EEVblog
Vložit
- čas přidán 6. 12. 2018
- Let's do some programming, early 1970's style with a teardown of the Canon Canola SX-100 Programmable Calculator
Forum: www.eevblog.com/forum/blog/eev...
EEVblog Main Web Site: www.eevblog.com
The 2nd EEVblog Channel: / eevblog2
Support the EEVblog through Patreon!
/ eevblog
AliExpress Affiliate: s.click.aliexpress.com/e/c2LRpe8g
Buy anything through that link and Dave gets a commission at no cost to you.
Stuff I recommend:
kit.com/EEVblog/
Donate With Bitcoin & Other Crypto Currencies!
www.eevblog.com/crypto-currency/
T-Shirts: teespring.com/stores/eevblog
Likecoin - Coins for Likes: likecoin.pro/@eevblog/dil9/hcq3 - Věda a technologie
I worked on these in 70s in Canada. The SX310 used mag card storage for programs, the SX320 used data cartridge. The 310 was the cheap version, the 320 the deluxe. They were not around very long. They were replaced by the AX-1 which used the same programming language and the BX-1 that used BASIC. I think there was even a floppy drive system available for the SX320.
Great video a nice trip down memory lane. Thanks.
Here in the US I worked with the SX series in my first job out of high school. I learned 6800 assembly on the AX-1/BX-1. I'm currently working with a guy in Portugal who has a BX-1 that mostly works. Printing on 30+ year old thermal paper looked good. We just got it to save programs to a hardware floppy disk emulator!
Ahh, I see you sprung for the Cockford-Ollie voice activated focus module. Good stuff.
i don't think so, dave's much too professional and white collar to be into a lowly working man's channel
Focus you fack!
Came to comment on AvE focus. See I am not the first.
My pop became an accountant in the 60s and only retired some five years ago. He used maybe 2 (3?) Canon adding machines all the way through. (He wore at least one out, as I recall.) The keys were dependable and had snappy rebound and the printers were pretty much indestructible. He would not switch for anything else. What we look at today as sloppy or bodgy was very much dependable at the time, and these large desktop-format Canon calculators were tool of choice for anyone that banged out numbers all day. I'll note that yours fired up just fine, which I found very pleasing indeed. Great teardown!
I was thinking of building one with Cherry MX Brown mechanical keys.
@@eggaweb Cherry MX Blues is where it's at.
Note that some of the "bodge" wires have dashed-line silk-screen, so not entirely bodged.
Some of the extra wiring looks as though it was added to reduce the resistances of traces on the board, instead of relying on plated-through holes to carry current from one side of the board to the other. This is important, as you don't know for certain just how thick the metal layer is. Still, they could always have just soldered pins into each of the through-holes, and the pins would have carried the lion's share of the current .....
Getting PCBs made was a slower and more expensive process in the 1970s! If you made a mistake, you would not just scrap the boards; you would modify the ones you had and alter the artwork in time for the next batch.
1980s mini-computers were often full of "bodge wires". They still worked well enough to be useful for designing the next generation of microcomputers .....
Yes. You got the date right. My Dad was an accountant and got one of these brand new in 1976. It was great entertainment for us. He would spend five minutes entering codes and numbers, then it would finally print out my name.
I should add that the purchase price was NZ$4000.00 !!!!!!
I used this programmable calculator to determine the size of the timbers and nail plates for house roof trusses in the mid 70's. It was provided and programmed by Pryda (New Zealand). Pryda manufactured the Gang Nail plates used in construction of timber roof trusses. It provided all the cutting list and angles for the timber cord and web components and each gang nail plate size at every join. It worked for all spans and roof and ceiling loadings (Concrete Tiles, Roofing Sheets and Plaster). Over time I learnt how to modify existing programs which was essentially an early form of the Basic programming language with up to about a thousand lines of code. This machine was the catalyst to my computer analyst career.
These were not office machines or personal calculators. I used to see these and others like this from names like HP and Tektronix when they were used in scientific and industry to run experiments, test equipment or some process machine like a low cost PLC of the day. What you are missing from that back panel is the I/O connectors, I expect everyone who bought one of these had a special cable that they made up to connect to the hardware they were automating.
PLC?
@@squatchhammer7215 Yep. It's a mini plc or Arduino of its day.
@@squatchhammer7215 Programmable Logic Controller :)
We used HP98000 series "calculators" at school. In the later 1970s, HP brought out the HP 95/97 desktop calculators. The HP97 was the desktop version of the HP67. Programs, and the magnetic strips, could go between the two interchangeably. The HP97 did have a built in thermal printer. The HP 9800 series higher end "calculators" were extremely capable, being able to be used as GPIB controllers, and a couple of the models could do linear algebra.
Back in the day, I tore apart an early 70s or maybe late 60s video terminal. The entire character set was formed by a board sized array of individual diodes. The screen memory was a huge delay line coiled up in a metal box.
Jim Steele Boise
@@krazykarl0 I think they were made by Conrac... at least that's what we called them... in polite company. :-)
Thank goodness they developed the pocket calculator relatively quickly because the pants with pockets that were tailored to hold that thing weren't very stylish.
REMINDER TO DAVE: The ROM based state machine sounds interesting, do a video on it ;-)
To give a little teaser. It is a complete not simplified Karnaugh map. So it is input mapped to output without the simplification. Which if you don't have cplds or fpgas (and they didn't back then) is the next best thing.
That sounds like something Ben Eater would do. He made a 7-segment decoder with an EPROM in his video named "Using an EEPROM to replace combinational logic".
czcams.com/video/BA12Z7gQ4P0/video.html
@@dentakuweb interesting video. It just doesn't write it as a sum of products, and use that as the memory for ORing and ANDing.
that is actually a lovely bit of gear
A Joy Forever, brings a teat to tied eyes! I love the craftsmanship.
Great teardown. Thanks!
Imagine being in college in the 70's and showing up to an exam with that bad boy.
Those kids would have a HARD time trying to smuggle one of those calculator during the surprise math exam. ;)
When you set that on your desk, you don't have to worry about it moving...
Both
At 10KG, that ain't going anywhere.
Bet you I'd still manage to lose it!
great video Dave!
Seriously a wonderful video topic, right up my street you bobby dazzler!
Gotta say it's pretty well equipped with scientific functions for the time, let alone the programmability which could lead to some pretty complex algorithms running on the thing.
REMINDER TO DAVE: The ROM based state machine sounds interesting, do a video on it ;-) (bis)
... My dad still uses his.
Edit:
To expand on this, he uses the SX300, (and still has spares for parts) he was an accountant, and being able to work with large numbers and have the written history of all the calculations was invaluable to him.
The information on the clear CPU feature is very interesting, as I remember him being very upset about that, as it took out 3 days worth of programming something.
I will check if he has a different manual for it to what you have.
> focus, you bastard
Heheh, we're all on a wavelength aren't we.
12:50 for those who hurry
It's slightly more advertiser friendly than AvE's line.
@@Anonymouspock I don't want to live in an advertiser friendly world!
If you have the perfect ROM, everything looks like a lookup table.
This was the "first and only", machine in my years - 1970's at Wesley College in Perth WA, now I'm 57, what a blast from the past, hope it works!
The model number is called Canon Canola SX100. Canola today means a type of cooking oil and SX100 is a model of Canon PowerShot camera.
What a thing of beauty.
One of your more interesting teardowns. : - )))
The blue wires aren't bodges - they're marked on the silkscreen!
I reckon they added them to the silkscreen instead of redesigning the board, still a bodge really
its possibly so they could do discreet testing on the board, powering up a test area ???
Thumbs up! What a beautiful calculator. And the thermal printer too.
I think that printer just needs some cleaning and lubrication and it will run fine again.
yeah you are right. I love vintage calculators
I love hand drawn documentation, done by real draftsman. My father used to do it, and I started out doing it.
It's fun to find at my job to find the old drafting drawings.
Very similar to the Olivetti Programma computer considered to be the very first '' personal computer '' of that era (1965 to 1979). The slot was for magnetic cardbords. On the Olivetti it was the size of a punchcard but one size was fully covered with ferric oxyde like the strip on credit cards. My dad who was an engineer had one of these before he bought his HP85. That Olivetti was used to compute armed concrete resistance to build TV signal relay towers. It was before the age of satelite TVs. When it cames in the late 70s, the HP85 was a huge revolution compared to the Olivetti Programma though. Indeed the Olivetti style computers are now seen as programmable calculators, but they where seen as real computers back then. It was possible to use all kind of peripherals such as paper tape or endless loop magnetic tapes. But the lack of alphabetical keyboard limited it to engineering and they are not known to the general public. With the exception of the Apollo fligh computer I guess.
The first “personal” fully programmable computer would have to be the LINC en.wikipedia.org/wiki/LINC , from the early 1960s. It can be considered “personal” because it was cheap enough to be approved as a departmental-level purchase.
Lawrence D’Oliveiro It was a mini-computer, not a micro-computer. The Olivetti Programma was the first commercial desktop individual computer available in Europe. It was the first time small engineering companies where able to provide their engineers with a individual computer they could use anytime they needed it rather than a dum terminal connected to an external mainframe located in a university or IBM facilty. As a result the computing was free of charges and not delayed.
@@srfrg9707 I didn’t say it was a “micro-computer”.
Lawrence D’Oliveiro Indeed, but you presented it as a personal computer. In fact mini computers came along with dum terminals and shared their CUP time among users. The very concept of a personal computer is that you are the king of the hill. No time sharing. You put it on you own desk and it's all yours. Olivetti just did that.
@@srfrg9707 Which is what it effectively was. No timesharing -- dedicated to a single user at once, hence “personal”.
LOVE those freehand curves of the tracks!! No 0°, 45°, 90° 'rubbish' of the modern-day CAD! I'd love it if I could still design modern PCBs like that!! Solder mask?! Pfft Shoulder flask! ;) all that steel!! No wonder its heavy!! And then... Along came miniaturisation!
The multi-chip "microprocessor" is just the old fashioned mini-computer era way of making a CPU. The whole board is literally the central processing unit. Actual ~$100000 computers like the PDP-11 were built in the same way.
Structural engineers used these things. My boss used a HP one in his early years in the late 70’s. I have seen structral calculations for precast floors from 1973 where the engineer listed the name of the computer on the documents: Olivetti. Can’t remember the model number though but it was very fancy pancy to use such a piece of equipment for structural engineering calculations in those days.
Make sure you lubricate the gears with canola oil.
Aww this is beautiful! I love how it displays seemingly random numbers as if it's displaying its calculations or something. Super aesthetic in my opinion.
Could you do us a favour and upload a small video on your second channel of this baby making all her lovely sounds? Just prop the mic nearby and write a simple program on it for 30 or so seconds. Would be certainly appreciated.
It wasn't too uncommon back then for the display to use a portion of the system RAM (since it was really expensive back then), which would produce patterns like that as it performs the calculation.
@@BlackEpyon Ahh, fascinating. Cheers for that bit of insight, mate!
I love old calculators. When I was a teenager back around 1971 I came across a box of circuit boards at the local electronics disposal store. Bought the lot for about $10 and discovered they were all the boards from a scientific desktop calculator, can't remember what brand it was. Bottom of the box had the case and keyboard and a wonderful 16 digit Nixie tube display board complete with tubes. I put it all together and it seemed to work except for a few weird faults. The boards were all diode logic with a few 14/16 pin DIP packages. The boards had hundreds (thousands?) of little glass diodes standing on end, soldered together on the top. After MANY hours messing around, I managed to get it working without fault. Turns out there were a few cracks on come circuit boards which when soldered fixed the problems. No idea why it was scrapped because the fault was not that hard to fix and a machine like that back in the early 70's would have been worth a small fortune. Soon after, I was falling behind in my high school exams and my dad tossed out (and sold to pawn brokers the more valuable items) all of my electronics parts and projects, including this calculator. I never forgave him for that!
I love that display
22:57
You should demo that on a breadboard if you don't have the original hardware. That would be an interesting follow up to Ben Eater's EEPROM demonstrations. He's got some great uploads explaining the basics of using an EEPROM as a 7 segment HEX display driver. It's part of his breadboard computer series.
I'm not aware of any other well explained EEPROM state machine demonstrations on YT that are beginner friendly. Given the current surge in popularity of vintage hardware you'll probably get quite a bit of interest with an upload like this ;)
-Jake
32:17
15 digit readout, now that's accuracy
OverUnity7734 Early Texas pocket scientific calculator had that too (rechargeable NiCd battery pack).
From before decoupling was invented by the look of it!!
Me: It's sexy. You, .3 seconds later: It's pretty sexy looking.
Interesting machine!
I used one of these when I was a graduate student. They were unpopular in the lab and shortly replaced with Wang 700s.
Don't know about this one. Wang made pretty good stuff, competing head to head in word processing against IBM. We had a Wang VS mini with COBOL and Fortran some years after this 700. Too bad it went out of business.
That is NumberWang
My Wang is a little dodgy.
My take on the blue wires with dashed line silkscreen markings was that there was no more room on the board layout for a given trace, so a jumper wire was necessary. Not a bodge, not an incomplete update.
ICBM launch facility would like their launch calculator back ;) LOL
Dave just had Christmas arrive early.
It's always interesting to see what people were using back in the day.
13:00 BRASS gears, not that plastic rubbish.
An HP rep brought one of the first production units to the Kansas state summer science honors camp when I was a freshman [9th grade] in 1970.
We wrote several programs that pushed it hard. By '75 these were pase' HP 65s and TI 41s were fully mag stripe programmable in your pocket. The only thing the desktops had by then was the printer.
I still have my TI30, TI58 and TI59. All in working order.
That looks similar to the Wang Numeric programmable that we had in computer lab in 1978. As I remember there were 4 of them in the classroom along with two Tandy TRS-80s for programming in MS-DOS and the main frame terminal that allowed us to execute FORTRAN and COBOL programs.
My mates and I used the SX-100 and another Canola 1614P for mathematics and physics in high school. c. 1974-75. I believe they were donated to NSW public schools by Canon.
I wrote programs for the SX300 model in high school around '74-'75. It had a card reader on the right side that read manual punch cards to load programs (you pushed out squares on the card to create the program to load).
I think the SX310 and 320 were later models that used the external drives you referred to.
A lovely machine.
I guess you cant compare pocket calculators with this thing. It would be like comparing a tablet with a notebook, or or a notebook with a desktop. :)
Nice teardown!
Bobby Dazzler... I can't help but think about "The Curse of Oak Island" when I hear this now
There was a late 1960s machine that I learned to program on, called the Olivetti Programma 101. It only performed arithmatic functions plus square root as I recall. This machine is much more sophisticated.
We had something that looked similiar to this, in school back in the late 1970's.
However, it had no printer, but had a small b/w CRT where the printer is.
If memory serves, it had 8 general purpose registers, with 20 (osh) places of accuracy.
The two slots to the right were insert and output of hand fed cards which held the actuial programs.
It was capable of only nesting loops to a depth of 3.
At the time, it was a wonder to use. :)
We had one of these at school to learn programming. Balgowlah Boys High in the late 70s.
My guess is the "tape reader/writer" uses a paper tape (or cards) with a magnetic stripe on the back, and the 24 column setting is for printing metadata on the front side.
That's about the same size as the HP-97 iirc. Pretty standard form factor for a deal calculator, and it probably cost about $300 or so in 1970s USDs. I had a TI-59 with the printer cradle, and that was a little bit smaller, but the print tape wasn't nearly as wide.
Neat! From the same period as the HP-90 or 9800 series I guess? Seems like a worthy competitor (save for not using RPN of course :D)
The Olivetti P101 was worth about $9000 in NZ in the early 70s. At that time a tech's wages were $80 a week.
Those "calcs" were intended to last, and the mag card readers allowed control data and transaction storage to ensure data integrity of the magnetic surface on the card. Pretty cool redundancy checking when there wasn't a full print of the transaction data.
The Canola looks much the same tech with its mag card reader.
I could see this being used in the physics lab component at a college in the 70's.
externally your Canon Canola SX-100 resembles in GIGANTIC SIZE the HP-97, both programmable calculators.
My dad said they had one of these in one of his classrooms when he was a kid.
Even after the advent of the pocket calculator most people in the business world still used adding machines. I am guessing if you were doing accounting of some sort you would find it easier to use that beast then a calc with a tiny screen. plus it does have the printer as you stated.
lesdmark indeed, my understanding was always that having a literal "paper trail" was a requirement in certain professions (accountants presumably).
This one was a few years before the scientific pocket calculator. This one was for engineering, science or economic cals. Don't need all that power for accounting, but why not if you have that on your desk.
This one was used for engineering work, and for people who had to do a lot of fast and complex computation.
I imagine that the magneetic program recorder used something a light cardstock sized like an IBM punchcard, except with a magnetic stripe.
Absolutely amazing! My brain hurts. I'm estimating it cost around $17,000.
No way. HP 9810, much mightier, much earlier, of much better build quality was $2500-$5000 in 1971. You can compare to SX100 data www.calcuseum.com/SCRAPBOOK/BONUS/72669/1.htm
Oh, programming in those days must have been rough. You'd have to have a big notebook with the whole program printed or written out, with layers of whiteout or the paper worn thin from being erased in spots.
I almost envy them, the thought of flipping through physical pages while reviewing the code...it sounds fascinating.
I used to do this in 2009-2010 with my graphing calculator (TI-83 with BASIC). It was really bad to type on, so it made more sense to write the whole program in a notebook, then enter it all in.
I used to do a lot on the TI-59, and we'd print out the code and/or output on the narrow thermal printer and tape it into a notebook, or photocopy it for engineering reports . Of course thermal being very unstable you couldn't read it even a few years later unless you had a photocopy.
Yep, pretty much!
When I looked at this thing I thought "hey, If i've been working in the office and didn't need portability I'd rather have a nice keyboard and display than dingy microscopic pocket calc". Then I looked at the display :D
Awesome but almost certainly coincidental Dr Strangelove reference "OPE"
Nice one. I wonder if the slider next to the display that you didn't touch is actually the brightness control.
MITs Altair is introduced in January of 1975. The Altair used the 8080. Probably not available, when this unit was designed. So it was their own, "in-house" "processor" design.
CPM was better than DOS then. Gary Kildall, RIP.
That's a beast, non of this pocket rubbish. Don't need to write your name on that puppy.
The glare on the CRO almost looks like a waveform
dave, please dont get rid of the ESD mat, it looks so sexy on video
Flying Spaghetti Monster damn, that is one sexy unit. It would fit perfectly under my tertiary monitor. Too bad I'm not seeing any for sale...
I wish they did manuals like that these days...
Did the manual give any clue as to what the mystery chip was? (IC6270)
i found this service document.
www.wass.net/manuals/Canon%20SX-300%20Service.pdf
Nice tour!
≈ 4m 10s: "And on the back of this beast, we've got a weird-ass mains connector ..." --- "mains" = main power??
Yes, that's a pretty common style; one I've seen lots of, I think even still today, mainly on printers. Why do you call it weird?
Now, keep in mind, I'm talking U.S. machines; which of course, this was.
Fred
3:48 Mankind lived in such ignorance before the coming of Swedish Rounding ...
Don't forget: a post "man on the moon" era calculator
I'd love to walk into an exam with one of these
Dean Schreiber More fun bringing (and using) an actual slide ruler. (With a modern device as backup).
My guess is one major advantage it had over the HP-65 and the other pocket programmables coming on the market was *price*. Those pocket calculators were hugely expensive - around a thousand dollars - about 4600 in today's currency. I can't find any price specifics, but I bet the desktops were significantly cheaper.
looks like the cash register the store that use to be open down the road had when i was 5 or 6
cool old tank
Full size keyboard probably allowed faster and more accurate data entry than a pocket calculator. If I had to perform masses of calculations I'd prefer to use this than a pocket calculator. All those 8 leg metal cans in the card reader board looked like 74HC series logic chips.
Did you ever figure out what the IC 6270 was? Was it a delay line?
I have a Canon Card LC-32 pocket calculator.
Most important question should be: "Will it run Doom?"
Bit-slice processor design?
Not in the rigorous sense. But some sort of "roll-your-own" processor, since it is programmable.
Calculators were often using 4-bit BCD internally, with numbers expressed in scientific notation; so it is probably implementating an ALU that works serially on 4-bit BCD "words".
The roller on the magnetic (bar strip) strip card reader disintegrated exactly the same way the one did on my HP 9100B. Seems stuff just doesn't last for 50 years anymore ...
rubber doesnt last forever. It drys up and perishes. youll have to replace the roller.
Pretty much everything on these things should last forever except the rubber, really
I had to replace the belts and roller on my HP-85 as well, sadly the roller is still a huge pain though
Frankly, I'm disappointed the caps didn't eject off the PCB.
The chip closest to the processor I suspect is the memory controller. (HD3542) The chips have the same layout as an Intel 8088.
"all the good stuff is made in japan" and then crack it open to see a nightmare
Back in the old days they often used ROMs as FPGAs before FPGAs were a thing. In the 80s PALs and PLAs were the thing, then they were superseded by FPGAs.
I am surprised it is fast. The program execution is slow, but this might be because the printer is enabled. The factorial, sin, cos, tan are very very fast. 4MHz is also pretty high clock, even if it is divided few times.
This 69! result was essentially instantenious. Even on modern calculators it can take seconds.
Its fast probably because everything is done in ROM or with 74 logic.
I wonder what'd happen if you fed that reader a paper fare card like a lot of US East Coast transit systems use. Bet it'd read and write just fine!