to draw PCB traces in the '70s. It worked like this:
1) design the PCB layout on graphics grid paper (in reverse, since this will be the bottom of a 2-sided PCB);
2) tape the paer over PCB, drill holes manually;
3) use an ink drawing pen (that holds tiny amount of ink droplet between 2 adjustable blades), draw traces between holes with household paint and a ruler;
4) wait till paint dry, then wash PCB in chemical (forgot, maybe Fe3SO4?, I bought it in yellow powder chunks), to eat away unpainted copper area;
5) wash off paint with solvent (gasoline from my CB100's fule tank).
It worked very well, though time consuming and you could only make one at a time.
Back in the late 80s, we had an in-house PCB layout tool that allowed you to plot to a desktop HP 5-pen plotter in which one of the pens was replaced with a sharp-pointed steel cutter -- something our mechanical techs built that looked like a weaponized HP plotter pen :) Only the outlines of the traces were plotted, of course, so the post-plot peeling of the rubylith was pretty simple...unless the cutting pen was in need of sharpening.
I remember at the time thinking how cool and high tech it was to be able to draw a PCB layout, complete with all those surface mount components precisely placed, and then just plot to rubylith, peel it, send it to our in-house etch shop, and get a finished PCB back within a day or two.
Ahhh Rubyliths...brings back memories. I can remember one of the first jobs I had as a young twentysomething new hire at Hughes Aircraft in Los Angeles was peeling traces on Rubyliths. This was the time when we would design microwave circuits using Touchstone, then do the layout in Autocad, convert DXF to Gerber, then print the Rubylithh using the old HP floor-standing plotters.
We were lucky in that we had a wet etch facility that could perform the reduced exposure and etching process, with quick turn around. This technique was standard for many years with the microwave integrated circuit (MIC)crowd. By the mid 90's, there was a move towards contracting out of house to use companies that would supply actual sized mylar liths for etching. I remember one could also pay more and get the circuit on glass sheets for dry etching as well. Dry etching was popular for the higher frequency (Ku/Ka band and above) because of the reduced geometries of the circuits. Fun stuff...
Was musing on this some more ….. Does anyone remember "Decon-Dalo" pens? They were like a felt-tip pen with an etch-resist substance in them. You drew your tracks and pads directly on the raw PCB and then etched it. I got one in a "PCB Design Kit" I saw in Practical Electronics for a couple of pounds when I was about 14. It had the pen, a few small bits of PCB, a small bag of Ferric Chloride etchant crystals, and a couple of sheets of design ideas. Pretty crude, but it got me started.
When the Ferric chloride ran out, I couldn't get any more to etch my PCBs with. My physics teacher was into electronics and he gave me this recipe.
Take a bit of swimming pool acid (In Zimbabwe everyone had a pool so that was easy - it was fairly strong Hydrochloric Acid HCl) and dissolve some old Iron nails in it. This makes a greenish liquid which is Ferrous Chloride FeCl2. You then pour some Hydrogen Peroxide (seem to remember that is H2O2) into it and after some bubbling and smelly stuff you have Ferric Chloride FeCl3. I don't remember the reaction, any chemists out there? It made great etchant, it lasted me for years and never seemed to lost its strength.
Another great product was Positiv 20 made by Kontakt Chemie in Germany. You sprayed this onto your PC board in subdued light, dried it in mom's oven (set on very low) and you had a photosensitised board ready for exposure with a positive master. A couple of minutes in the African sun did the trick, then develop in a 7% solution of Sodium Hydroxide - Mom's drain cleaner - and etch in my home-made Ferric Chloride. The beauty was that if you stuffed up, you just sprayed your board again.
Hardly commercial PCB production this, but it was great fun, got good results and was not that different from what I do today, presensitised board and Ammonium Persulphate etchant, but otherwise much the same process. I doubt if Duane would give me a job on the strength of my years of PCB production expertise, though?? ;-)
Nice article Aubrey. I remember Rubylith, I produced a PCB for a +/-5, +/- 12 V Power supply for a breadboard with it. I might still have it around somewhere. I did a lot of work with Bishops Graphics (I saw but never used the red/blue double sided stuff) and remember liking it, even though it was fiddly - when you got your finished PC it was almost entirely your own work and gave considerable satisfaction (especially if it worked!)
These days my job does not involve PCB design but I still produce my own at home. I have never got into the CAD programs (laziness more than anything else I think) and still use a program which is essentially a "Bishops graphics on the screen" type of thing - PCB Designer by Niche software, a British outfit who I think went out of business a few years ago - anyone else know anything to the contrary?
It's certainly come a long way - look at today's multilayer PCBs and tiny tracks and the reliability of the boards - it's certainly a feat of technology. I've read some of Duane's (and others') articles on PCBs and they are as high tech as the chips themselves these days.
I was always told that -- in ye olden days -- you never saw PCB Layout Designers wearing woolly pullovers...
...this was due to the fact that if a thread dropped onto the layout it could end up being imaged and causing a short circuit in the final layout...