Date : Wed, 10 Nov 2010 18:45:05 +0000
From : list-a_cloud9.bbc-micro@... (Theo Markettos)
Subject: Electron Ferranti ULA reverse engineering progress
I got a chance to look at the Electron ULA Mark sent me under the microscope
today (with thanks to Sergei):
http://www.cl.cam.ac.uk/~atm26/acorn/electron/ula/
asic_1 is with a 10x objective, asic_2-4 are with 5x.
After a bit of staring we figured out that it's a bipolar technology (which
is obvious when you think about it), and that vias (holes drilled between
layers) can be filled with metal or N or P dopants. So you can't always see
one side of an NPN junction, but have to infer it from the context.
This article:
http://www.docstoc.com/docs/22805065/PROGRAMMABLE-LOGIC-DEVICES-(PLD)
(page 86, .doc page 26)
gives the layout of a Ferranti C-series cell. The layout on the above
doesn't correspond directly to my photo, but it looks topologically similar.
I haven't traced through all the parasitic bipolar transistors to see if
there's any difference. On the photo it looks like dark green is N-dopant,
purple is P-dopant and light green is metal.
So the bad news is this isn't a sea of gates, it's a sea of transistors.
That means it's not simply a case of drawing up a netlist and throwing it at
a synthesis tool - logical functions need to be drawn up from the sea of
resistors and transistors. Quite possibly reverse engineering tools won't
cope so well with that.
Some other references. Sorry if some of the stuff below is behind paywalls,
I can't tell the restrictions from a university machine.
This New Scientist article (7 Jan 1982):
http://books.google.co.uk/books?id=o2Woah0axxYC&pg=PA4&lpg=PA4
says either the Serial or Video ULAs used the Ferranti R-series platform.
Their cell structure is given in:
http://www.sciencedirect.com/science/article/B6V0X-47WTWY2-2C/2/a553fd812b073c97d40e804cb9759fa1
That's different in that it provides a quad-emitter current source and two
extra double-emitter transistors.
This article is by Peter Robinson who designed the video ULA:
http://portal.acm.org/citation.cfm?id=800740
and mentions C-series ULAs - but it's 1983.
Now I need to write some code to control the XY stage so I can take aligned
pictures...
Theo