EEVblog Electronics Community Forum

EEVblog => EEVblog Specific => Topic started by: EEVblog on December 07, 2018, 09:59:33 pm

Title: EEVblog #1153 - 1970's Programmable Teardown
Post by: EEVblog on December 07, 2018, 09:59:33 pm
Let's do some programming, early 1970's style with a teardown of the Canon Canola SX-100 Programmable Calculator

https://www.youtube.com/watch?v=eOWtrO04NOk (https://www.youtube.com/watch?v=eOWtrO04NOk)
Title: Re: EEVblog #1153 - 1970's Programmable Teardown
Post by: johnlsenchak on December 07, 2018, 11:37:55 pm
My educated  guess,   looking at the trigonometry  keyboard keys,   that  ancient  calculator  was  used  in  either  civil  engineering or  architecture design

Why would you use  Sine, Cosine  and  Tangent  in  accounting?  That's  more of  a early  engineering  calculator  with  crude  computer  circuity  inside
Title: Re: EEVblog #1153 - 1970's Programmable Teardown
Post by: WN1X on December 08, 2018, 03:12:58 pm
That was a very cool teardown! I love 70 and 80's era hardware  :-+
Title: Re: EEVblog #1153 - 1970's Programmable Teardown
Post by: SeanB on December 08, 2018, 04:45:57 pm
The grounding wires on the daughter board are there probably because the designer made such long ground traces, you can see it was run from the one socket on the back, all to the left, then down, then in the middle to the right then down and to the left again. Probably found during the first production run that the board had significant ground bounce on the IC's at the far end, and this would occasionally result in errors in either tape read or write.

Thus rather than respin the boards ( expensive parts then, you would have to remove all those expensive TTL IC's and solder into the new board along with the passives) they went with the production line rework and probably a service brief with associated field kit ( probably also a few other bodge repairs to fix other logic race conditions, which probably is half the added on parts on the main board) for the ones out in the wild. The underside of the board being all spotty shiny is simply old flux residue from the wave soldering, improperly cleaned off, or they had not cleaned it at all other than the cleaning applied to the edge connector masking tapes, or there was rework and they used SK10 spray on flux.

Worked on similar boards with only one decoupling capacitor per board, they worked fine with TTL with limited slew rate edges, though if that capacitor failed it was going to give weird failures, as the supply was decoupled with a 22uH potted inductor. I came across one board with a failed capacitor, and the previous guys had instead soldered a 100n dipped ceramic across every IC and superglued them to the package. Removing those would likely have resulted in damaging the PM3 packages, so they stayed there with the new capacitor on the board, I was not going to risk lifting those fragile traces with unsoldering those flatpacks, even with the right soldering tip in a Weller body.

Funny enough it now lives in a Solomon soldering iron head, I bought it myself and it left with me, and nobody got to "borrow" it without me being attached, bring your board and I would replace the flatpack no problem. I also had my own electric screwdriver, seeing as I had to remove 84 M4x16 allen setscrews to open the case. I wore out hex driver bits, and also a planetary gearbox set of gears. Got the powder metal replacement gears from Black and Decker easy enough, along with a near free very expensive ( at the time they were brand new on the market) variable speed reversible drill that just needed a stator rewound and putting the kit back together from the bag.
Title: Re: EEVblog #1153 - 1970's Programmable Teardown
Post by: SilverSolder on December 08, 2018, 05:23:32 pm
Guess it's easy to forget that the manufacturing precision just wasn't there, back in the 60's and 70's.  Same is true of cars - all the gaps and clearances had to be bigger back then.   There's a reason modern engines last 200K - 300K miles/Km, whereas a rebuild was very typical before even reaching 100K miles in the "good" old days!
Title: Re: EEVblog #1153 - 1970's Programmable Teardown
Post by: kenryan on December 08, 2018, 06:40:48 pm
You mentioned possibly doing an episode on using PROMs for state machines.

Back in my early career (circa 1993) I worked on a design that used a set of Cypress UVPROMs that were specifically designed for state machines - the CY7C259.  They included a 2Kx16 registered PROM array but you could internally feed data pins into address pins, so as to choose how many state machine outputs vs. available states and conditional inputs, with no external logic and running at a full 100MHz (nice speed for that era). 

The datasheet can be found on the web as 1994_Cypress_Programmable_Logic_Data_Book.pdf" (I'm newly registered, not sure yet of the link posting policy).

Love your channel, been watching it for years and finally compelled to post!

ken
Title: Re: EEVblog #1153 - 1970's Programmable Teardown
Post by: ggchab on December 10, 2018, 12:29:14 pm
It's really art! I like this  :)

(https://www.eevblog.com/forum/blog/eevblog-1153-1970s-programmable-teardown/?action=dlattach;attach=593098)