Say, if transistors weren’t discovered yet, would we have modern computers with lots of vaccuum tubes?
I remember reading a SF story years ago where the computers were run entirely by vaccuum tubes. There was a large crew assigned to just continuously replace tubes and also an enormous air conditioning system for cooling. I’m sure it would be a monster!
I thought you were going to ask how many tubes were in old computers!
There are a couple of problems with tubes for modern computers.
-
The computers would be very slow, not only because of the slow response time of the vaccuum tubes, but because of the long length of the wires between them. The delay due to the speed of light, IIRC, is about a nanosecond per foot. a lot of the work of designing a microprocessor is placing blocks to minimize “long” wires - where long is inches. Routing signals to gates feet away would be disastrous.
-
You’re quite right about the heat. But volume is also a problem. Designs with hundreds of millions of transistors are now common. I leave it as an exercise for the reader to compute the volume of a processor with 100,000,000 tubes.
-
Reliability is the most important thing. The mean time to failure of a single transistor is extremely long. There are various mechanisms by which they fail, and one can compute the expected failure rate over time - and the parameters of new processes are adjusted to keep this under 100 fails per billion hours of operation. Given that the mean time to failure of tubes is always going to be a lot shorter, you can compute the number of tubes above which the processor will basically always be down. I suspect the number is smaller than the number of transistors in the chip that plays a song in a greeting card.
So, the answer is no.
Not only would it be, it was!
The old SAGE system was a vacuum-tube computer built to monitor air defences back in the 1950s. Worse, it was completelt redundant, so that if a tube blew they could switch to the other system while it was fixed. The tubes were held in arrys big enough to allow air to circulate and technicians to walk through. It was monstrously huge. Until it closed down a decade ago, the Computer Museum in Boston had erected a portion of it on one floor.
More on SAGE:
http://www.computerhistory.org/events/lectures/sage_05191998/sage_xscript.shtml
SDomwe pix:
More info and pix:
55,000 vacuum tubes in each of the redundant systems.
There’s a good, representative drawing about halfway down the page of the SAGE computer, “more than 200,000 vacuum tubes @ 1,000,000 watts,” showing the multiple diesel generators and banks of air-conditioning units needed to operate. It says “Today it can be replaced by a faster $7.95 throw-away calculator.”
Note that many commercial computers before the transistor was invented used relays. For the reasons alluded to, systems involving large numbers of vacuum tubes were not practical for most applications. Electromechanical systems were faster, more reliable and more compact than you might have guessed.
If we still used transistors for computers, nobody would be disputing that man was affecting global warming.
On the other hand, those who had stock in liquid nitrogen firms would be cleaning up. 
We aren’t? Sure could have fooled our technology group.
yabob, could you give me a cite for a computer implemented with physical relays, not relay like structures implemented by vaccuum tubes? I guess phone switches, if considered as computers, might be one example. IIRC correctly Shannon’s MS thesis inventing digital logic used relays, but he didn’t talk about the actual implementation.
Here’s one:
http://www.ipsj.or.jp/katsudou/museum/computer/0100_e.html
Note that it was used until about 1965. And I was going to bring up phone switches. The link to the ETL Mark I from that account says:
You should be able to find further examples of relay-based computers built in the 1950’s.
Thanks for the link! I’ve actually read nothing about early Japanese computers. Interesting.
My copy of Bell and Newell vanished a long time ago. Looking at the edition with Dan Siewiorek as first author reminded me of the Manchester Mark 1, which was down with the Williams tube, which basically used a CRT as a storage medium.
Most computers, though, did use vaccuum tubes - the Illiac 2 and the LGP 30 for two. Not all transistorized computers were faster than vaccuum tube ones - I used the LGP-21, the transistor version of the LGP-30, which was slower.
I remember reading some early science fiction where big computers were built in space so the vacuum tubes wouldn’t need to be tubes, just exposed to space. Maybe it was Doc Smith…
Anyway, you can find all sorts of anachronistic stuff in old science fiction. I’m think particularly of Heinlein’s “Starman Jones” where starships is piloted between wormholes using mechanical computers and hard-copy log tables, and a major plot point is that the ship is trapped when someone steals and hides the books of log tables.
My mind is agog! SAGE was actually built! I hope it made a difference.
Question: Snip “Some of the missiles that operated under SAGE had a serious social problem: they tended to have inadvertent erections at inappropriate times.”
What the Hell? How can a missle have a social problem? Also:
“Weighed 250 tons, took up twenty thousand square feet of space, and was delivered in eighteen large vans. It had 50,000 vacuum tubes, more than 150 CRT monitors, needed more than a million watts of power, and took up two stories of a building.”
This is indeed impressive. I haven’t had time to check all the links above, but are there any pictures of it?
Thanks for all your responses.
Hey, don’t I say in posts #4 and #5 that there are pictures?
Sorry, missed that link. SAGE is indeed impressive. Thanks for the link.
One of my favorites was called “Cosmic Engineers” I think, by a real rocket scientist. It was in Astounding in the early '50s, and the cover had a slide rule over a rocket. (I can look up the details tonight - I have that issue.) The story is all about the need to generate a perfect square wave for some reason.
I wonder if any slide rules have ever flown in space.
Anachronistic stuff like using cams as programming elements for space ships (In Heinlein’s “Rocketship Galileo” and others, and George O. Smith’s “Venus Equilateral” series), or using slide rules (Heinlein again, with “Slipstick Libby” – he used his head, of course, but the nickname implies others were using slide rules) > Arthur C. Clarke had a story where the flight computer broke down, and the astronauts improvised with home-made abacuses (abaci?)
i’ll bet somebody’s taken a slide rule up, but I’ll bet they didn’t plan to use it.
I’ve still got quite a collection of them myself.
I would be surprised if at least some of the Gemini or Apollo-era astronauts didn’t bring some kind of slide rule up with them, or there wasn’t a specialty slide rule (of circular or radial design) used for backup calculations. Most of the Mercury Seven were degreed engineers, and IIRC some of the later groups of astronauts have candidates with advanced technical degrees, so they’d all be familiar with the operation of a slide rule. However, I doubt there’s much you could actually do from the cockpit with a slide rule (other than to backcheck some simple calculations) if the guidence computer gave a serious heave-ho; they were tethered to ground control for interpretation of guidence and astrogation information.
Let me know if you ever want to liberate some for cash.
Stranger
Fascinating links - I particularly like this picture. My father loves to ridicule the ‘flashing-lights consoles’ of old sci-fi films and programmes. That photo shows one in real life. (And in the context of this thread, I guess the flashing lights, or lack of flashing, could help pinpoint dud tubes?)
I think you’ll find you’re misremembering something.
The only story called “Cosmic Engineers” I know of was in Astounding, all right, but it was serialized in 1939 and none of the issues had a slide rule on the cover. It was by Clifford Simak, who was a life-long newspaperman. I don’t think he had any kind of degree, much less a technical one.
Aha, Found it. The story was “Galactic Gadgeteers” [!] by [G.] Harry Stine in the May 1951 Astounding. Here’s the cover. High on the list of “they don’t make 'em like that any more.”
He was indeed a rocket engineer, working at White Sands and popularizing model rocketry.