Starting at the beginning with Mirror Screws

Forum for discussion of narrow-bandwidth mechanical television

Moderators: Dave Moll, Andrew Davie, Steve Anderson

Starting at the beginning with Mirror Screws

Postby Panrock » Thu Sep 01, 2011 8:00 am

I'd really like to have a go at a "high" definition mirror screw for my next project. I'd get one of Darryl's 'universal' converters for the source and get in a position to use a mirror screw set as my main household colour tv. A handsome mirror screw set in the corner of the room would make quite a talking point.

Another thing... I could actually carry a mirror screw in to the NBTV Convention! I've stopped coming to conventions the past few years because my mechanical colour camera and monitor are just too heavy and bulky to shift... they really need a fork lift truck!

Trouble is... I'm still trying to get my head round the optical concepts of mirror screws - which should be simple. But alas, not for a brain like mine.

Starting right at the beginning....

Mirror screws all have this 'three section' corkscrew appearance. I guess that's how it always turns out when you have a complete single turn built in to the stack of mirrors - yes?

I'm trying to work out whether, for example, it would be possible to make a 90-line mirror screw to give a 4:3 12-inch horizontally scanned picture. Doesn't this mean the mirror screw would have to also be 12 inches in diameter? But if 90 lines meant a mere 4 degrees of rotation per line, wouldn't this also mean a tall thin picture resulted? Or by placing the line of light - and the viewer - far away, can this be overcome?

Is there a formula (or something) to work out the dimensions and aspect ratio of the picture vis-a-vis the number of lines and the mirror screw dimensions?

So far, I've been experimenting with a single polished aluminium bar on a pivot, and an LED bar (car maintenance) light, to try to get the hang of what happens.

To make the mirror screw's slats, I propose to design some sort of interlocking plate that automatically locks to other similar plates at the right 'stagger' angle. These would be laser cut from stainless steel. I have a good laser cutters nearby - they produced the 20-inch discs for my last NBTV rig.

Any help appreciated. Thanks! :)

Steve O
Panrock
Green padded cells are quite homely.
 
Posts: 870
Joined: Mon Feb 05, 2007 8:25 am
Location: Sedgeberrow, England

Postby kareno » Thu Sep 01, 2011 8:37 am

Hi Steve,

I've been through this (mis-)understanding process and I found it easiest to first consider a simple plane mirror on a rotating platform.

What you see in such a mirror is the world rotating, but at twice the speed that the mirror is rotating!

All of the scanning action happens when the mirror is more or less facing you and the angular segment of the world you can see in the mirror is a clue to the number of lines it can support.

If you can see 10 degrees of the world (in other words, the view in the mirror is what you'd see if your eye was at the mirror and you had blinkers that only allowed you to see ten degrees) then the screw can support 72 lines. Why 72? 360/10=36! Because the view of the world swings around twice as fast as the mirror rotates so its 720/10.

Getting the number of lines up requires that you move further away from the screw. You may find that, for 625 lines and a 12 inch image, you'll need to sit in the garden!

That's one drawback of the mirror screw actually - the aspect ratio is only right when you're sat at a certain distance from the screw.
kareno
 

Postby Panrock » Thu Sep 01, 2011 8:11 pm

A typically lucid account from you Karen. Thank you.

So, I've been thinking about what happens as we walk away from a working mirror screw - assuming horizontal scanning. Correct me if I'm wrong... I could well be!

1) The display looks smaller and smaller, 'cos it's getting further and further away.

2) From (1), the picture height decreases linearly with distance.

3) But the picture width decreases at TWICE that rate with distance.

4) So the picture gets narrower and narrower as we walk away.

5) Does that mean we then start seeing double/triple etc. pictures - side by side?

6) Say we are standing at the distance that gives a picture half the desired width (when there is a square picture). Could we then restore the aspect ratio to square, by inserting (and correctly angling) twice the number of mirror slats?

7) If so, we would have a picture with twice the number of lines. But the problem with a mirror screw seems to be that the more the lines - the further away you have to be. So the higher the definition of the picture, the harder that definition actually becomes to see, due to distance. If you make a bigger mirror screw to compensate, this is no solution, since then it won't support the same number of lines.

8 ) So this means, for a given aspect ratio (let's say 'square') and perceived picture size (ie. the angle it subtends) there is a set number of lines a mirror screw will deliver.

I've probably got completely lost here. Thanks for your patience. ;-)

Steve O
Panrock
Green padded cells are quite homely.
 
Posts: 870
Joined: Mon Feb 05, 2007 8:25 am
Location: Sedgeberrow, England

Postby kareno » Thu Sep 01, 2011 9:12 pm

Hi Steve,

You've pretty much grasped it I think.

It is easiest to consider a single mirror to start with. The task then becomes one of ensuring that the light source does a transit across the mirror (one scan line) in the required angular movement of the mirror.

Imagine you are in a round room with a circumference of sixty metres and a rotating mirror is standing at the centre. If, where you are standing you can see one metre of the circular wall through the mirror, then you can make a 120 line screw with it (remember, the image in the mirror swings around twice as fast and so the image will do two rotations (120 metres of wall passed by) for every turn of the screw.

The full screw is then a matter of arranging the mirror segments with the required angular difference. With the right angular difference the light source will leave one mirror as it enters the next.

An important point to understand about mirror screws: the reflection of the light source (strip) in each mirror segment is almost certainly from a different part of the light source but you can't tell because it is so uniform.

And yes, if you get too close to the screw the images become narrowed and you see multiple images much as you do with a disk monitor when the illumination spills over the top and bottom edges of the picture. This was exploited in order to widen the field of view of the mirror screw. The 'extra' images are a line or two our of sync but you don't notice that much on 120 lines or more.

You can imagine: if you only wanted to see 1/312th of the wall through the mirror (for 625 lines) you would have to stand a long way back and the mirror width must not exceed 1/312th of the wall.

There are other problems too:

1. The image brightness falls as the number of lines goes up.

2. The precision required in the screw will be immense for high line counts.

3. Sometimes the mirror plates cast shadows over one another leading to reduced viewing angle or missing parts of the picture.

A way around the large seating distance requirement it to use curved mirrors. These act as a kind of telescope thereby picking out a smaller piece of our imaginary wall. I suspect you'd have to do this for high line counts. I hope you've got CNC plant accurate to micrometres!
kareno
 

Postby Panrock » Fri Sep 02, 2011 1:15 am

Don't worry - I have no intention of trying anywhere near definitions such as 625 lines! :)

This has been helpful. One thing I don't yet understand though... why does the brilliancy of the picture go down with greater line count? Or at least why does it, if you keep to the same width of 'line of light' for the modulated source?

The mirror strip thicknesses are reduced for higher line count, and the more that have to be crammed in. So each then reflects less light - sure - but equally there are more of them present, so this effect would cancel out.

I understand you would tend to use a slimmer line of light with a higher definition display, in order to render the higher video frequencies, and (everything else being equal) this would then reduce the light available. But without changing the line of light, I don't see why more lines spells less light.

What have I missed? :roll:

Steve O
Panrock
Green padded cells are quite homely.
 
Posts: 870
Joined: Mon Feb 05, 2007 8:25 am
Location: Sedgeberrow, England

Postby kareno » Fri Sep 02, 2011 7:32 am

It's the old story, Steve: the average picture brightness is the brightness of a single spot. With a narrower mirror plate you see a smaller portion of the light source reflected in it and so the spot reflects less light.

Stray light becomes a problem too if the picture becomes too dark. The slightest chink of light around the curtains can lead to a bright line down the picture (that's assuming you scan horizontally i.e. have the screw axis vertical.)

Mirror screws are perhaps the ultimate in mechanical television. Many will dispute that so I'll qualify: It is perhaps the ultimate direct view mechanical televisor (direct view meaning that the light spot is a reflection/refraction of the light source.) A mirror drum and a very bright light source can produce a bigger, better picture on a ground glass screen.

An issue that I think Klaas mentioned is the peculiar apparent distance of the reflected light source from the eye. I'll assume a screw that rotates on a vertical axis:

If you perturb your viewpoint sideways i.e. move your head side to side while remaining facing the screw (imagine that peculiar neck movement some indian dancers can perform!) In that case, the light source appears a long way behind the screw. In fact, it appears as far back behind the screw as the light source is from the screw.

But if you perturb your head up and down in a similar fashion, the light source appears to stay on the mirror. In fact, you are seeing a different point on the light source when you perturb your viewpoint vertically but you can't tell because the light source is uniform. From a vertical parallax point of view then the light source appears actually at the mirror screw.

All's fine if you keep your head bolt upright (or perfectly horizontal if you are viewing from your chaise longue :) But if you tilt your head then the parallax becomes a mixture of both and the brain cannot make sense of the images presented to your eyes.

No amount of eye crossing or divergence can merge the images under these circumstances and it makes for very uncomfortable viewing indeed. That is a serious drawback of the mirror screw. You can't lean your head againt the side of your armchair and watch TV at the same time!

I'd stick with a CRT for your living room, Steve :)
kareno
 

Postby Panrock » Fri Sep 02, 2011 1:05 pm

Only on the internet could you find so much experience freely on tap about such an esoteric subject! Someday too, it would be interesting to hear your views about the Mihaly-Traub scanner. I once made a couple of these, expecting a free ride, but ran into raster distortion problems caused by the distance to the central polygon's reflecting surface varying as it rotated! Some other time perhaps.

I'm still having difficulty understanding why (say) twice the mumber of mirror strips, each of half the thickness, means less light intensity in the picture (rather than less light overall from the smaller picture) - but hey it's 3am and puzzling about this will get me to sleep!

Next, I must start thinking about a 'standard slat' design for the interlocking mirror elements.

Steve O
Panrock
Green padded cells are quite homely.
 
Posts: 870
Joined: Mon Feb 05, 2007 8:25 am
Location: Sedgeberrow, England

Postby Panrock » Sat Sep 03, 2011 12:16 am

This is my first go at a system that might lock the mirror slats together. Just three slats are shown - in practice there would be as many as the number of lines, with their stagger angle adjusted accordingly by the peg and hole positions.


I presume that the shaft centre not being coincident with the flat reflecting surface centres - but slightly behind - doesn't matter?

Steve O
Attachments
mirror_screw_locking_idea.gif
mirror_screw_locking_idea.gif (20.87 KiB) Viewed 31600 times
Panrock
Green padded cells are quite homely.
 
Posts: 870
Joined: Mon Feb 05, 2007 8:25 am
Location: Sedgeberrow, England

Postby Panrock » Sun Sep 04, 2011 6:51 am

On to designing a practical mirror screw. I've been juggling the various parameters around. In the end, the practical limits were set by
what sizes are available in stainless steel sheet and the tolerance of the laser cutting process. I wanted a large picture of good quality, viewable at a comfortable distance and with plenty of brightness.

The best compromise looks something like this. See what you think.

Picture/Screw: 192mm wide by 144mm high.
60-lines 25 fps 4:3 horizontal scanning - 1935 French standard.
Stainless steel slat (line) width: 2.4mm (using doubled up 1.2 mm plates).
Line segment: 12 degrees.
Minimum viewing distance 91cm.

One horizontal (square) pixel on the picture is equivalent to 9 minutes of arc. So I reckon the precision of angular positioning of the slats will have to be better than ±2 minutes of arc for a fair picture. I seem to remember the positional tolerance on my old laser cut discs was ±0.1mm. Therefore, if my mirror slat locating 'holes and pegs' are (say) 180mm from the centre, this amount of slack also gives an error of ±2 minutes of arc, so results should be acceptable.

I could build in a second set of positioning holes to the slats so if the precision of this rig in practice proved more than adequate, I could later change it to 120 lines (to be viewed from further away) without too much work.

Steve O
Panrock
Green padded cells are quite homely.
 
Posts: 870
Joined: Mon Feb 05, 2007 8:25 am
Location: Sedgeberrow, England

Postby Panrock » Mon Sep 05, 2011 5:26 am

I think my understanding still has some way to go... :oops:

1) The curved backs on my mirror slats are perhaps a bad idea. Won't they tend to shade the active mirror surfaces when the person viewing is looking in from above or below the mirror screw?

2) These are mirrors - so they'll reflect everything - not just the line of light! This includes the objects in the room and the people present. However I guess all this gets integrated into a sort of 'mush' and it reduces the contrast of the display? Looks like the mirror screw should be housed in a deep box with black internal surfaces...

3) I still need to understand this business about mirror stagger angles. For 60 lines, surely this should be 6 degrees, not 12 degrees.... it's just that the next mirror slat 'takes over' sooner than one would expect, because it scans at twice the speed of the screw rotation.
Panrock
Green padded cells are quite homely.
 
Posts: 870
Joined: Mon Feb 05, 2007 8:25 am
Location: Sedgeberrow, England

Postby dominicbeesley » Mon Sep 05, 2011 9:50 pm

Hi Steve,

I like the peg idea, it will be insteresting to see how you get on.

Having the axle behind the lines doesn't appear to matter in practice. I did do some calculations ages ago and it made for a small difference in the picture projected but if I recall correctly it was just similar to moving your viewing position forward or backwards, I don't remember it affecting distorition.

1. Not really, if you think the backs won't start to obscure the fronts until a few lines further up so wont cause any bother, remember the backs might get in the way after say a 1/4 turn but at that point the light won't be coming from the obscured line! Also to see the whoe picture you'll need to be a good few feet back I'd say for 60 lines and reasonable picture size so unless people are lying on the floor or swinging from the lampshades they wouldn't even shade the lines anyway...

I think rounded backs is a reasonable idea. It keeps the bulk of weight near the centre of rotation and the rounded backs (once painted matt black) should reflect less sharply?

2 Yup, effectively if the screw's sides are exposed they'll reflect everything and you just get a lower contrast picture. A deep black box helps....or turn off the lights...

3). Eh? One complete 360 twist is what you want or you'll end up with two pictures one above the other. If you have two mirrors facing the same way they will reflect at the same time, unless you come up with a clever two light system...So 6 degress sounds right to me

Dom
User avatar
dominicbeesley
Anyone have a spare straightjacket?
 
Posts: 685
Joined: Sat Sep 13, 2008 11:32 pm
Location: Hebden Bridge

Postby Panrock » Mon Sep 05, 2011 10:17 pm

Thanks Dom. I'm at Tamworth motorway services at the moment - without my glasses - so more on this later... ! :)

Steve
Panrock
Green padded cells are quite homely.
 
Posts: 870
Joined: Mon Feb 05, 2007 8:25 am
Location: Sedgeberrow, England

Postby Panrock » Mon Sep 12, 2011 6:03 am

I will be keeping notes on this project HERE though if it proves too expensive I'll either cancel it or tackle it very slowly.

Steve O
Panrock
Green padded cells are quite homely.
 
Posts: 870
Joined: Mon Feb 05, 2007 8:25 am
Location: Sedgeberrow, England

Postby Viewmaster » Tue Sep 13, 2011 1:28 am

I cannot quite understand why you have transferred this project to the BHTV forum, as all the experience on NBTV is on this forum, the BHTV seemingly dead on NBTV.
However I am quite impressed with your project and would give my 2p's worth for what it's worth!

You haven't mentioned on BHTV that you plan to scan at 24FPS although you have mentioned 1500RPM.
If at 1500RPM then the centrigugal forces acting on any imbalance will be high.
For example just 1 gram rotating on a radius of about 5 inches (your outermost tip of 240mm picture width) will be nearly 1 lbs. and pro rata.
SO........
Will the centre bearing shaft hole in stainless steel also be lazer cut and if so, how well will it fit the shaft to prevent inbalaces ? Also, how central to the slit length will each hole be? All this will contribute to out of balance forces creating rumble and vibration at 1500.

When I made my but 750RPM mirror screw I had rumble on that due to the shaft holes not being EXACTLY in the middle of each slat.

Also as the tips of the slats will be travelling at around 60 ft/sec if rotating at 1500RPM so you may need a soundproof housing with say, a polycarbonate sheet for the viewing window.

Silver steel may be the way to go for the shaft as it is ground to a tolerance of +- .00025 inches.and it has a fine surface finish.
But it's difficult to machine and die a thread on it if required. (3/4 or 1 inch dia perhaps )
A 13 inch length of 3/4 dia is about £9 plus postage.

To reduce noise further a simple ball bearing, housed in a centre drilled cone in the bottom of the shaft as a thrush bearing would be quieter (and cheaper!) than a ball race thrust bearing.

Oilite plummer blocks as bearings for the shaft will be quieter again than ball race bearings.
(I used them for my Mirror screw shaft)

Hope none of the above has put you off but just given you some food for thought.
Here endeth my 2p's worth.
User avatar
Viewmaster
Frankenstein was my uncle.
 
Posts: 1306
Joined: Fri Apr 06, 2007 4:50 am
Location: UK Midlands

Postby Panrock » Tue Sep 13, 2011 2:47 am

Hi Albert,

Viewmaster wrote:You haven't mentioned on BHTV that you plan to scan at 24FPS although you have mentioned 1500RPM.


I was planning to scan at 25 frames per second (1500 rpm) because this is one of the standards supported by Darryl's World Converter.

If at 1500RPM then the centrigugal forces acting on any imbalance will be high.
For example just 1 gram rotating on a radius of about 5 inches (your outermost tip of 240mm picture width) will be nearly 1 lbs. and pro rata.
SO........
Will the centre bearing shaft hole in stainless steel also be lazer cut


Yes. The 10 or 11 ounces of centrifugal force per gram should be set against the 38lb weight of the whole assembly. Nevertheless I'll have to be careful and right of the start am planning for a ball race both above and below the mirror screw.

and if so, how well will it fit the shaft to prevent inbalaces ?


I'm hoping for an interference fit. One thing I wanted to do was obtain the shaft before submitting the CAD for the laser cutting of the mirror slats, then measuring its actual dimension and working from that. I envisaged a suitable shaft for this weight of assembly might be of about 1-inch in diameter.

Also, how central to the slit length will each hole be?


Should be within 1/10mm -ie. the tolerance of the laser cutting.

When I made my but 750RPM mirror screw I had rumble on that due to the shaft holes not being EXACTLY in the middle of each slat.


I can well imagine. And with the weights I'm playing with, all the more reason why my screw should be substantially mounted, top and bottom. How did you achieve dimensional accuracy on your screw?

Also as the tips of the slats will be travelling at around 60 ft/sec if rotating at 1500RPM so you may need a soundproof housing with say, a polycarbonate sheet for the viewing window.


Could well be.

Silver steel may be the way to go for the shaft as it is ground to a tolerance of +- .00025 inches.and it has a fine surface finish. But it's difficult to machine and die a thread on it if required. (3/4 or 1 inch dia perhaps ) A 13 inch length of 3/4 dia is about £9 plus postage.


Good tip. Thanks.

To reduce noise further a simple ball bearing, housed in a centre drilled cone in the bottom of the shaft as a thrush bearing would be quieter (and cheaper!) than a ball race thrust bearing.


Interesting idea. The positioning and the profile of the internal cone(s) would clearly need to be spot-on for this. Would there be cones both above and below the support ball - for location?

Oilite plummer blocks as bearings for the shaft will be quieter again than ball race bearings.
(I used them for my Mirror screw shaft)


Reducing noise is always a good idea, but wouldn't bearing noise (especially if mounted on sound deadening material - the bearings in my big 30-line rig are mounted on rubber) be somewhat overwhelmed by the wind noise expected from a large mirror screw like this?

Hope none of the above has put you off but just given you some food for thought.
Here endeth my 2p's worth.


I thank you for your attention. :) You have a lot of experience with mechanical assemblies. Yes, there are many things that should give one pause for thought with a project like this - and I'm the sort of person who tends to bite off more than he can chew. I welcome all advice.

Steve O
Panrock
Green padded cells are quite homely.
 
Posts: 870
Joined: Mon Feb 05, 2007 8:25 am
Location: Sedgeberrow, England

Next

Return to Mechanical NBTV

Who is online

Users browsing this forum: No registered users and 30 guests