The story: I wanted a +D, but how to get one? I could cruise eBay until I found one,
or, I could try to build one, and so improving my knowledge of electronics in
the process. After all, I had the circuit diagram, courtesy of WOS, and it didnīt
look that complicated. Ahem. First I had to locate the required components.
There are two tricky chips on the +D, the WDC1772 floppy disk controller, and the Pal20L8.
Both of them are no longer produced, however the WDC1772 can usually be obtained from dead
Atari STs, or from Atari specialist shops in the internet. Note, usually the 1770 may be used instead
of the 1772, they differ (AFAIK) only in their stepping speeds. However the +D insists on a 1772.
The Pal proved to be a bigger problem.
I managed to get some of them from a US radio ham who had a few in his spares box, and
is prepared to pass them on as long as the project you want them for is interesting enough!
It should be possible to substitute the common Gal20V8 instead of the Pal, but I havenīt tried
I wanted to get my design as close to the original as possible, so a decision was made to produce
a board with the same footprint and feature set as the original, but having a new layout.
With the help of a colleague (Harald Brunig), we got a layout together using the
superb Protel software. The layout we devised is
not a copy, but a new layout which duplicates the function of the original while attempting
to keep the components in their original positions.
After this we got two prototype PCBs made, using the services of
PCBPool. You only have to send them the Gerber and drill files from your CAD software, and they produce a ready to solder PCB, drilled, cut to size, all vias through plated, etc. Better than mucking around with etching chemicals.
I then soldered the components in place, blew an eprom containing the +D eprom image, and also a Pal
with the contents of the Alice file. I connected my +D clone to a Speccy, and the damn thing didnīt even
power up properly. Bugger. I contacted Peter Rennefeld of the SPC who kindly gave me an original +D on
loan. I started to compare the two circuits and it became apparent that the circuit diagram on WOS was
incorrect in a number of places. Most importantly: the chip select line of the RAM and EPROM chips
were swapped, and the address lines of the FDC were connected to A0 and A1, they should have been attached to A3 and A4 instead. After
correcting these errors I had a working +D clone! A bit of aluminium bending and spray painting later (thanks also to Peter Schunke!) and I had a smart looking case to finish the whole thing off.
Partially finished +D board
Finished +D board
The completed article, original and clone, can you tell the difference?