section layout
My Apple II Rev 0 Replica Project

PCB Layout

This was the most interesting and time consuming part of the replication process. The photo above shows shows a section of a PDF output of my layout overlaid over a photo of an actual REV 0 board.  I used photoshop to carefully align the replica layout with the original. This example gives an idea of how closely I've followed the original layout.  I've been lucky to obtain quality images of a rev 0 board from Geoff Harrison, which has greatly simplified the process of recreating this design.   In additional the famous "Red Book" contains a poor photocopy of the front a bare Rev 0 board, so that traces hidden by mounted components can be followed and replicated.    See the "construct your own replica" page for side by side examples of the finished replica and an actual rev 0.

My approach in producing the artwork was to use my Macintosh and an excellent program called Osmond PCB. There were several steps involved.

The first run of six boards have been fabricated and work perfectly.  Follow this link to view the bare PCB.  A higher resolution image is here.

Here is some interesting data that I have gathered in the process of laying out the board.

Mistakes found

Other than the 20/24K addressing bug and the issue with cassette output signal.  Apple's rev 0 copper layout was remarkably bug free.  However, I found a couple of small things during my layout process that might be considered small errors.

transistor packages

Unexpected differences between Rev 0 and Rev 1 motherboards

Expected differences between Rev 0 and Rev 1 motherboards

Errors in Schematics in Red Book!  

I found that these schematics are incomplete with the soft-5  (A2-8 and A2-11) inputs to various chips are not documented.  In other places connections to GND are not documented.

That Mysterious USER 1 Jumper

user 1 jumper location
I spent some time figuring out the location and use of the the fabled USER 1 Jumper.  The Apple II Reference Manual has the location wrong in two different pictures, and the errata on the Apple web site, doesn't help the matter.  This two vias are nearly obscured by slot 7, and can be seen in this photo.

The top via connects to USER1 which is connected to pin 39 on all the edge connectors.  The bottom via connects to H12 pin 6, which an enable input to the I/O address decoder.  This enable input is normally pulled up to +5 volts by one of the 3K resistors in the SIP near slot 7.  Connecting these two vias connects the USER1 signal from the bus to an enable input of the decoder at H12.  If driven low, I/O decoding is disabled, except for the on board ROMs.  I suppose this could be used to reuse some of the I/O space used by onboard I/O devices.  As far as I know this capability was never used by any production I/O card.  USER1 was reused on the Apple IIe as 65C02 SYNC and once again for M2SEL on the Apple IIgs.

Confusion About Rev 0 Statements in  Jim Sather's book "Understanding the Apple II"

cassette changes

Overall Impressions

First of all, digitally recreating a board of this size that that was laid out by hand is no simple process.  I estimate that I spent an average of 10 hours a week for 4 months on this.  I might have spent much more time than this, but I don't really want to know if that is the case.  Grant Stockly on his Altair Kit page recommends using a PCB scanning service for such tasks.  I didn't have a rev0 to start with, and if I did, I would have been reluctant to rip it apart to get a good scan.

However lengthy the task, it could have been far harder than it actually turned out to be.   It was quite remarkable how the vast majority of the 15 mil traces lie precisly on a 25 mil grid.  The majority of the traces that run diagonally,  run precisely at a 45 degree angle.  The chips are in precisely alignment and best of all, there were only two layers to contend with.  The square ends on some of the power and ground feeds were tricky to contend with in Osmond, but eventually I figured out a way to get it right.

I ran across a few odd things during the process.  A few of the pads on the bottom side of the board were smaller than normal to allow traces to slip between them.  Check out slots 0 & 6 and ICs D2 and A8 for examples of this.  One of the hardest portions of the recreation, was the silkscreen.  I can understand why Grant Stockly didn't recreate the original silkscreen on his Altair Kit, though he had a very poor silkscreen to start with, compared to the Apple  II.

Back to Mike's Hobby Home Page