New subject: End of restoration of the early French computer SEREL ODP-505(germanium CPU) Same message, but cleaned!

30 Mar 2008

| > ECN: Engineering change notice. Indicates revision of PCB.| | Thanks, I 
did not realise you had already translated it to English.| >| > The ODP-505 
is a pure binary machine, not designed for arithmetics,| > statistical 
purposes or BCD computations.It is a real time (!!!)  | > computer.| > Get 
some datas (voltages, frequencies, switch states, motor speed,| > 
temperature...) compare to thresholds or values, and speed up /  | > slow 
down| > motor, or move a cadmium bar (?) in a nuclear reactor...| | I see. 
The Elliott machines were the same (but not the ICT1301),  | keeping the gun 
of a tank level and aimed in the right direction,  | accepting data from 
gyros and inertial sensors and tracking the  | position and orientation of 
an aircraft or ship, cross checking with  | compasses and other navigational 
systems, analysing sonar data and  | tracking enemy submarines etc. I think 
we are right. I never worked with this machine, so i do expectations (with 
my good knowlegde of micro-processing) | By the time of the 920ATC we even 
used  | a special real time programming language called Coral66, based on  | 
Algol60 syntax but with cut down semantics to give highly efficient  | run 
times. Before that everything was done in assembler (called  | S.I.R.) or 
machine code.I have no information about software. I don't know if they had 
even an assembleror if they do the job with a pencil and a rubber. (as i 
explain in my IBM 650 section)Have a look there: 
http://pichotjm.free.fr/IBM650/fileUS.html| >> | The architecture of your 
machine reminds me of the first machine I| >> | was allowed to operate. It 
was an 18 bit binary machine, it had  | >> 8k of| >> | memory built in, 
expandable with external modules and was made  | >> by the| >> | Airborne 
Computing Division (ACD) of Elliotts, it was an Elliott  | >> 920B| >> | and 
was a compact, flyable version of the commercial Elliott 903.| >> | There 
had been an earlier model 920A which I think would have been| >> | Germanium 
and was roughly the size and shape of a carpenters work| >> | bench. This 
may have been contemporary with your earlier machine I| >> | think. 
Unfortunately I never saw one in the flesh, but there was a| >> | bench in 
the computer room which I found out later, was the empty| >> | chassis of a 
920A. Behind the bench was a large panel full of| >> | electrical 'chocolate 
strip' connectors where the analogue and| >> | digital input and output 
signals of the 920A could be connected up.| >> | Apparently some of the 
analogue outputs had been connected up to an| >> | oscilloscope to provide a 
visual display unit, though it used a  | >> fair| >> | bit of processor time 
to keep it refreshed, even with the long| >> | persistence phosphor of an 
oscilloscope. The panel had been covered| >> | with board with pegs to hold 
mylar paper tapes and until the day it| >> | was scrapped I had not seen 
what was behind it.| | > Very interesting!There are some video connector in 
the earlier  | > machine and| > the Serel companywas specialized in high 
tech video solutions.| | Possible application for computers in video back 
then were  | superimposing images and standards conversion and just possibly 
| titling and graphics.You seem to have not browsed completly my website. I 
show about 200 beautiful photos describing these kinds of applications. We 
did superimposing images and standards conversion for ship, nuclear 
plants,helicopters test, meteo ... have a look there 
http://pichotjm.free.fr/Techno73/Techno73.htmlI have not translated the 
pages (need time) but if you click on every linksyou will see these 
beautiful pictures i have taken in the 'Le Bourget' aeronautic international 
fair in 1973.I have no returns about these publishings, so i am not very 
encoraged to translate them... If you want I will try to take time.I just 
verify it's translated! 
http://pichotjm.free.fr/Techno73/Techno73US.html(probably badly, but 
sufficient) | > I have read| > somewhere in docs, thatthis computer have a 
screen output...| | If it has two or preferably three digital to analogue 
converters then  | a vector display is fairly easy, but maybe it was more 
complex than  | that, and you only need one digital to analogue converter 
for a  | raster display, provided you have enough CPU power to both drive 
the  | display in real time and do any processing required as well.| | I 
should explain what I said about two or three A/D converter, you  | have one 
to drive the X plate and one for the Y, ideally one for the  | brightness 
(Z), but with a fast enough converter, you can deflect the  | beam so fast 
so it will not mark the screen significantly and so no  | need to turn the 
beam off before moving to the next item to be  | displayed.I have a picture 
of the display terminal for the earlier one: 
http://pichotjm.free.fr/Serel/Notices/Serel%20computer%20dept/img_3673.html… 
following english page indicates 'Oscilloscopic characters display' | > I 
have found,| > last week a small notice describing microprogramming on| > 
ODP-505.http://pichotjm.free.fr/Serel/ODP505/MicroProgrammation/ | > 
MicroProgrammation.html(doc| > found in photomultiplier doc!)I have found 
commercial document  | > describing| > displays and analog memories...I 
don't know the date... (1970?) I  | > have to| > study these documents...| | 
I had a look but my understanding of French is terrible, and  | technical 
details in French are even more difficult.Do you know how I am tired after a 
translation to publish in english? ...| >> I have a| >> earlier machine from 
the same company SEREL, named OA-1001. Built in| >> 1959/1960.I need to 
restore it. It lays on the floor (horizontal| >> position)... The blue one| 
...
 >
here:http://pichotjm.free.fr/Serel/Photos/Photos.htmlI will start| >> 
restoring| >> next month (with the Sun!)| >| > | What is involved in the

restoration? Do you intend to make it work,| > | this would be very hard 
without the schematics.| | | > As you know, i am alone,| > here. I want to 
make an esthetic restoration:Photos and notes,  | > dismantle,| > photos and 
notes, wash, dry, fix the rust, protect with Rustol,  | > photos an| > 
reassemble. I hope to be capable to do that... may be one year, may be| > 
two...| |   As I understand it, that is what the big museums call 
preservation.  | There has been a lot of discussion of the relative merits 
of  | preservation and restoration. Restoration is what they call it when  | 
we return a computer to running condition, preferably so it can be  | 
demonstrated. There is some truth that restoring a computer always  | 
destroys some of the original and that preservation is, to the  | purist, a 
better thing to do. I know... but i can use a bad english word... It's why i 
precised esthetic!| As I have two 1301s, I have the luxury  | of being able 
to restore one and preserve the other, but restoration  | is a lot more 
enjoyable to me.May be, but i know the time needed. And I have many projects 
for the future.I don't used to watch back. I prefer future [i intend to live 
in!  ;-) ]I do a lot of program in television, compression, splitting, ECC, 
FTP, transmission...(I have a son living in the USA, so he can watch french 
TV in broadcast quality... chutt!)| The other thing of course is simulation. 
This can be done at various  | levels, simple simulation of the instruction 
set, more complex  | simulation at gate level and in theory the ultimate 
would be to have  | a 3D model of the computer you could walk around and 
inspect the  | inside of, and attach a virtual oscilloscope to and watch the 
| analogue signal levels and be able to do hardware patches and have  | 
simulated hardware faults and diagnose and repair them whilst the  | lights 
and switches all worked like the real thing and the sounds and  | 3D models 
of the peripherals would also be like the real thing. Maybe  | I'd leave out 
the sound of a card deck being mangled and thrown into  | the air by the 
card reader though.Yes i know all these kinds of technics... and can do the 
work (not 3D)It would probably be better to spend his time in creating a 
special software to do reverse engineering, (and more exciting for me) with 
a mixt of taking photos, and piloting schematics and routing 
simultaneously...| >| > May be some reverse engineering to get 2 or 3 
schematics. (need one  | > week for| > a board! I have 2 boards in 
process...)JM Pichot| | Good, Yes, but i have no courage to finish the 
task...| are you able to identify the function of each type of board  | 
(such as 'And' gate, Flip-Flop etc)Probably, with the help of my lectures of 
the sixties...| If not I could perhaps compare with my 1301 schematics and 
see if  | they have any similarities. How difficult is it to identify the 
power  | supply connections?For the small computer, i have to analyse 
schematics (to be done!)For the earlier, cables are preserved and connected. 
I have power supplies, so it is allways possible to trace wires... (not 
really exciting!)| Are you planning on doing schematics for the board 
interconnections?  Not today! or this year. First i plan to preserve the 
machine and publishphotos.| I imagine this would be very difficult, so 
probably best not to try.There is no national interest, today, in France, to 
preserve or studyold machines... It's a huge task. I can not do that... 
alone.So i will wait (or my sons..) that French speak, make meetings, 
makereports, ask for money... may be in 20 years. Museums are for paintings, 
ceramics, statues...I have to be more positive!!!By the way, if you are not 
still fed up, i spent last days in publishingan analysis about a display 
terminal designed in 1969, which use 32 TV delay lines as memory!  Have a 
look there: http://pichotjm.free.fr/DisplayDL/DisplayDLus.htmlNote. I have 
not a perfect english, so i can misuse words. I never intend to be 
agressive, unpolite or to hurt any body. I wish you a good 
evening.Jean-Marie PICHOT 

End of restoration of the early French computer SEREL ODP-505 (germanium CPU)