Well, here we go -- here I hope to record the things that happen as I embark on my attempt to build a mechanical TV. I should start by pointing out that I have absolutely no electronics skills whatsoever -- I'm very much a software guy, so this will be interesting.
I have been reading everything I could find on the 'net, and in particular I will be referring to the excellent construction tutorial on Peter Yanczer's site, along with the information on the NBTV.ORG website in the 'Handbook' section.
I have already purchased a Nipkow disk from Peter, along with a hub as shown on his website, and I believe I'm getting a motor and drive belts along with it. I'm based in Australia, but these parts are in the USA at the moment. I expect to see them in a month... or three.
There's a gorgeous mechanical TV that I'm using as my 'fondest desire' target. If I can make anything like this, I'll be thrilled...
The above by Alan Rosser, as recently reported on the NBTV site.
Here's how I understand the system will work, in summary.
A CD player providing stereo input via an RCA jack, with the left channel holding NBTV 32-line video, and the right channel holding sound. I'll use a club CD for test patterns initially, and develop my own software for conversion of arbitrary video later.
The sound channel is fed through a separate sound amplifier controlled by a volume knob.
The video signal consists of synch pulses of 0.3V 'height' (and please bear with me, I know nothing about electronics), merged with an actual picture signal that ranges from the top of the synch pulse at 0.3V up to 1.0V. That is, the range of the video itself is 0.7V.
The synch pulses are stripped out, or separately handled -- these are used to synchronise the speed of the Nipkow disc (which has 32 holes around the circumference to assist with timing). A separate IR detects these holes and generates a separate signal that is somehow compared with the synch pulse from the video signal and that comparison is used to either speed up or slow down the motor that drives the Nipkow disk.
I do have a desire at some stage to have colour, but I'm thinking that at my first attempt I should just stick with a stock-standard greyscale monitor.
OK, so the video signal (with the synch pulses removed?) is amplified and this is then used to drive an array of so painfully bright that you can't even look at them LEDs ganged together in an array. It's the scanning of the holes in the Nipkow disk over this pulsing LED array that gives us our actual pictures.
Now I've seen a lot of stuff written about the LED array, and it seems that as few as 6 ultra bright LEDs are sufficient. I note that the light needs to be diffuse, so these LEDs have a diffusion barrier (glass/plastic??) between them and the Nipkow disc.
My first step, of course, is to understand just a little electronics so I can build myself a reasonable light array. I'm OK with the idea that LEDs run on current, and that there's a voltage drop across each LED -- so I need to make a decision based on the rating of the LEDs I use that will mean that I don't exceed the power supplied. See how I'm randomly pulling up electronics terms? This is because I don't really have a clue.
Step 1. Understanding how power works...
I'm going to need to provide power that I can't kill myself with. Various sources I've read suggest that 15V or 12V are reasonable output ratings for transformers suitable for this project. I have handy access to computer power supplies, and they output at least one 12V supply, probably more. So my first decision is to use a readymade computer power supply, which will be something like 200W.
If that's a bad idea, please stop me right there before I do any damage!
I will need to power the sound amplification circuit (and speaker that I assume is powered from the same circuit's output?), the video amplification (and I assume the LEDs are powered again from that circuit's output?) and of course the motor control/synch circuit (again, the motor being controlled from that?).
Finally, once that is all working (easy, right?) I'll need to have a pretty case, much as the above. I've already scouted out a few woodworking/cabinetmaker friends so that will be fairly straightforward, I think. I'll make a barebones first, then worry about the pretty bits.
Now, looking at
http://www.nbtv.wyenet.co.uk/syncsep.htm this seems to me to be something that I'd like to use. What I understand this will do is that it will amplify the video part of the input signal (although I'm a bit confused because the club standard suggests to me that a standard signal has a black to white range of 0.7V, yet this circuit block diagram says 3x amplification and has an output of 1.4V (ie: 2x the standard). So there's something I"m not understanding there (and what's pp? pulse to pulse?).
But, assuming that it magically amplifies the input signal and there's a simple video out on it, there are two other ouptuts -- a line and clamp pulse, and a frame pulse. Obviously the line and frame pulse are used to synchronise the disc rotation with the signal -- another circuit I'll get to later. And the clamp pulse? I think I read something about this being used to generate true blacks (like giving the base-level of black, perhaps???). Ok, that will have to wait till later, too.
But it looks to me like I'm going to want this particular board, to start with, right? A pity it's unpopulated, as I'm sure my soldering skills are pretty much like trying to solder with a hot housewife-type-iron.
THe next thing that looks like I might be interested in is
http://www.nbtv.wyenet.co.uk/vidamp.htm a video amplifier circuit. Clearly this is intended to be connected to the 'video separator' circuit, and I note with the block diagram it also seems to amplify the signal by a further 2x (giving 2.7V not 2.8 -- can't electronics circuit designers multiply?
) I have just noticed in the block diagram on the above page, the synch pulses still seem to be there in the signal, so the LED array is actually going to 'display' these, and trying to understand what I'm seeing here, it looks to me like the 'clamping' gets fed to the transistor at the right, which seems to merge the clamped value and the video value before driving the LED array.
I note the LED array is 'powered' at 100V, so presumably we can whack some pretty large number of LEDs (or power hungry) at this point. Clearly I need to understand LEDs a bit better. I do think I understand that if you string them together in serial, you sum up the voltage for each, and you have to provide enough V to 'get through' the whole string, or none will light. I also think I understand that you have to put a limiting resistor in series with the LED string, so that they aren't fried by excessive power. I guess I just don't understand what's the real-world difference between a string of LEDs with a single resistor, and a parallel array of LEDs, with a resistor for each. The LEDs still light up, right?
Anyway, more ignorance later....
So I'm pretty sure I need to get the above two circuit boards (but I'm a bit concerned that now I'm dealing with 100V, not 12, so need to really investigate what this means in reality).
That leaves the speed control circuit, and the sound amplifier.
I would have to assume that the sound amplifier is a pretty stock-standard electronics thing, and that I can find that pretty much anywhere.
And the big thing I've been thinking about (particulalry since the article in the last NBTV newsletter about 'Phase Locked Line Synch' seems to suggest that synchronisation is by no means a simple issue) is the synchronisation of the disk rotation with the signal.
I'm aware that the standard technique is to mask off one of the synch holes on the Nipkow disk, and that corresponds to the frame synch. So the idea, I guess, is that we have (from the synch circuit above) a line synch corresponding to each one of the signals from the holes in the disk, and a frame synch corresponding to the absence of a synch hole in the expected position on the disk. Somehow this has to 'fire up' in a state that will cause the motor to start, regardless of what position it is in, and once started get some idea of if it should speed up or slow down. Interesting. I guess I'd do it with some sort of state machine, if I were programming. I wonder how it's done in electronics. More stuff to learn!
OK, that will do for now. Please please feel free to comment on the above. Obviously my understanding will be riddled with mistakes, but one does have to start somewhere.
[/url]