Increasing image size using a masking disc

[Andrew Davie]

If a Nipkow disc were so constructed that instead of a single spiral of holes, going around the disc once, it had two (or three, or more) turns around the disc, and the holes were therefore doubled in apparent visual vertical distance from each other, and therefore twice the size (or three, etc.), spaced twice the width apart, and the disc spun at twice (or three, etc.) the speed, then we would have the effect of a much larger disc.

To prevent the incorrect hole showing (because with a 'double' spiral, we have two holes showing over the monitor area at any time), we place a masking disc in front of the Nipkow disc, running at such a speed that for every rotation of the Nipkow disc, we have half a rotation (or 1/3, etc) of the masking disc. The masking disc so designed that it covers the hole(s) that we should NOT see for the particular orientation of the Nipkow disc. In essence, the masking disc runs at 750rpm, and the mask 'enables' the particular scanline that we SHOULD see (by the presence of a 'hole' which is in fact a whole scanline long), and the Nipkow disc runs at nx750rpm, where n is the number of spirals used.

This system would allow much larger images to be produced for a given Nipkow disc size, also allowing brighter images due to the increased hole size, but at the cost of increased 'keyhole' distortion due to the greater use of the disc area.

Cheers
A

[Stephen]

This is a great idea, Andrew. Another possibility is to use a couple of LED arrays, one for each spiral, and electrically switch between them for each spiral in synchronisation, such as with a commutator along the hub of the disc. This would have the added advantage of higher efficiency, in that all the amplifier power would be directed into half of the picture at any one instant instead of spread over the entire picture.

John Logie Baird described a similar but different approach in his British Patent 294,267, filed 21 January 1927. In this patent he shows a couple of modulated light sources, with each light source directed to the inner or outer spiral by mirrors or internally reflective tubes associated with each aperture. With respect to the embodiment that describes the use of internally reflective tubes, I believe he anticipates my proposal for an optical fibre disc by 80 years. His "optical commutator" has a different location, about the edge of the disc instead of about the hub, which I believe is less advantageous, but it is otherwise much the same in principle.

Mr. Baird describes yet another, and very intriguing, concept in his British Patent 326,251, filed 10 October 1928, wherein he proposes a shutter-like disc with five zones and five sets of light sources, for a display, or photosensors, for a camera. A lens disc, with six lenses each sightly canted at different angles to achieve six scanning lines, shifts the image onto the shutter disc, with five revolutions of the lens disc for each revolution of the shutter disc. Thus, the lens disc provides six scanning lines for each of the five zones, providing 30 lines of resolution for the entire picture.

For the 32 line system, such a device could use an eight lens disc in front of a four zone shutter disc and four modulated light sources switched on in synchronisation with each zone of the shutter disc. Each lens of the lens disc would capture and focus all of the light that it receives from the modulated light source turned on at that instant, providing an incredibly bright display, since all of the light from the modulated light source passes through its respective shutter aperture in the shutter disc.

I have posted these patents at http://www.taswegian.com/NBTV/images/GB294267A.pdf and http://www.taswegian.com/NBTV/images/GB326251A.pdf for reference.

[Stephen]

As a variation of the Baird design, it seems that the shutter disc would be superfluous. Imagine four different laterally spaced LED arrays behind an eight lens Nipkow disc. Each disc lens has a slight radial displacement from its adacent lenses so that it represents a different one of eight scanning lines. The video signal energises one LED array at a time, sequencing from array to array for every revolution of the lens disc. The lateral spacing between the LED arrays is such that the eighth scanning line generated by the lens disc for one array is adjacent to the first scanning line generated by the lens disc for the next array.

That is, when the rightmost or first array is active, the lens disc completes a single revolution to produce lines 1 through 8. The second array is active for the second revolution of the lens disc to produce lines 9 through 16. The third array is active for the third revolution of the lens disc to produce lines 17 through 24. The fourth array is active for the fourth revolution of the lens disc to produce lines 25 through 32. Consequently, a complete frame comprises four revolutions of the lens disc. Thus, the lens disc would rotate at a rate of 3000 rpm for the 12.5 frames per second system.

An optical fork may sense the revolutions of the lens disc and activate a counter that sequences the LED arrays to keep them in synchronisation. Since there are only eight lenses about the perimeter of the lens disc, they may have a relatively large diameter and therefore have a correspondingly high light gathering ability to produce a bright display.

[Stephen]

Another intriguing variation of Mr. Baird's system, which he mentions only in passing in the patent, is substituting a mirror-type scanner for the lens disc. In this case, the camera or display could use, for instance, a Weiller mirror drum with only eight mirrors which work in combination with four laterally displaced photosensors or modulated light sources. This would allow for a simple and light weight mirror drum that would be much easier to rotate and align.

[Steve Anderson]

Gents,

Two or three years ago I embarked on building a very similar device, a drum monitor with two spirals. (see Pic.1, the only one I took at the time). However due to time constraints it never got much further than the 'rotational' stage. It did look promising though, maybe I'll resurrect it.

I started to write an item for the newsletter and here's the beginning of it....

A direct drive NBTV drum based display.
Steve Anderson.

I decided to build a drum based monitor for NBTV. For a reasonable sized image of say 32mm x 20.5mm conventionally requires a drum of over 300mm in diameter. That’s not an easy task even with access to a large lathe to turn it.

To reduce the size and mass allowing the use of a smaller motor a dual spiral of holes on the drum was considered, reducing the diameter down to 174mm. The downside is that it now has to rotate at double the normal speed at 1500rpm. However, this is exactly the speed a PAL VCR head drum assembly rotates at.

In addition, the light source needs to be split between the two spirals, the first (right-hand as viewed) for the initial sixteen holes, then the second (left-hand) one for lines 17 to 32. Because there is relatively wide spacing between the two spirals of holes it’s not important if each light source spills over a little into the adjacent spiral of holes.

The mechanical bit.

The basic form of the drum was cast and then turned on a lathe out of Aluminium, once the lathe work was commenced it was not removed from the lathe chuck until complete to ensure true concentricity and balance, the picture hole drilling was done on a specially constructed jig.

The intention was to get the hole drilling as accurate as possible within the confines of what can be achieved in the usual NBTV den. Two items needed to be addressed; one is accurate angular placement of the picture holes on the circumference of the drum, and the spacing from line to line.

The drum was fixed horizontally on a drilling jig constructed solely for this purpose. With 32 holes arranged in two spirals (now a helix) results in a subtended angle between holes of 22.5º. To get an accurate angular measurement a large square was marked out on a smooth concrete floor, the diagonals were checked to ensure accuracy. The two centre-lines and the two diagonals provide eight of the drilling points, the intermediate ones were simply carefully measured and marked. A laser diode pointer was fixed to the drum assembly and each time the drum needed rotating by 22.5º the bright red spot from the laser was moved from one marker to the next one, the drum then being clamped firmly. The whole affair resembled a model of Stonehenge!

With a tolerance in placing the laser spot within 5mm results in an angular error of less than 0.04º, less than 0.1 pixels! The majority of the errors would come from play in the bearings of the drill so a good quality one must be used. The drilling jig needs to be bolted to the floor, quite literally, to stop it moving.

The thickness of the drum rim is 3mm and the holes were to be 0.75mm giving a small degree of overlap between lines. But at first a larger and sturdier drill bit was used to provide a countersink of 2mm in the drum rim to reduce the risk of the smaller drill bit wandering off centre on the curved surface when its time came. Both drill bits were ground down in length leaving just enough length extending from the drill chuck to penetrate the drum, less than 5mm to reduce flexing to a minimum.

The drill body was insulated from the chassis by mounting it on machined Teflon blocks; this provided a means to ascertain when the countersinking drill bit made contact with the drum using a battery, resistor and a LED. A M6 screw was used to then push the drill motor slowly forward 2mm, at a thread pitch of 1.0mm; two turns gave the correct depth of countersink.

At this display size the spacing between lines is 0.66mm and this was the tricky one to solve. A M6 ISO (course) screw has a thread pitch of 1.0mm, so a ‘sinking derrick’ was arranged for the drum using some long M6 screws that were rotated 2/3 of a turn (240º) each time a hole was started or drilled. The screws used were stainless steel which have a better surface finish on the threads than the usual zinc or nickel plated variety, where tapping was required a new M6 tap was used. M6 (fine thread) screws could have been used which have a finer thread pitch of 0.75mm, but they’re unobtainable here.

This exercise required concerted concentration, a check-list was created and items crossed off as each one was done. I gave the family some cash and sent them off to a movie and then dinner. I also unplugged the phones and turned off the mobile. It was done at night, and thankfully no-one called round. The beer stayed in the fridge until the task was completed.

The whole of the drum was sprayed matt black with the exceptions of the where the drum is affixed to the motor, and the 0.75mm holes which were plugged with toothpicks to keep them free of paint; just a 0.1mm coating could reduce the light output quite a bit.

For frame synchronization a small area on the outside of the drum was polished then masked prior to spraying and using a reflective type opto-sensor provides drum position information. It should be noted that as this drum will be rotating at 1500rpm, two pulses from this opto-sensor will occur for each NBTV frame, more on this later.

The usual holes or other arrangement for line sync pulses are not required as the motor that drives the drum is a brushless DC motor and is truly synchronous and drives the disc directly with no intervening gears or belts. (See Vol. 29. No. 1). There are three Hall-effect devices internal to the motor that can provide this function if required. (Output of Hall-effect device is 200Hz at 1500rpm)

A rigid base was folded out of 3mm aluminium and a cross-member added on the underside for stiffness, this also houses the minor amount of electronics that are needed to be located relatively close to the drum assembly.


.......there's a lot more but it's very far from complete...

In the photo you can just see the VCR head-drum motor supporting the main drum. The small bracket at the front was to mount the drum position sensor. The photo was just to record progress...but the monitor is now in storage somewhere!

Cheers,

Steve A.

[gary]

That looks great Steve, it will be a real shame if you don't get to finish it off!

I was wondering if you did the casting yourself? (I am in the process of setting up a small foundry myself for the purpose of melting aluminium and there are a number of ways to go, decisions, decisions, etc., so am on the lookout for tips and pointers).

Cheers,

Gary

[Steve Anderson]

Hi Gary,

No I didn't do the foundry work, it was sub-contracted out to a local company. The lathe-work was done by my brother-in-law, the sheetmetal bashing by myself. And in 3mm ally it's not that easy!

Steve A.