Back in the Game!

[Ralph]

The original goal of this particular project was to construct a Nipkow televisor (my first, if you ignore the trivial task of assembling a MUTR "kit") that would provide a realistic appreciation of the trials, tribulations, and occasional triumphs of those late 20's and early 30's experimenters building basic equipment with manual speed control. There is no doubt that, with the addition of the simple "string braking" system, an enthusiast could have watched and enjoyed the relatively short programming intervals of the period. Note that I said an "enthusiast"! To get spouse and family to pay attention would have required a high-end televisor with the ability to lock the display automatically.

The final steps to operational status with this project involved some simple modifications:

(1) The previous post reviewed the solution to the "diffuser problem", achieved by providing adequate space between the LED's and the diffuser material.

(2) Increasing the light intensity was a simple matter of the black set bias control of Klaas's one transistor LED driver. I had set the potentiometer to the excitation point of the LED's while observing the LED cluster directly. In fact, the proper way to do it is with the diffuser in place and the the disc running. If the black level is set that way, there is a lot more light available with normal video drive.

(3) With a bit of experience, the simple string braking system for framing and phasing is so effective that it is possible to take longer exposures, providing better photographs of the display. The image below shows an earlier image of my friend Bugs Bunny (A) - great focus, but with mottling as a result of the diffuser issue and barely adequate light for the single-frame photograph - 1/12.5th of a second. (B) shows the current situation with the new diffuser arrangement, a brighter image as a result of the bias change, and a 1/2.5-sec. multi-frame exposure (an even 5-frames).

Although the unit is now operational, the inside of the light-box is far from elegant, with PC-board pieces from two earlier iterations lashed together in a distressingly informal way! By the end of this week what will probably be the last contribution to this thread will be a new integrated pcb for which I will provide documentation and board-layout data.

Ralph

[Ralph]

Well Phase 1 is finally done and the televisor is completely operational! The first image below shows the third (and last!) iteration of the circuit board for the light-box. By the time I had the diffuser problem solved, the light-box was a rat's nest of wiring with pieces of two previous boards. It was fifty years ago that I first thought abut building a Nipkow-disc televisor and it has taken that long to get around to doing it. I wanted everything built to a high standard, so the point of the third board was to eliminated the "mess" in the box.

The circuit board (A below) has all the active circuits and is mounted on the inside of the lid of the project box that serves as the light-box. With the "lid" facing to the rear (unlike earlier versions), the box can be opened for tests, adjustment, or servicing without the need to remove the scanning disc. The BRIGHTNESS potentiometer and 4-conductor POWER/VIDEO connector are just below the circuit board. The lower half of the board incorporates Klaas's one-transistor LED driver (the BUZ11 mosfet is near the center of the board), with the 12-LED array occupying the upper half. In this version there are four series strings, each with three white LEDs. I was able to include an additional LED in each string as I am using a 24VDC power supply. This also let me eliminate a 20V regulator that was used in the earlier versions.

The board is designed to be divided into independent video and LED modules for remote mounting, but is kept in one piece for this project. The main part of the project box is completely empty except for the framing mask and diffuser on the side facing the disc. The exterior of the light-box is shown in image B.

The second picture shows the completed televior. Although a casual look would suggest that is simply the televisor mechanics, all the necessary circuits, with the exception of the 12V supply for the motor an the 24V supply for the light-box, are unobtrusively built-in. Nice bright pictures, easily viewed in normal room lighting. If I want to simulate a neon light source, I can simply drop a suitable orange filter between the light box and the disc.

It has been fun!

Ralph

[Panrock]

A first rate effort. Well done!

Steve O

[Ralph]

Thank you Steve! I am much more comfortable designing circuits and writing code, but the mechanical engineering and fabrication of the apparatus is one of the fascinating elements of this mania we all share.....

Ralph

[AncientBrit]

Well done Ralph,
Very nice construction.

What's next?

(No pressure there then !!)

Cheers

Graham

[Ralph]

Graham,

What next is the really big question at the moment. I suspect the first thing will be the design of a more-or-less universal precision motor drive. Got an idea that would be an interesting combination of the familiar and unfamiliar, but I will hold off talking about it until I can check out a few ideas.

Assuming I succeed in the first endeavor, the next project will be a mirror screw. I have the Club mirror screw kit and will have, after the Nipkow adventure, all of the essential circuit elements. All I will need to find is the patience to polish the mirrors!

Ralph

[AncientBrit]

Ralph,

The precision drive sounds very interesting.
I wish you luck.

Thanks for a very informative set of articles.
A very good 'read'.

Kind regards,

Graham

[Ralph]

The Nipkow televisor is officially complete and along the way I have had the chance to look a a variety of ways to control the Nipkow disc speed.

Basic PWM DC Speed Control

This was the track documented in this thread and as of interest to me because that is how so many early enthusiasts tried to view pictures. It can work, particularly with the addition of a simple "string brake", but it is tedious and labor intensive in terms of operation.

Synchronous Motor Simulation

Synchronous motors were widely employed on my side of the Atlantic "back in the day", but suitable motors are hard to find and quite expensive. I was able to simulate synchronous operation with my DC motor by using a portion of Peter Smith's motor control circuit, replacing the NBTV sync pulse input with drive from a crystal oscillator/divider chain, providing precise 750 RPM speed control at much higher torque than would have been possible with a real synchronous motor (see the Motor Seed Control Issue thread). This required adding an optical sensor to track disc speed. I chose to try a basic reflective system (see A in the second image below). Initially it would not lock properly, but Gary and Steve A suggested that slow rise and fall-times from the optical encoder might be the problem. Adding Schmitt triggers to both reference inputs to the 4046 solved the problem. It works quite well and led to additional experiments with using precision computer generated tones to supply both the basic sped reference and the ability to do precise framing and phasing. The results of those tests will probably be submitted as a short piece for the Newsletter.

Conventional Sync Triggering

Given that I had already added an optical encoder to the system, it was a modest step to go the Smith/Club speed control board with a few minor modifications. My board is shown in image B below. Since there is no really convenient way to derive sync from Klass' one-transistor video/LED driver, I used an LM538 dual op-amp to implement DC-restoration followed by a sync stripper. It is non-critical in terms of setup and, as expected, works very nicely. Th sample of NBTV images below includes four from my ROMScanner NBTV source and the the test pattern from Darryl Hock's standards converter.

The televisor is performing very well and I am pleased with both the clarity and grayscale of the images. I can also demonstrate it in term of manual speed control, synchronous operation, and sync-triggered modes.

I may have gotten a bit carried away in terms of tinkering, but I now have a complete set of modules to use with my still-to-be-polished mirror screw.

Ralph

[AncientBrit]

Hi Ralph,

Very interesting to follow your developments.

As a slight aside I note you have been experimenting with motor control.

I've been playing around with an aperture drum which is mounted on a 2.5mm ali backing plate.

Out of interested I offered up a pair of magnets as a 'saddle' to the edge of the disc.
The magnets were in in assistance, ie N pole facing a S pole, and hand held.

The induced eddy currents were sufficient to appreciably slow down the disc speed.

A single magnet had hardly any effect.

Might be of future use.

Kind regards,

Graham

[Ralph]

Graham,

I have used eddy currents to damp the response of my seismograph, but never considered possible applications for NBTV. That's a topic deserving of some thought.....

Ralph

[gary]

In principle, it ought to be possible to rotate an aluminium Nipkow disk using eddy currents (as in an electrical meter). I haven't heard of anyone trying this method so perhaps there are some physical limitations that prevent it working effectively at high RPM (such as windage, friction, etc.) that I am unaware of. It is a compelling idea that has been on my backburner for a long time.

[AncientBrit]

@Ralph,
I suppose you could replace the permanent magnets with a pair of electromagnets if electrical control (rather than mechanical) was adopted for speed control.

@Gary,
Like you I have many projects on back burners!
Interesting idea about using an eddy current motor.
I suspect it wouldn't be powerful enough for larger discs/drums.

Kind regards,

Graham

[Ralph]

Given the fact that the next project would be a mirror screw, I decided to follow through with an integrated controller that would work with both the Nipkow and mirror screw televisors. Given that all the circuits were thoroughly tested and debugged, I felt this would be a trivial extra step.

The only thing new that I needed was a simple video preamp. Klaas's LED driver has been working flawlessly, but I could never get quite enough gain to optomize the images from the Club CD's. The "simple" solution was a one stage non-inverting amplifier (max. gain ~6). that would have the CONTRAST pot at the input and which would drive Klaas's circuit. The summary of this arrangement is shown below. To give an idea of the total context, the controller interfaces to the televisor mechanics with a 4-conductor cable:

1 - Ground
2 - LED-
3 - Motor-
4 - Optical sensor output

The controller has its own 12V supply and the televisor has high-current 24V (LED) and 12V (Motor) supplies, both referenced to the common/system ground.

The system fired up and worked perfectly for 3-4 minutes. Great adjustable contrast for any of the sources in my inventory, great motor speed control, and flawless video. After a few minutes I rapidly lost dynamic range and the LM358 op-amp stage failed. The other stage in the chip (properly terminated because it was unused) was just fine, as was the LM358 op-amp used for the sync-stripper.

I have tried additional bypassing, resistive loading of inputs and outputs, all with no effect - the buffer op-amp fails after a few minutes operation. There is never anything on the video output to suggest that the stage is oscillating or is handling any other spurious signals and all the other IC's on the board (including a total of three other op-am stages) are unaffected.

Op-amps and I have had an amicable relationship over the years, but this is a puzzle. I suspect the problem may be something simple/fundamental, but my aging brain cells have run out of ideas. Any suggestions?

Ralph

[Steve Anderson]

Ralph, you've omitted any biasing to the +ve input of the op-amp. A resistive divider to produce +6V, each (say) 220k, one from +12V to the junction of the 1uF (I assume, 1mF is actually 1000uF) and the 10k resistor, the other from there to 0V/ground.

It works for a while only due to the very small bias current charging the 1uF cap, once charged it falls over.

You will also need to put a cap in series with the 10k resistor (pin 6) say 10uF.

Steve A.

[Ralph]

Hi Steve!

I had finally come to the conclusion that the capacitor alone was simply charging the + input and, between the value of the capacitor, the 10K resistor, and the impedance of the input circuit, that was setting the time required to build up a fatal voltage at the + input. The Texas Instrument application notes suggest a 100K to 1 meg resistor from the + input to ground to stop that from happening.

I am trying to figure out your suggestion. If I were using an inverting amp I would put a voltage divider at the + input to center the output swing when using a single-ended supply. However, I have never seen that done with a non-inverting amplifier. An why a capacitor in series with the 10K resistor to the - input? If we have a non-inverting amplifier with a voltage gain of 6, it seems to me the output will lock up at the positive rail with the 6V at the + input, irrespective of the AC input.

The amplifier as shown delivers excellent video until the last few seconds when the + input fails. I have run out of op-amps at the start of our Memorial Day weekend so experiments must wait until next week. It looks like a simple resistor from the - input to ground should solve the problem, but in the interim I would like to understand your ideas with respect to biasing the stage.

As far as the capacitor values, I'm "old school" on the western side of the Atlantic:

1 mF = 1000 nF
0.1 mF = 100 nF
0.01 mF = 10 nF

Cheers,

Ralph