[ChuckMcM]

You might find this note[1] interesting to read. I would love to see a commercial off the shelf cathode ray tube (CRT) that could do 480 Hz, so please share with us a link if you can find it in your notes.

A CRT monitor, with a refresh rate of 480 Hz, allows just 2 mS (2.083333 mS to be precise) for excitation of the phosphor. As you know, (here is a similar explainer [2]) the phosphor is excited by the arrival of electrons which contribute their kinetic energy when they impact the phosphor to the electron energy of the particular phosphor. When the energy decays back to base level, it emits a photon at a frequency that is characteristic of the band gap between the excited and rest state of the electrons in the outer orbit of the crystal.

The more electrons you can get excited in the phosphor crystal, the brighter the display. But the more electrons in the beam gives the beam more inertia and so rapidly moving it from one side of the screen to the other requires higher voltage potential.

I haven't seen any high refresh rate CRTs but back in college we had a device which was a "flying spot" scanner which had a very fast phosphor but it was enclosed in a black box and basically relied on photographic film for persistence (it was made by a company called "Dicomed"). The shorter the persistence of the phosphor gave it a very high dynamic range of brightness which allowed us to transfer digital photos with a high dynamic range to film without losing fidelity.

I haven't really followed the evolution of cathode ray tubes post the Sony Trinitron era, I would really love to play with the tube you mention, or at a minimum get the engineers specification for it. I'm assuming it was full color? (the Dicomed's screen was a sort of yellowish white and it had three filters that it would place over the screen to scan out a color image).

[1] https://www.nasa.gov/centers/dryden/pdf/87778main_H-609.pdf

[2] https://www.phosphor-technology.com/how-do-phosphors-work/

[willis936]

P22 phosphors are the ones used in all trinitron PC monitors. I get conflicting numbers on the actual decay time of these phosphors, but with a 240 fps camera the decay time is difficult to judge by flipping through frames. I’d have to make a proper set up to try and even give an estimate.

I can safely say the P22 decay time is below 1 ms. Here [0] it is stated to be 100 us for G & B and up to 1 ms for red. A commonly referenced list of phosphor decay times [1] lists it as “medium”. In general, CRT phosphors seem to have sub 1-ms decay times unless you want a long persistence.

The limiting factors will be in a monitor’s max horizontal scan rate (typically 130 kHz or lower) or the DAC’s maximum pixel clock (it used to be 400 MHz but most VGA adapters these days struggle to keep 200+ MHz stable).

0. https://www.epanorama.net/documents/video/phosphor_decay.htm...

1. http://www.bunkerofdoom.com/tubes/crt/crt_phosphor_research....

[ChuckMcM]

Excellent data! Now you need to add the time to saturation (TTS) which gives you how long the beam has to sit on the phosphor to achieve the maximum amount of light. Different tubes do this differently from moderating the energy in the beam itself to modulating "time on spot".

What you end up with at the end of all this research is a really nice ideas of the "hard bits" around engineering a CRT for a particular application. When we looked at CRTs in depth in the EE program one of the things that came across was how so many of the things you had to vary were interconnected. The professor suggested that was why there were so many different models to choose from, even when they were all the same form factor.

Thanks for the links too, I've added the bunkerofdoom one to my collection.

[ChuckMcM]

And to be clear, the time it takes for a phosphor dot to go from "off" to "full bright" to "off" again is its frequency response. And every dot of phosphor has to be "visited" by the electron beam on every frame. With those two numbers you can get the absolute fastest you can scan through the dots with full dynamic range.

In the CRT designs we looked at in college they typically used a fixed "dwell time" of the beam, so it was on each pixel for a fixed amount of time, and modulated drive power to get different intensities. (some phosphors are non-linear in their response so you can correct that in the drive table).

And as I was reminded in email, the beam has to visit every pixel AND get back to the top of the screen for the next frame. So on a 480i screen that is 640 x 240 per frame at 30Hz, that is 30 frames per second / 640 * 240 pixes per frame so you have just 195 uS to spend on each pixel. (less than that actually since with overscan you have 720 pixels per line but it is a very short time).

[buildbot]

Somewhat tangential, but Dicomed also made pretty special large format digital backs at one point [1] that really have yet to be eclipsed in terms of chip size (definitely in quality). Interesting they also produced what sounds like a really high quality digital to film solution too.

[1] http://www.epi-centre.com/reports/9604cs.html?LMCL=xIZMrP

[matt-attack]

FYI the SI unit of the second is abbreviate with a lower case “s”. Only units named after humans have uppercase abbreviations.