Homebrew PCBs

At 08:02 AM 5/3/02 -0500, hans@... wrote:
>Been following all this about perf. boards and milling etc.
>
>I like the idea of milling but it seems the prep. work is tremendous.
>
>Doing pref. boards then PCB seems like hard labor.
>
>I'd like to have comments as to why pre-sensitzed PCB and etching is not
>mentioned and what you consider is so bad with it ?

We do both.

The photo process uses DuPont Riston 4215 laminate film, a much modified 
Ibico laminator supplied by Kepro, developer is Soda Ash (Potassium 
Carbonate), etch is Ammonium Persulphate, stripper is Caustic Soda (Sodium 
Hydroxide), drilling is by Gordon Robineau's PCB drill.  This is a very 
mature process - we have made thousands of boards with the process over the 
past 18 years or so.

The downside is the prep time.  The raw board has to go through a 5 step 
cleaning process before being laminated.  The negative has to be done at a 
local print shop - someone takes over a floppy and comes back with a 
negative.  Drilling used to be done by hand but is now with Gordon 
Robineau's seriously cool PC drill - I simply feed it the drill file from 
my CAD program.  The bottom line is that it is a half day process from 
start to finish - 1 board or 20 boards takes about the same amount of 
time.  The boards turn out perfect - it is a great process.  But it ties up 
somebody for that half day.

One of my techs purchased a dead engraving machine from the local repair 
outfit.  The computer part of it was completely dead but the mechanical 
stuff was just fine.  He cut the stepper motor drive section off of the CPU 
card - he now had a stepper controlled X-Y-Z mill, complete with stepper 
drivers.  I wrote a simple PIC stepper controller (12c508) that takes in 
step and direction commands and generates the step sequences for the 
stepper drivers.

Ryan spent a couple of months in his spare time learning how make the 
system work.  He now uses a somewhat modified version of Kevin Carroll's 
Stepster as the G-code interpreter - its what drives the stepper controllers.

It takes Ryan about 40 minutes to process the plot file from my CAD package 
into G-code suitable for feeding to Stepster.  I don't know the exact steps 
involved: I do know that he uses Corel Draw, Adobe Photoshop, and Desk-NC 
in the process.  I'll document the procedure sometime soon and post it.

The milling bits are standard high speed steel engraving machine bits.  For 
milling PC boards, Ryan has the bits ground to what the re-grinder calls a 
.005" flat.  That means the tapered bit does not come to a point but 
instead has a .005" wide flat surface at the very tip of the 
bit.  Apparently any finer than that results in a tip that is too fragile.

Milling a board takes anywhere from 20 minutes to an hour, depending upon 
the size of the board and how much copper has to be removed.  The boards 
turn out OK - Ryan can get a trace between 2 IC pads reliably.  The sides 
of the traces are somewhat jagged - not perfectly smooth like the photo 
process.  The isolation path is probably about 0.015" or so - it means you 
have to take a little care when soldering to make sure you do not bridge 
the gap.

The boards are drilled after the isolation paths have been milled.  Ryan 
simply changes the milling bit to a drill bit, loads the (edited) .NCD 
file, hits GO, and walks away.  When all the holes for that size have been 
milled, he changes the bit to the next size and the machine drills 
those.  The PC board is not moved between any of those steps so 
registration remains perfect.

We now rarely use the photo process!  The milled boards don't look quite as 
nice as the photo etched boards but it takes less time to get a board made 
and the milled board takes only about 1/4 of the total man-hours of actual 
labor.

I've been planning on doing a bit of a write-up on the whole process - lack 
of time has stopped me.  But I'll try to get some pictures of actual boards 
made on the machine - soon.

dwayne

PS - Ryan has done one hell of a great job in turning a dead engraver into 
a wonderfully useful CNC tool.  He was the driving force behind this - I 
assisted him with some things but Ryan deserves all the credit.  I want to 
be clear on this - it was his idea, he found and purchased the dead 
engraving machine, he figured out how to make the software available to him 
do the job he needed.

dwayne


Dwayne Reid   <dwayner@...>
Trinity Electronics Systems Ltd    Edmonton, AB, CANADA
(780) 489-3199 voice          (780) 487-6397 fax

Celebrating 18 years of Engineering Innovation (1984 - 2002)
  .-.   .-.   .-.   .-.   .-.   .-.   .-.   .-.   .-.   .-
     `-'   `-'   `-'   `-'   `-'   `-'   `-'   `-'   `-'
Do NOT send unsolicited commercial email to this email address.
This message neither grants consent to receive unsolicited
commercial email nor is intended to solicit commercial email.