Here are some of the machines I helped to develop whilst at Shandon.
All the pictures came from their web site
apart from their Varistain XY which didn't !
This
is a Hypercentre XP. It's used for processing biopsy samples. Based on
a Z80 running at 2.5MHz (whoosh!) That was all written in Z80 assembler.
The original source took 49 minutes to assemble on a Mostek Matrix running
Uniplus III+. When I got a 386/33 running DOS I could reassemble the lot
in under 4 minutes. It's a shame it took the management so long to buy
me a faster computer. I inherited about 200k of uncommented source and
had to add whole raft of bells and whistles that the marketing buffoons
dreamt up aswell as some serious and worthy updates. I hoped I never touch
the code again but I had to add 6 lines of code in 1998 for Y2K fixes.
Damn!
This
is a Cytospin 3. This one was great fun to do. Yet another revamp of existing
code but it needed tweeking to accomodate more marketing dross (4 digit
display rather than 3 digit) and to adapt it to run on a new family of
processor. Original based on an Hitachi HD63701, this one runs on an HD6303X
and all the code is in assembler. The centrifuge runs at upto 2000rpm but
the development machine could run with the lid open and the overspeed disabled.
The removable centrifuge head self-removes at about 5000rpm and does a
good impression of a flying saucer ! Co-developed with my old chum Alan
Bridge (who still works for them.)
A
Varistain V24-4. Yet another rehash of an existing product the V24-3 but
all new electronics and a few mechanical tweeks. This was entirely new
code again for the HD6303X again all in assembler. In the 7 years this
was in production we never received a single software bug report. A few
specials exist which support continuous operation or extended immersion
time for special tissue staining (bones and such things).
Cadenza - ugh! A brilliant machine that works really well but the software
was overtly silly. I had to rehash this crock several times. It was originally
written by a bloke who wanted to write his own UNIX. Hence if you know
which buttons to press you can get a Cadenza to dump its program and pop
a csh style shell. The individual real-time tasks appear as virtual files
on a UNIX-like filesystem. You can cd and ls all over the place. What is
worse parts of the scheduler were written by contractors in a special dialect
of C, COC or Contractor Obfuscated-C. Truly loasthsome to work
on but a cool thing to watch working.
The Varistain XY. A more expensive and marginally quicker version of the
V24-4. This exists because of a competing stainer called the Sakura DRS-60.
We got a DRS-60 and ripped it apart, figured out how it worked, spotted
all sorts of problems and produced a spec for a new DRS-60 beating staining
machine. The MD (CEO for those in the USA) just wanted a clone. So we started
to produce the clone. Then Sakura released the DRS-601 which included all
the improvements we wanted to put into our machine. Now the MD wanted a
DRS-601 clone which is what he'd have got but 6 months quicker if he'd
listened to us in the first place. Oh, and we had to make ours a mirror
image of the 601 so nobody would notice it was a copy! Based on the Cadenza
guts (68008 @ 8MHz) with the MMI by my mate Alan again and the motors and
real time bits by me (but in C not assembler).
The last major product I worked on for Shandon. The Pathcentre. All new
everythings ! This was the buisness. Compact electronics with an 80186
@ 20MHz, 1Mb Flash ROM , 256k SRAM, RS232, RS485 links. Software upload/download
to a remote PC. Remote control and diagnostics over a modem. Sexy LCD screen.
Once again Alan did the MMI and I did the gurgling, rotating, whirring
and other realtime bits with a chap called John Temple overseeing the whole
shooting match. By the time this was finished and the majority of
the hardware reliability and software sillies were sorted I'd had enough
of Shandon management especially with a short, arrogant prat known to all
as "Whoar Sid".
To be continued...