A Nipkow disc with no holes.

[Steve Anderson]

I’ve had this idea rolling around in my empty skull for some time. A Nipkow disc that doesn’t have holes. Well, not for scanning anyway.

The idea is to use a linear sensor array to scan the image. The TSL201R seems being ideal (pdf attached). However, if its active position is at 3 o’clock it scans across the lines. But if positioned either at 12 o’clock or 6 o’clock as shown in the sketch in red, this will scan an NBTV line of 64 pixels.

The smaller ellipse is a representation of a lens, a 50mm focal length lens from an old SLR camera would be ideal.

The output is analogue and needs to be ‘stretched’ as one complete frame would be scanned in 2.5mS.

There is also the issue of integration time. With a single sensor on a disk rotating at 750RPM there is very little time for the device to capture the ambient light upon it. It’s within spec. but I suspect that a great deal of light would be needed. Here that’s not a problem; we receive three times the solar energy that Europe does.

So to combat this I considered increasing the quantity of sensors to three which would reduce the speed of the disc to 250RPM and increase the integration time by a factor of three. Using three sensors spaced at 120 degrees would allow the use of AC synchronous motors that run at 250RPM (RS & Farnell). The problem is that the analogue signal still needs to be stretched.

To get a continuous output waveform that doesn’t need stretching would require a lot more sensors and doing some simple maths they would simply not fit next to each other, there’s just not enough space.

There is also the problem of getting the signal in and out of the devices as well. They need two inputs, the output plus power and ground. Slip rings are a possibility (yuk), the other is opto-coupling with power provided by batteries on the disc.

This is all very conceptual, and partly why I posted the ‘PCB discs’ thread some time ago. The electronics would be quite complex, but I think possible.

Steve A.

[AncientBrit]

Steve,

Nice idea.

An alternative to slip-rings or opto is magnetic coupling.

Probably 2 separate systems, one for power transfer at around 50kHz, the other for signal transfer on a carrier.

GL

[Steve Anderson]

An alternative to slip-rings or opto is magnetic coupling.

GL

Yes, I had considered using the rotary transformer within the VCR head drum assembly for that purpose, but I haven't pursued it further. Maybe in the near future.

Steve A.

"I have seen the evils of procrastination, and I vow to change my ways tomorrow."

[Steve Anderson]

Going on from the above here's another variation of the same idea.

Using a horizontally mounted drum would do the same job, but with the advantage(?) of no curvature in the lines or the 'keystone' distortion of the previous version. It would look rather like a cake-baking tin.

Your thoughts gents?

Steve A.

Thinking further if a single sensor was used or a small number of them the drum size could be reduced to about 50mm in diameter or less.

[Steve Anderson]

I did play around with the idea of a small drum. With a drum of 50mm diameter it is possible to get 16 of these devices around the circumfrence, but the internal wiring would be almost impossible.

So shown below is the plan view of a 50mm drum with eight devices mounted on it. It's drawn to scale and there's enough space for the wiring.

With eight sensors this would require a rather strange rotation speed of 93.75RPM.

The other reason for not using 16 devices is they are not cheap!

Steve A.

[AncientBrit]

Were it not for the loss of sensitivity you could change the concept to having a single fixed sensor.

Mounted in front would be a vertical rotating mirror that would "sweep" the image across the face of the device at frame rate.

All cabling would be static with no need for slip rings.

Not a new concept, already used in image recognition devices.

[AncientBrit]

On reflection I think you would also need a frame store since the frame scanning does not yield a continuous video stream.

[Steve Anderson]

On reflection I think you would also need a frame store since the frame scanning does not yield a continuous video stream.

Mounted in front would be a vertical rotating mirror that would "sweep" the image across the face of the device at frame rate.

Yes, agreed, it's what I have been trying to avoid (the frame store). But your idea of a static sensor sounds entirely plausable. Let me cogitate on it further.

It also requires just one sensor (far cheaper) and consistant output, no device-to-device variations.

Steve A.

[Steve Anderson]

I've done some hand-waving and even some AutoCad work on using Grahams suggestion. This would require a horizontally mounted drum and the mirrors would all be in the same physical orientation. The sensor array would be in the vertical plane.

The quantity of mirrors is not important as long as the next one comes along 80mS later. It's tempting to go for the fewest number of mirrors possible, but as the mirror swings by the distance from the optical centre of the lens to the sensor varies which means if the centre of the picture is in focus, the edges won't be, and the converse. It does depend on the depth of field of the lens, on the presumption that we want as much light as possible to reach the sensor I guess the aperture of the lens would be wide open

A sensible minimum quantity of mirrors might be ten, although 20 might be a better figure. With 20 mirrors the drum would only have to rotate at 37.5RPM. With 17 mirrors the RPM becomes 44.1. Using an old synchronous driven record player turntable could well be feasible, if driven with a mains frequency of 49Hz. It's also horizontal.

But as for doing the mirror calculations...it's doing my head in! Help!

Steve A.

[AncientBrit]

With the lower resolution of NBTV would variation of path length (and focus) as the mirror rotated be a problem ?

[Steve Anderson]

With the lower resolution of NBTV would variation of path length (and focus) as the mirror rotated be a problem ?

Quite possibly not, and no matter how many mirrors one employed it would never be 'perfect'. I used to have a copy of 'The Ilford Book of Photography', within which it explained how to calculate the depth of field, the circle of confusion and so forth for any lens.

But being the twit I am, I lent it to someone and never got it back.

I guess there's always the 'net to come to the rescue...

http://www.cambridgeincolour.com/tutori ... ulator.htm

Steve A.