Ferric chloride gone into liquid

[Richard_FM]

I remember Electro Supplies in Stockport used to sell bags of Ferric Chloride beads. I remember at least one of my friends considered making their own PCBs but couldn't afford the equipment!

[djrm]

I used to make my own boards using dil transfers and a pen for the tracks. Then etch by floating the board upside down in a developer tray or pyrex dish. Never made anything very complicated. I think the last thing was a ql memory expansion with edge connector.

[Craig]

I laboriously did the circuit board for the classic Texan amplifier that appeared in PW in 1972. Being a schoolkid and chemistry "lab boy" I had heard that dilute hydrochloric acid was a good copper etchant.

It certainly was. I watched in horror as it stripped the board clean - all the laboriously painted tracks floated off.

First and last time I tried that!

Craig

[ppppenguin]

While HCl may attack whatever glues the copper to the board and may attack what you're using as a resist it certainly doesn't dissolve copper.

What did people use for resist back then? I think I tried the special (Dalo?) pens, sticky backed plastic cut out with a craft knife.

Later on I realised that DIY PCBs with those methods was too much trouble. I used a lot of matrix board (like Veroboard without copper strips). This is the prototype PALl/NTSC encoder that was my first professional freelance design in 1987. I think I still have it and it probably still works. The main purpose of that design was to give a PAL output from a genlocked BBC Micro. A production board, on a PCB, is in service at the Dulwich museum, encoding the output of the COW (Computer Originated World)

[CambridgeWorks]

In 1981 I made the Elektor 2m to 70cm transverter.
A workmate used the spray and photoresist method, I etched and drilled it. Double sided, quite a task.
I couldnt get the tx alignment to work, until I borrowed the depot 1GHz (Texsacan Wavetek??) spectrum analyser.
Fitted nicely inside an eddystone diecast box.
Still have it, though not used for 35+ years.
Transverter for the 70cm band (1981) | Elektor Magazine https://share.google/HWYNYVWg7BYZ7x3GE

Rob

[BusyBee]

I have not etched many circuit boards. I found that I never really got on with them, finding the method too tedious. I remember designing  a fairly complex board at college, then realising a mistake, luckily before working on a board. I dropped the design and went to a more maintainable multi board veroboard design.

Having a quick look about I have found a (messy) early circuit. A light switch, I think. My usual method was to use enamel paint, hand painting the tracks but this one looks a bit gouged so I am unsure about it. I never made more than a handfull of etched boards anyway, quickly finding that I could easily implement a circuit diagram on veroboard directly without an interim layout.

   

[ppppenguin]

I've never tried it, but you can mechanically mill the copper rather than etch it. A CNC mill can automate the process. BusyBee's little board in the last post looks like a crude version of of milling. BusyBee also reminded me that I too used paint (car touch-up) as a resist.

I grew to really dislike standard Veroboard. Give me plain matrix board any day. Sometimes I've used "colander" board which is like matrix board but with a ground plane. I've occasionally used boards designed for DIL chips which have power buses. I've still got the wiring pens that use very thin solderable enamelled wire. haven't used them in years.

[Mike Watterson]

In my last place we needed fast turn around of UHF (800 to 950 MHz) PCBs and used a kit based mill. PC controller and bigger version of a Dremill as power tool. An included utility converted the plot files from Eagle for the PC controller application. We used a retired "Pizza box" 19″ rack mount "server" as the PC.

Steps:
1: Buy a suitable piece of material and fix it to bed. Then mill it.
2: Tape board to alignment marks and drill two peg holes through board and base
3: Mill one side. Avoid large empty areas in design. Might need two mills for performance. Smallest one for tracks between legs or SO devices.
4: Other side.
5: Drill holes. May need bit changes. Used SMD as much as possible.

I gave up etching in 1980 and used bureau in Ireland till 1990s. My first etched board (1970?) used mum's nail varnish as resist.

I've done a lot of construction on track side of vero/stripboard with often an added copper ground plane. SMD and regular parts (bent to sit on track). Also slit track at the holes to have x2 tracks for SO packages.

I'd never etch again. Even in 1990s I did some consultancy for local idiots with bubble tanks. I told them to outsource that. There are actually good local ones for volume. China if only wanting 5 to 10, which is cheaper than making one badly at home.
Online sellers with PTH etc and track widths infeasible for home etching. I'd only consider mill solution again, for same reason, experimenting with printed coils, striplines and capacitors. RF from 500 MHz to 20 GHz. Even at 800 MHz using 0.8mm is better and above 2GHz approx more exotic materials than regular fibre glass.

EDIT:
Last manual taped / "letraset" style symbols was 1991, a double sided Z80 single Eurocard controller. All PCBs after that used CAD (FutureNet, Orcad, Eagle and two others). Schematic to PCB checking of net-list and parts list is a must!

[BusyBee]

I've never tried it, but you can mechanically mill the copper rather than etch it.  A CNC mill can automate the process. BusyBee's little board in the last post looks like a crude version of of milling.  BusyBee also reminded me that I too used paint (car touch-up) as a resist.

I can remember finishing boards by hand but do remember totally cutting by hand was possible but error prone and tedious, so I think that that board was probably a combination. I have never really had access to machine tools except at college.

I still prefer veroboard, or similar or my modern take on etched board which is to use manhattan layout with copper faced board and glued on islands.

Tracy

[DrStrangelove]

I don't like veroboard much.

I found a matrix board version of the Stewart tape recorder design in a biscuit tin in the garage.

Point to point & much easier, with no requirement for a twist drill to remove connections.

One of my classic design cockups in the letraset days had an inverted 7660 negative bias generator due to some sort of brain fade.

Fortunately it was an 8 pin TO-99 and apart from looking like a spider, worked ok, which is more than might be said for the thing it was in.

About 20 or so years ago,on a contract in the Forest of Dean, the company had just acquired a milling machine for pcbs.

I wasn't around long enough to find out if it worked or not.

I also recall a UK pcb company who produced a pcb with no pads without bothering to ask if that was the design.

And another bunch who supplied us with "tested" 4 layer pcbs, some of which had shorts on the databus due to whiskers, and another that had a short from Vcc to Gnd due to a misplaced drilling. All good clean fun.

I've always wondered what "tested" meant. Checking for shorts didn't seem to be in the program.