Thanks, all.
If you look closer you could see the cable that connects to the deflection could, the outer metallic case are the deflection coils.
I had use those in the past in rx imaging.
I discovered (per
https://www.ebay.com/itm/LI-441-LI-441-vidicon-cathode-ray-tube-USSR-1987-1-LOT-1-PCS-/264260419173?oid=264216720419 ) that the LI-441 is electrostatically focused, but
magnetically deflected.
So that explains the external "deflection system"; the LI-441 seems to be not a good choice for this.
Does anybody know if the Russian LI-475 is electrostatically deflected? Or where to get a genuine electrostatic vidicon? (Can't seem to find any on eBay.)
This is really spoilsporting, but if you are interested in the image artifacts from various scan mechanisms you could just simulate it.
I've considered that - I think I'll try it before doing a build.
Here's my real interest in this:
A conventional TV system has a fixed frame rate (typically 25 or 30 Hz, ignoring interlace) and fixed resolution (# scan lines, or #X and Y pixels).
What if instead of scanning the image in a repeating pattern (as with a raster), you moved the scanning beam around the image in a
completely random pattern?
The time between re-visiting each spot in the image would be random, instead of 1/25s or 1/30s. So in some places you'd get more temporal resolution (more than 30 Hz) and in other places less. There would be no "frame flicker" effect. And no visible raster pattern.
If you did it fast enough (bandwidth limited), persistence of vision (and maybe phosphor persistence) would make it look smooth.
Of course it's also a less efficient scanning pattern because you re-visit some places more often than necessary.
But what would it look like? Would it have advantages re seeing quick movements that would otherwise be lost between frames?
I want to try it and see.