Rooms Full of Computers
Go watch some old World War II movie, something big and fancy with war-rooms and generals and all that; or, if you want to be specific, something like The Imitation Game. Now, you’ll notice from time to time, that they might end up in a room with some thin, pasty, white-coated guy, absolutely surrounded by giant banks of metal and gears. Those of you who have been to a museum or two will know that these were computers, in fact similar to the one I’m currently using. Although, perhaps that comparison is a bit disingenous, given that a modern toddler’s iPhone is vastly more powerful than the combined computational forces of both the Axis and Allies during World War II.
So how did we move from those gaping vaults to the tiny plastic squares of today? Well, nothing in particular. Or rather, a particular nothing: Vacuum tubes! Now, I’m not going to go too far in depth, given the sheer amount of genius that went into inventing these things, but I can give a brief explanation.
Vacuum tubes work a little bit like lightbulbs: you put a filament inside a vacuum inside a glass tube and run current through it, along with electrodes and ‘grids’ which controlled the flow, direction, and amplification of the current. These allowed for fine-tuned, variable control of the circuitry that drove computers at the time. What this led to was something called ENIAC: Electronic Numerical Integrator and Computer. It was basically used to calculate artillery paths, a task which it accomplished 2400x the speed a human could. In order to do this, of course, it needed memory and a way to compute. Now, it didn’t have any persistent memory–if you wanted to save things between runs, you’d have to punch a card and carry it until you wanted to boot up again (well, eventually they did get magnetic memory on it, but that was much later). Then why need memory, you might ask? Well, you can’t exactly go through calculus without remembering what number you’re on or what operation you’re performing. And yes, that is a type of memory.
ENIAC utilized ten-position (0,1,2…) ring counters to store digits, with each ring counter requiring 36 diode vacuum tubes to function – the intended function was to replicate the way mechanical ‘spinning-wheel’ computational machines worked, with digits ‘ticking up’ and rolling over to the next when they hit their max. On top of this, they had twenty ten-digit ‘accumulators,’ devices composed of the above ‘ring counters’ that could run 5000 addition or subtraction operations between themselves and some other memory bank every second. This was a marvelous feat of engineering that certainly gave the US a well-needed tactical edge. Still, there was far to go.
A Rather Mercurial Invention
Next on the line was something called “Delay Line Memory,” which relied on two compounds with rather interesting physical properties. First we had the quartz crystal, specifically, its ‘piezolectric’ quality. Piezolectricity is an electric charge that accumulates in certain objects when the object is exposed to mechanical stress. In a further subcategory of these elements, specifically those known as to exhibit ‘direct piezolectricity,’ such electric charge can be released in the form of physical expansion or shrinkage in the host material. What this means is, via the application or removal of electric charge, you can expand and contract a physical object at a very high speed, as long as you can alternate the current quickly enough.
When placed in mercury, the quartz crystals expand and contract, producing compression waves in the mercury around them, travelling through the container at the speed of sound in mercury (1450 m/s) until they hit the other quartz crystal, producing an inverted response on that end. By repeating the signal every time it came around, you could ‘store’ these pulses in the mercury; of course, given the high speed at which the waves moved through the medium, the number that you could fit was limited.
As Seen in TV
Finally, at least in our short history of incredibly outdated volatile memory, we come to Cathode Ray Tubes, or CRTs. You might recognize these from ‘old-timey’ desktop computers and televisions circa the 90s. Picture the big, slightly convex, rounded screens that behaved strangely when you put a magnet next to the display. So what exactly is a cathode ray tube? Well, in the general sense, and the one you’ll see above, it’s a device that allows a beam of electrons to be blasted against a surface at the end of the tube, lighting it up and causing an image to appear. Combine a bunch of these together and you have whatever you want: an oscilloscope, RADAR, or TV. Yet, it turns out there was more to these devices than just flashing lights, as a man named Freddie Williams discovered. If you power the electron beam above a certain threshold, rather than just impacting and lighting the phosphor that made up the display surface of the CRT, it would induce a burst of electrons that rise away from the impact point and quickly fall back to the surface of the phosphor display.
This is closely related to the photoelectric effect which made Einstein famous, where electrons impacting a surface energize the electron orbitals and induce a release of photons. Now, the immediate effect of this is to create a ‘potential well’; the area where the beam hit is electrically negative, wheras the areas around it are electrically positive. These wells faded over time, as the electrons re-distributed to a more electrically neutral configuration, but if you lined up enough of these charge wells on a single sheet, well, you had bytes! To read them, you placed a metal sheet on the side of the phosphor opposite to the electron beam (a small one about the size of the electron well) which would receive an induced charge if there was a potential well (representing a ‘1’) at that spot.
However, this would induce a potential well there even if there wasn’t one before. As such, you’d need an electron ray to re-write the data whenever you read it, if you wanted to use it again. Another issue was that of cross-talk: if two electron wells were too close together, they’d get mixed up and become useless. And furthermore, as mentioned above, the charge would dissapate over time, requiring the computer to constantly re-write the memory to their spots, a property known as ‘volatility’.
Well, this ended up going a little long; we’ll probably continue with more modern computer memory in a post in the near future!