Ken has just posted part I of his blog about this tube circuit: www.righto.com/2018/01/examining-1954-ibm-mainframes-pluggable.html And part II: www.righto.com/2018/01/ibm-mainframe-tube-module-part-ii.html
You could imagine the men working on this, imagining all the amazing things they could do with electrical circuitry. But they come across a wall of historic proportion, the need for a clean signal. Once they realized they could create and guarantee that signal they knew they could make a device of rapidly flawless calculation. The true birth of the digital age.
I am always amazed to see these old IBM machines chugging along. The fact that these computers were so expensive, and companies managed to justify the total cost of ownership implies that they had to make money using them. I just can't imagine how... I am guessing data storage was one application. So , maybe one of these machines could replace 10 - 20 people doing the same job?
Data processing and computing was the big thing. Storage was at first very limited and expensive, though it's ubiquitous now. Mass storage was done first on punch cards or punch tape, then magnetic tape as seen in the photo, and later hard drives and floppies.
By the condition this may have been one I liberated from the from the Florida State University 704 about 12-15 years ago before it succumbed to the scrap man. It was in very poor storage conditions for over 30 years I still have a handful of modules and tubes left.
I remember buying some of those assemblies back in the mid 60's. Only paid a couple dollars for them and came with the tubes. Came from discount obsolete part dealers in New York.
Very nice little presentation on the 5 channel tube switch de-bouncer. I guess it took 21 of these for a standard 101 button keyboard. The power requirements just so you could run the inputs for the keyboard was probably pretty significant, much less the rest of the computer and it's accessories, like printers, monitors and indicators. Thanks for sharing just a bit of our past with all of us, it was a nice Christmas present. Wish I could have been there in person, I love this kinda of fun. Work in an electronics repair lab repairing aircraft test sets. I get to reverse engineer and build test jigs and adapters to equipment that is no longer manufactured and is unique, it's a lot of fun. I got a laugh at the equipment age difference itself, between the counters, power supplies and the oscilloscope. If it had not been for the oscope this could have been an engineering retro video from the 80's. LOL Anyways thanks again for the fun. Belated Merry Christmas and Happy New year. Best Wishes n Blessings. Keith noneya
I cut my teeth learning about thermionic emission & hot cathodes in electronics at secondary school in the late 70’s then on to PN junctions etc...but you can’t beat the glow of the heater working....it’s always encouraging to see in an old circuit you’re working on
A lot better than my TI-30 calculator, that I used early in college, in the late 70s. Worked fine for about 2 years, then it bounced so much as to be unusable.
I have fallen in love with your channel and watched many videos, but I noticed my subscribe button was still off. Not anymore! This is absolutely amazing. Back access to a computer science museum, hooked forever. Thank you and a happy new year!
One of the many things I enjoyed about this video was that the background hum was eerily similar to the soundtrack of Forbidden Planet when they surveyed the Krell device. I suspect similar levels of power were being pushed through the systems :-)
Regarding the 'tube tester' bit: Growing up in St. Petersburg {Florida}, I remember a tube tester sitting near the entrance of the ECKERD DRUGS I used to shop at. I never had any reason to use it, but it was there. *_"Seems a thousand centuries ago."_* *-- APOCALYPSE NOW [1979]*
Remember, the programmable computer, starting with the special-purpose British "Bomb" to decipher ENIGMA codes, to the first small-word-length quantum computers and neural network brain simulators (the ones that are now beating everything in GO and CHESS) is less than 80 years, one slightly long human lifetime. This seems very strange to me, seeing that it took centuries for most other advances to work themselves up to such a high level of sophistication. Are we getting "help"?
Jeez. Makes me feel lucky that all I need now is a 555 or maybe a couple of transistors, some resistors and a couple of caps for a one shot circuit to debounce switch inputs.
It was done this way for decades, including with discrete semiconductor devices and then ICs. Then somebody figured out how to do it entirely in software (you read the input many times in quick succession and accept the predominant state).
Mike Cowen is it really that people only later "figured out" how to do it in software, or that there was eventually so much extra computing power that doing this in software was practical? I doubt there were enough spare cycles back then to deal with debouncing.
Bobby M You make an interesting observation. Typical scan rates for a perceived immediate input are only about 100 Hz, so that wouldn't steal that many cycles. I don't have any smoking gun proof, but as hobby projects came out for single chip CPUs, despite having a typical 1 MHz or better clock (plenty of hobbyist debounce power), they were still using hardware debounce circuits into the 90s. I can only presume that was what the designer was familiar with, and stuck with it. Old ways die hard. I was also working in the arcade industry during that period, and even they hung on to the very last second.
menoob: it also has to do with stability, price per object used, behaviour of the element, elements in the circuit.. afaik.. and i think.. that the flipflops, latches by tubes didn't infere with the program cycles (slippery moment)
Yes it depends on your early influences. The first computers i digged deep into was ZX-80 and ZX81. So for my own projects during the 80s and 90s, a simple keyboard scanning and "debouncing" in software was often the "self evident" solution.
Yes, actually billions of times. But for me, discrete diod-resistor gates and resistor-transistor gates/latches were the golden middle way, though. :) (7400-series TTL and 4000-series CMOS too, for that matter.)
Seeing those power supplies not aligned in a proper way is eye hurting. Why did you do that Marc!? Loved the video anyway, thank you and keep up the great content!
Scientist: "In statistical hypothesis testing, a result has statistical significance when it is very unlikely to have occurred given the null hypothesis.More precisely, a study's defined significance level, denoted by alpha, is the probability of the study rejecting the null hypothesis, given that the null hypothesis was assumed to be true; and the p-value of a result, p, is the probability of obtaining a result at least as extreme, given that the null hypothesis is true. The result is statistically significant, by the standards of the study, when p
That's a really big circuit for just debouncing one button on the console. Hopefully, that one module is used to debounce multiple buttons. Because if you had to have one of those modules for each button on the console. That would take up a hell of a lot of space. BTW: why such a complex circuit for debouncing? Won't a simple RC filter ran to a single triode do the job?
I'm a amateur diy electronics guy. Can someone explain the reasoning behind all the various voltage requirement. Does that kind of voltage regulation come years later?
I haven't seen the schematic but it could very well be that it is basically an RC followed by a schmitt trigger to produce the digital output. Edit: I got curious and followed the link provided in the description and discovered that yes, indeed, my suspicion was correct! :D
"osmelloscope" eh ? Someone has been watching the AvE channel... My first TI30 LED calculator (with its ultra-cheap keyboard and bad repeating contacts) sure could have used such a debouncer circuit. But it wouldn't have been convenient... lol !
Never heard that old Swedish proverb. But just wanted to make a small correction; "RIKTIGA DATORER LYSER I MÖRKRET". Observe the Ö instead of the Ø. Ø is not used in Swedish but in our northern neighbours Norway and Denmark with more or less the same pronunciation. Not a big deal ofc but perhaps it interest someone.
Very sorry about mangling your language. My Swedish is about as good as my Google translator, and I couldn't find how to do the Ö on my US keyboard. We French also get all uppity with our many accentuated letters...
How much wattage was there just for that one debounce circuit? It is no wonder these things were power hungry beasts that needed to be in air conditioned rooms. The fans on at least one of the power supplies was rather loud. It got nice and quiet again once everything was turned off.
it was under 10W per debounce circuit - the entire module consumed roughly 55W with 5 debouncers and 5 cathode followers. The debouncer used two triodes while the cathode follower is implemented on a single triode; assume 2/3 of the energy is used for debouncing. Heating the filaments constitutes about half the total consumption.
I was thinking about how just these few switch debounce circuits required 55W. Add in however many other debounce circuits, then add in what the main system needed. It is no wonder these machines were in separate rooms with heavy duty air conditioning and massively thick power cables running to the cabinets.
Why was such a module necessary? Why not a simple capacitor? Did vacuum tubes not have a clean "transition voltage" that changed computer state from 0/1 or vice versa (with semiconductors generally around vcc/2 or vcc/3). Did tubes need an exceptional amount of input buffering?
A simple RC filtering would not work. You need to follow it with a comparator with a Schmidt trigger action. So this is exactly what this is: an RC filter followed by a Schmidt trigger.
Clean high speed switching was required, and utilized by electron tubes operating at various switching speeds.the signals need to be clean and precise.
I had one of these modules at one time! I never knew what it was until now. Unfortunately I used it for parts and threw the rest away when cleaning out one day.
Interesting piece of retro tech to power up. sure shows how long tech has come since 1955. I heard that computers from that era took 100 foot long rooms.
To a none technical person what does it do? Well except for lightning the bulbs? Converting a keypress into.... a voltage or have the voltage conversion into a binary hold by bulbs/transistors to signify the character pressed? If so are there different voltages for different keypresses, it is not binary from keyboard to the machine?
Well thinking abit about it XOR, OR, AND maybe only usable in a binary context? Audible add where each pulse represent a wordlength one would have to construct some other sort of electronic filters adding words rather then binaries?
Is a ternary machine still discrete? Are machines using more then two states by default analog what signifies a "digital computer" when does it become digital, i mean that it hold other then binary values do not make it analog? Is it the clock that make it digital? I read about continous computation, a computation leveled adding words would not be continous, so i guess it is a digital computation afterall? I read about physical quantities mixed they call it continous computation, but anything using a wordlength can't be continous, but that does not mean it is by default digital, so what make it a digital computation?
Buttons and switches aren't very reliable - the contacts inside can bounce around when you mash them, causing one press to seem like multiple presses to the computer. This circuit just takes a noisy mess from a button press and produces one single stable pulse for the computer. Nothing to do with voltage level changing or key mapping Did that answer your question? I'm bad at explaining things :)
That wouldn't be done in this circuit, but every button would probably be on a key matrix and another module interprets them I'm not sure how they're interpreted in a tube computer like this, but for a modern computer just look up keyboard polling
So Cool! Thanks for this post ! Dig this old digital computer idea that used air to computer called the FLOWDAC Patent 3190554 . I suppose you would run it from a compressed air tank .
I could not find an exact count, but I found a refrence saying that the 704, which was a 36 bit machine, had 4000 such gates, so a few 10,000s of tubes?
Some resistors "fell off"? I take it that's a euphemism for "got borrowed for other purposes". Considering how sturdily components were connected together in those days, it's hard to imagine any of them literally falling off, even after many decades.
Could have been borrowed, but more likely that in handling over the years the resistors were banged against some object and came loose. no cutter marks, rather the connection points showed solder failure from strain
So the pigtails weren't wrapped around the terminals? I suppose it could happen in that case. Usually in valve-era equipment the wires are poked through holes and bent over, making them a pain to remove even if you wanted to!
The problem with bending the leads over around the posts, as I learned on another channel, is that it will stretch the leads, putting stress on the component so that the expansion and contraction from heating an cooling would eventually break the lead connection inside the component, leading to early failure. Ideally, you just solder the leads to the posts without wrapping around, that way the longevity of the device is reliant on your soldering skills rather than the heat expansion and shrinking of the components.
Id love to go back in time and talk with the guys who made this. This type of research must have been considered utter buffoonery. What made them want to make a machine like this. I know this is a small piece of a larger construct but its origins baffle me
Looking at pictures of the operator’s station, just by itself would require a cabinet full of these modules. With one of these in my house I could heat the place in the winter.
This just 'popped' into my head. {Yeah, that is how MOST THINGS occur to me.} A lot of people -- especially younger ones -- _POO-POO_ the Apollo Moon landings as being faked. {THEY WEREN'T.} One point I hear repeated is the poor quality of the TV signals sent back from the Moon, at least on the earlier missions. Those same people do not realize HOW FAR electronics have come since the early 20th century. At the time -- late 1960s to early 1970s -- there was simply NOTHING BETTER that could do the job under the required parameters: Light weight and portability, rugged {usable in space and on the Moon}, and relatively low power consumption. Just my 'two cents'.
@CuriousMarc >>> It has been a VERY LONG TIME since I looked at a powered-up vacuum tube. {Yeah, I am a BOOMER...🤭} Is the _'glow'_ produced by most -- or all -- vacuum tubes the product of _"incandescence"_ of some internal piece[s] glowing red hot? Or is it the product of something like the glow of an excited gas, like a neon bulb? This question just popped into my head, and I realized I either never knew or cannot remember.