Talk:Storage tube

The article says "Very low bandwidth needs compared to vector graphics displays". A storage tube *is* a vector graphics display, so shouldn't that be pixel graphics? I'm a bit hesitant to change it myself, because it's been in the article since the beginning and nobody else has considered it wrong. --84.143.60.121 00:01, 7 January 2006 (UTC)manoj is a student of hctm college.!SAJAN & Amit are his classmate.

The meaning is to compare it to a vector display that is refreshed continuously. There were large, expensive systems at that time, about a full rack in size, that could continuously draw enough vectors to create a useful image without flicker. Evans and Sutherland is one manufacturer of such a system that comes to mind.

On a storage tube, on the other hand, a command to draw a vector is carried out once and the resulting image stays on the screen until erased. Back when these things were in use, about 1970, a high speed serial connection was 4800 baud, and that was unusual, 1200 baud was more common. The drawing rate could be much higher if the display system was directly connected to a minicomputer. This was, in fact, how many sytems were used. However, you must remember that a typical computer of that era used core and had a memory and instruction cycle time of 1 microsecond, so that, even in the best case, composing a complex image took a noticable amount of time. The typical display resolution was roughly 1000 X 800 addressable points, but the image was composed of analog vectors starting and ending at specified points, not individually addressed pixels.

A practical storage tube terminal, for example the Computek terminal, contained both an arbitrary vector generator and a character generator, which could draw the ascii characters as fixed size vector images based on information stored in the display, typically in analog form.

--AJim (talk) 04:41, 27 October 2008 (UTC)

OK, I'm confused...
Right, I understand how a Williams tube works, I've seen them and am familiar with the concept...

But just how does the image persist in a display storage tube, without the capacitance plate and continual re-scanning and refreshing of each line (which, for a Williams, was generally pretty low density / low resolution, on the order of 32x32 locations or 1024 bits per tube)?

And most particularly how does it maintain a vector display - without some kind of logic allowing it to "follow" each line, wouldn't it end up rasterising the image to whatever its scan resolution may be?

There's presumably some kind of clever mechanism at work to keep the phospors glowing until the high-powered reset pulse, but it's not mentioned in the text that I can see... 193.63.174.211 (talk) 12:07, 7 February 2014 (UTC)
 * Didn't some of them use a physical character mask inserted in a corner of the tube to generate text - defocussing the beam somewhat, deflecting it through a set location in the mask (say, an 8x8 grid for 64-character/6-bit encoding), and then putting it through a secondary deflection to aim the masked beam at a particular point on the display? 193.63.174.211 (talk) 12:11, 7 February 2014 (UTC)
 * There is a flood gun which supplies electrons to the whole screen. That keeps the positive charged regions lit, somehow without discharging them. I believe lines are drawn using partly analog logic. There are ADC to generate a voltage for the start and end, and then analog circuitry to generate a line in between. I don't remember that it is stepped, but just drawn straight. (That I remember from using the 4010 years ago.) Gah4 (talk) 22:02, 26 June 2020 (UTC)
 * There is a flood gun which supplies electrons to the whole screen. That keeps the positive charged regions lit, somehow without discharging them. I believe lines are drawn using partly analog logic. There are ADC to generate a voltage for the start and end, and then analog circuitry to generate a line in between. I don't remember that it is stepped, but just drawn straight. (That I remember from using the 4010 years ago.) Gah4 (talk) 22:02, 26 June 2020 (UTC)
 * There is a flood gun which supplies electrons to the whole screen. That keeps the positive charged regions lit, somehow without discharging them. I believe lines are drawn using partly analog logic. There are ADC to generate a voltage for the start and end, and then analog circuitry to generate a line in between. I don't remember that it is stepped, but just drawn straight. (That I remember from using the 4010 years ago.) Gah4 (talk) 22:02, 26 June 2020 (UTC)

Disputed content
This article seems to have got its wires crossed. It appears to be attempting to describe (in part) an image storage tube but describes (also in part) the operation of a Williams tube but incorporating parts from an image storage tube (which it never had). The article is almost devoid of referencing anyway.

The two devices were physically distinct. The Williams tube was an early method of storing binary data (i.e. it was a computer memory for binary data). The image storage tube was a method of continuing to display an image on the screen after the signal that created it had ceased to exist (i.e. nothing to do with binary data). The former type did not require a phosphor screen and most Williams tubes did not have one (it would not have been possible view it anyway as a metal plate covered the 'screen' end of the tube which was essential for its operation). 86.174.156.70 (talk) 16:12, 13 October 2016 (UTC)
 * They should be distinct, but as well as I know the story, the Williams tube had the advantage that you could visually extract the bits. So, just like an image tube. And it is possible to read the screen of the 4010, as there is an associated printer which does that. It seems that they aren't as distinct as might seem. Gah4 (talk) 21:57, 26 June 2020 (UTC)
 * This statement is incorrect: "did not require a phosphor screen". The Williams tube did indeed require a phosphor screen. It was the slow decay of the charged spots on the phosphor that provided the storage capability, and the saturation of that signal that allowed erasure. There is a difference between the two systems, the Williams tube lacks the flood gun of something like the 4010, but the two concepts are very much related. As this appears to be the basis for the fact hatnote, I will be removing the tag. Maury Markowitz (talk) 14:07, 7 September 2023 (UTC)
 * Slow decay phosphors store energy as excited electrons and not as charge. But okay, there needs to be an insulating, or low conductivity, material of some sort. It likely wouldn't work to just use the glass, and so, yes, phosphors were what was known. Charge needs to bleed off at the right rate, not too fast, not too slow. It doesn't need to be a visible phosphor, but I thought I remembered stories of visual readout of the bits. I might have to find a reference for that. Gah4 (talk) 21:06, 7 September 2023 (UTC)
 * In the WP article, I dispute the statement that the reading of the Williams tube was done by measuring the capacitance between the metal plate and the phosphor(or front glass of the tube). What was read were the varying electrostatically coupled voltages produced during operations of the tube. These voltages allowed not only reading but produced a 'look ahead pulse' for regenration purposes.
 * Some of this could be cleared up by looking at the patents and info on the Williams Tube, and also the old white paper for the RCA 6499 Radechon (not discussed in the WP article), which like the Williams tube must be electronically refreshed, but uses a Barrier Grid (dielectric mesh) instead of a phosphor, and also looking at the manual for the first commercial storage scope, the Hughes Memo-Scope, which goes into much detail on how its CRT stores images, having a dielectric mesh and a flood gun, and requiring no refreshing. The H/P 140 and 141 storage versions would have a newer technology version of the old Hughes product. the TEK 4010 is also a mesh+flood display unit.
 * There is also a USSR tube, an improvement on the Williams tube, described in the 1960 book "The Cathode Ray Tube Memory of the High Speed Computer of the U.S.S.R. Academy of Sciences". English translations show up from time to time on A*. Now that I think of it, that tube is similar to the Radechon in that it uses a barrier grid.
 * So here are some online papers and documents that should help the maintainers with this WP page, and I hope these are useful:
 * Williams Tube info and RCA 6499 Radechon info: https://bunkerofdoom.com/lit/CRT-scopes/CRT_memory/index.html
 * Hughes Memo-Scope model 104-D manual, start on page 1: https://bunkerofdoom.com/lit/hughes104d/hughes_104d_pub.pdf
 * TEK 4010 maintenance manual, page 6-48 'storage circuits' (pdf page 160): https://w140.com/tekwiki/images/7/72/070-1183-01.pdf
 * The 4010 tube type is one of these two, depending on serial number:
 * Vl 154-0662-10 Snr. BOI0100-B059999 ELECTRON TUBE: CRT 80009 154-0662-10
 * Vl 154-0740-00 Snr. B060000-        ELECTRON TUBE: CRT 80009 154-0740-00
 * The 154- numbers are Tektronix part numbers for the CRT. The only person I know of who might be able to dig up a data sheet would be Peter A. Keller formerly of Tektronix.
 * Searching "Tektronix Peter A. Keller" will probably turn up more history of CRTs and Tektronix. Also see his book "The Cathode Ray Tube". CRT Hobbyist (talk) 01:31, 9 May 2024 (UTC)
 * Vl 154-0740-00 Snr. B060000-        ELECTRON TUBE: CRT 80009 154-0740-00
 * The 154- numbers are Tektronix part numbers for the CRT. The only person I know of who might be able to dig up a data sheet would be Peter A. Keller formerly of Tektronix.
 * Searching "Tektronix Peter A. Keller" will probably turn up more history of CRTs and Tektronix. Also see his book "The Cathode Ray Tube". CRT Hobbyist (talk) 01:31, 9 May 2024 (UTC)

Storage CRT oscilloscope.
I thought the Tektronix terminals came after the storage CRT oscilloscope. Tektronix was (and still is) big in the oscilloscope market. It is often needed to see things that only happen once, so the storage CRT was very convenient. Now they use digital storage. In any case, the terminals use much bigger (and presumably harder to make) tubes. Gah4 (talk) 22:05, 26 June 2020 (UTC)

X axis represented individual bits within a word
The article says: X axis represented individual bits within a word. I thought for the 704, they used 36 tubes, for parallel access. With only 1024 bits each, you will need some number of tubes. Gah4 (talk) 23:09, 24 October 2021 (UTC)
 * OK, it was the 701, where the 704 got core memory. The 701 has 2K 36 bit words, stored in 72 tubes. As the bits are in a 32x32 array, it would be complicated to put 36 bits in one row. Note that core is also normally arranged with one plane for each bit of the word, where X and Y address words. Gah4 (talk) 23:48, 24 October 2021 (UTC)
 * OK, it was the 701, where the 704 got core memory. The 701 has 2K 36 bit words, stored in 72 tubes. As the bits are in a 32x32 array, it would be complicated to put 36 bits in one row. Note that core is also normally arranged with one plane for each bit of the word, where X and Y address words. Gah4 (talk) 23:48, 24 October 2021 (UTC)