Mirror Drum Scanner

[kd6cji]

Hi,
I am planning to make a NBTV display using mirror drum. "Basic NBTV Techniques" booklet that I have mentioned that a few designs were published in NBTVA newsletter (Vol 4/1, Vol 5/1, Vol 3/4 and Vol 8/1). It may have been reprinted in other club publications as well. If someone has access to this newsletters or other design information for Mirror Drums, I will appreciate any help!

Thanks and 73,

Sergei KD6CJI

[Stephen]

If you are interested in a Traub-type scanner, see Mr. Traub's British Patent 425,552 on the second page of the "Patents and Articles" section of the forum. The intriguing thing about the Traub-type scanner is that the number of scanning lines equals the product of the central rotating mirrors and the surrounding fixed mirrors. That is, for a 32 line system you could have eight fixed mirrors arranged around a rotating hub with four mirrors.

[kd6cji]

Hi Stephen,
I was actually thinking more like "regular" drum with 32 mirrors. However, Traub scanner looks very interesting. I wonder if anyone (in modern times) attempted to build one... I might go for it, as soon as I finish my Mirror Screw project.
Have a great weekend,

Sergei KD6CJI

[Panrock]

Hi Sergei,

I built a couple of Mihaly-Traubs many years ago, one for the camera, one for the monitor. I used a laser torch (modulated) for the monitor light beam. Only one array now survives, in poor condition. I think I once posted a picture elsewhere on this forum.

I had a lot of trouble with raster distortions due to variations in the optical path as the central polygon rotated. Maybe someone else here can tell us what the correct geometric figure should be to dertermine the positions of the fixed array of mirrors? I used a circle, and it was wrong.

Steve

[kd6cji]

Hi Steve,

I found the picture of your scanner (well, part of it) in one of the posts. What is interesting to me is that you have 32 mirrors. Looks like Mihaly-Traub scanner had less fixed mirrors. If rotating polygon had 4 mirrors then 32 line image should require only 8 canted fixed mirrors. (32/4). The first scanner had number of fixed mirrors equal to number of scan lines in circle with double sided rotated (at twice the speed) mirror in the centre of the circle.

Could this be a cause of the problem?

Regards,

Sergei

[Panrock]

No, I was 'greedy for lines' and my original system used 35 mirrors surrounding a central polygon. I've forgotten how many sides this polygon had but I think I was aiming at a 120 lines or so from a compact mechanical display, with easy adjustment of line spacing with the stationary mirrors.

This was later cut down to use just a single double-sided central mirror because of all the trouble I had had with raster distortion using the polygon.

The only effect of using more mirrors than the original M-T inspiration was to produce a smaller picture at a given distance from the array. The reduced sweep angles thus involved should reduce any raster distortions, not increase them.

Steve

[kd6cji]

Hi Steve,

I don't know the details of your project, but reading Mr.Traub patent, I think the key is a double reflection from rotating polygon. Even though I am not a specialist in optical geometry, it appears from patent's drawing that the light beam when reflected back from fixed mirrors to the upper part of polygon and then to the screen would have more or less equal path for any scanning point. The distances from fixed mirrors to polygon are the same and the distance from polygon to the screen is fixed. Any thoughts?
I for one getting more and more interested in M-T scanner...

Thanks,

Sergei

[Stephen]

I think that you are correct about the double reflection feature, Sergei. John Logie Baird had an earlier patent, British Patent 374,564, that also offers a multiplying effect with rotating and stationary mirrors. This patent is also in the Patents and Articles section of the forum.

Although Figure 1 shows a single canted revolving mirror, Mr. Baird indicates on page 4, column 1, lines 51 through 68 that a multi-faceted mirror with a pyramidal shape would disect or reconstruct an image with a number of lines corresponding to the product of the rotating and fixed mirror surfaces. This design is simpler than Mr. Traub's scanner. However, I suspect it would have much more optical distortion without the double reflection off of the fixed mirrors.

[kd6cji]

In fact, upon careful study of Mr.Traub patent, I found the following on page 2, lines 64-71:
(if beam of light is reflected from fixed mirrors to the screen) ... "This arrangement gives rise to considerable distortion of the picture particularly when (as is preferable for mechanical reasons) the number of reflecting surfaces on the rotating member is made small compared to the number of fixed mirrors. Such distortions are not produced by devices in accordance with the present invention". Sometimes it helps to read...

Sergei

[Panrock]

This is beginning to look very interesting indeed...

Maybe I will have to return to my array and see what actually happens with a central polygon and a double reflection. I would guess the extra reflection would naturally result in a wider-angle picture, though this would be countered by the tendency of the returning beam to miss the central polygon altogether. I can see a lot of faffing around, experimenting with all the different possible sizes and spacings of polygon and mirror.

Has somebody ray-tracing software to work all this out? I have RayTrace somewhere. A complication would be the light path changing in three dimensions though the 'action' in the third (vertical) dimension is relatively simple. So much easier to build to a plan rather than do it by trial and error.

Something to work on for the next convention perhaps!

Steve

[kd6cji]

I don't think that size of fixed mirrors are critical. In horizontal plane all fixed mirrors are at right angle to the centre of polygon, so unless mirrors are miss aligned there is a little chance that return beam will miss polygon. The size of polygon mirrors probably should be 2-3 times higher than fixed mirrors. The width of polygon mirrors are not critical either and can be rather narrow. Let us assume 48 line image. We need 8 fixed mirrors on arc of 1/3 of a circle. If we put those mirrors on a 4" radius circle, their size will be about 1" X 1". Assuming polygon from 2" diameter circle, polygon mirrors will be about 1" wide and 2" to 3" high. If to follow Mr. Traub suggestion to use two mirrors for each polygon side (lower at right angle and upper canted), the lower mirror don't have to be high, may be 0.5"?
But you are right it would be nice to do some modeling in software before starting to cut mirrors... Unfortunately I don't have any software of that kind.

Sergei