It’s well known how PPI (Plan Position Indicator) Radar CRTs made their way into computing, via Whirlwind, its Charactron display and the protoypes for the SAGE system. However, there was also an alternative technology, dark-trace CRTs, with designation P10, which would draw dark images on a backlit screen. These scotophor CRTs had some advantages, like a permanent image, much like E-Ink, but also major disadvantages, like using heat for drawing and cooling for erasing the images, which generally provided poor update rates due to the thermal mass involved.
However, there is a film showing a rather large live-updated PPI display, the “Skiatron”, used in WWII by British Radar (at 15:03):
(I guess, the acronym “GTI” used in the commentary is for Ground Target Indicator?)
Here’s a still of that scene, showing the Skiatron:
Post-war, these were apparently (according to Wikipedia) used in a few storage oscilloscopes, but I never heard of this technology being even considered for computer displays. As the film shows, this could have been a viable display technology for that purpose with the major properties of E-Ink.
Does anyone know more about the background of this and why these were not considered for computer displays?
Interesting technology! Sounds like it might be too slow for random access… and indeed, it’s like flash, in that erasure is the whole surface. Wikipedia draws from this PDF:
A variety of methods were used to erase the skiatrons. UK radars used fans to cool the tubes which were being heated by the stage lighting of the projectors. Simply turning off the fans made the tube begin to warm up, the erasure taking perhaps 10 to 20 seconds. German examples used a thin, transparent layer of tungsten deposited on the front of the tube, which heated up when current was passed through it. This provided much faster erasing [5 to 10 seconds].
I see the Williams-Kilburn tube was patented in 1946 - in time to be considered and used for computer storage. Perhaps one could regard this as a successor, and in this case suitable for rapid random access (admittedly write-after-read.)
I’d say, it shares some principles, like the electron wells caused by the secondary emission. Notably, these are crucial for its functionallity, since it’s those electron wells, which are detected by the read cycle of the Williams tube and also produce the visible traces of the Skiatron. (So, instead of being based on the primary emission, both are based on the secondary emission and local electon depletion.) But, in contrast to the Skiatron, the CRT of the Williams tube is phosphor based. Having said this, knowledge of the Skiatron and its working principles may have well contributed to the invention.
(This may be worth an ammendment to that post. – Update: I added a postscriptum to this regard.)
The erasure problem. In the film you can see that the pattern is being added to, not erased. There were some experiments with heated layers and other solutions but none of them really worked. Even with significant heat applied, erasure took something on the order of seconds.
And while a system using klieg lights and fans might be OK for a radar station, it seems less useful in non-industrial settings like a terminal room. So it’s fine for a series of still images, but moving images not so much. The solution used for the Star Wars animations is far more practical.
Actually, seeing this short glimpse of it in that film, I’m not so sure. Apparently, persistence is a function of the amount of current, but also much more a function of repeated intensification. In that film, we can see that the image fades rather fast away, where we see the moving radius, maybe even faster than the kind of slow phosphor that was otherwise used for PPI displays.
In contrast to TV applications, this should lend itself pretty well to computer control. Also, early X/Y displays all feature slow phosphor (e.g., in the preferred dimmed light viewing condition, P7 phosphor may be visible up to 10 seconds), the main concern being flicker and it was all about stable images. Fast, varying displays weren’t really a major application – and even Spacewar! profited from the iconic trails.
If you watch the film, you can see that the image decay time is about 10 seconds - the scan rate of the radar, an AMES Type 7 in this case, appears to be set to the fastest 8 RPM, but would more typically operate at 4, or even lower during low-activity times or high winds. At 8 RPM the image is being drawn every 7.5 seconds, and you can see the previous image is still quite visible. The half-life in this case is on the order of 15 to 20 seconds, which it had to be to work at the slower refresh rates.
Because the radar was revisiting the same place on a slightly faster basis than the decay time, the image never fully fades. At this RPM, an aircraft would leave a trail of dots, with the latest one being the “brightest”, which is precisely what you want as you can immediately tell the direction and measure speed with a scale.
So it is fine for images that are largely unchanging and being periodically scanned, like radar, but not so useful for things like computer displays. If the display changes more radically, like scrolling text, you would have to wait the entire fading period or you end up with “mush”. One can improve that, but only at the cost of making the display have less contrast.
Moreover, you have to continually update the display within the half life, which is fine, but does mean your expensive computer is doing something best left to the hardware. IBM went this route with the 2250, but at $280,000 in 1964 dollars this was not inexpensive!
Tektronix’s storage tubes were a much better solution. Cheap DRAM even more-so.
I was going to mention these! They seem quite miraculous, and again could possibly have been pressed into service as memories, except that once again the erase is of a full screen, so you’d need at least a pair, or perhaps N+1 screens, to save the data that’s about to be erased.
My point was rather made from a late 1940s, early 1950s perspective: nobody was really thinking of scrolling text and the great problem was image stability. (They actually went for the slowest phosphor available.) Text at all was quite a problem. It wouldn’t have been possible without the Charactron. That text was a problem, was also a consequence of the problem of image stability, as cooling times between beam deflections was of utter importance, which was also the major bottleneck, which in turn aggravated the flicker problem.
But, when the major application was plotting data, with X in the accumulator incrementing across the screen and plotting the data against this in the IO register, which also provides the Y coordinate, I believe, the Skiatron may have been a viable option. (We may argue, whatever happened next, or didn’t happen, was because the choice for fluorescent displays had been already made.)
(On the other hand, this may be also the short version: things strated with the Charactron, and this was what people at MIT, like Olsen, became familar with. Why consider anything else, if there’s already a well working technology, which you’re also already familar with?)
Edit: What I’m here referring to regarding poperties is:
Screen characteristics have been measured by a photo-electric cell and a recording galvanometer. It is shown that the intensity of coloration increases with excitation until a saturation value is reached. Coloration produced by a pulse of excitation decays in several seconds and is referred to as “transient,” while that produced by a series of pulse excitations, applied at intervals, may take several hours to decay and is therefore known as “persistent.” Increase of temperature reduces contrast but increases the decay rate. Decay rate also increases with illumination, from zero in the dark to a saturation value at about 5 000 ft candles. The overall efficiency of the screen increases with beam voltage. The decay rate of the persistent colour can be greatly increased by periodic electron bombardment at low intensity. The presence of impurities also increases the decay rate, but the effect is not permanent.
(King, P. G .R., Gittins, J. F: “The Skiatron or Dark-Trace Tube”, Journal of the Institution of Electrical Engineers, Vol. 93, Part IIIA, 1946, pp 171.)
So, I guess, with a combination of things like a heating element under electronic control (like the tungsten film in the Blauschrift-Röhre, which on its own already achieved a decay similar to P7) and exposure under computer control, including a low-excitation erasing exposure, workable parameters may have been not totally out of hand.