Construction Diary -- Part 2, Spinning the Nipkow Disc

[Steve Anderson]

For the packaged LED half... when I measured across the 680K resistor, I got nothing for the packaged type. When I measured from GND to the anode (+) pin of the transmittor LED, I got 12V. I considered this LED, then, to be blown (LEDs short, right, when they die).

Andrew!

UNITS!!!

R1 is 680 ohms, NOT 680k!! Another thing that frequently trips up the newcomer is the resistor colour-code. This takes practice, at this stage when you think you have a 680 OHM resistor in your hand, confirm it with the multimeter. Quite often the colours aren't very good and hard to tell the difference between red and orange for example. I'm assuming that your colour vision is OK.

The datasheet for the ZD1901 is a good example of a bad datasheet. The pinning arrangements are very ambiguous. I have no idea if LEDs go short or open when 'blown', I would have thought open.

My long-term goal is to redo all circuits with no ICs, and no prepackaged stuff. As many discrete components as possible, all 1920s stuff, with the exception of LEDs for which I'll make a concession because they're so useful

..all 1920s stuff....hmmm....this means no semiconductors, all valve/tube....and you're worried about working with 70V! Few valve/tube circuits work with less than 200V!

Steve A.

[Andrew Davie]

R1 is 680 ohms, NOT 680k!! Another thing that frequently trips up the newcomer is the resistor colour-code. This takes practice, at this stage when you think you have a 680 OHM resistor in your hand, confirm it with the multimeter. Quite often the colours aren't very good and hard to tell the difference between red and orange for example.

I DO use the multimeter! I can't tell the colours apart anyway; the red is as good as a brown in the light I work in, so I always use multimeter when measuring LED values. And yes, you are correct that I should be careful about units. But in this case, the mutimeter reading was 0.680M ohms, which I translated to 680K ohms. The real problem is I should be using a 680 ohm resistor, shouldn't I!!!

Thanks for pointing out something that I just wouldn't have spotted.

My long-term goal is to redo all circuits with no ICs, and no prepackaged stuff. As many discrete components as possible, all 1920s stuff, with the exception of LEDs for which I'll make a concession because they're so useful

..all 1920s stuff....hmmm....this means no semiconductors, all valve/tube....and you're worried about working with 70V! Few valve/tube circuits work with less than 200V!

Steve A.

Well, not valves... but ICs seems like 'cheating', wheras transistors seems OK.

Cheers
A

[Steve Anderson]

The reason I didn't start from square 1 was that I only have one prototype board, and it is totally occupied by the circuit. I didn't want to rip the whole thing out only to find it wasn't necessary. Bah.

It's best to strip the whole thing down and work from left-to-right, bit by bit confirming that each section is doing what it is supposed to. Building up the complete circuit and hoping it will work first time is asking for problems.

By the way, this circuit uses a chip...

Steve A.

[Andrew Davie]

It's best to strip the whole thing down and work from left-to-right, bit by bit confirming that each section is doing what it is supposed to. Building up the complete circuit and hoping it will work first time is asking for problems.

Although the Apollo program successfully used a concept called "all-up testing" which basically built the whole 500+ foot-tall rocket and fired it off without extensive subcomponent testing.... and it worked... my efforts to date at this sort of approach have been less than successful.

I will probably take your advice, start from scratch and build up the circuit bit by bit.

As to having a chip -- my goal has always been to get a working system first, then go back and change the things that I feel need changing. That's why I did the circuit with the LED driver and sync boards from the club, then when they were working I built my own single-board version. Once I have a *totally* working mechanical TV, I expect that I'll go back and do yet another version of it without ICs.

It's just that when I show my circuits to people and say "this is what they had for TV in the 1930s", people point straight to the ICs and say "no they didn't!".

[Steve Anderson]

Neither did they have transistors in the 1930s...

I'll admit I'm being somewhat antagonistic here...please excuse me...

Steve A.

[Andrew Davie]

OK, after a LOT of blown up IR components (yes, I did add your 1K resistor to the circuit), I now have a working IR transmitter/receiver circuit. I don't really know what I did wrong, I just know that I only have one IR transmit/receive combination that actually responds to an obstruction between the components. When interchanging with the 'dud' components, the circuit stops working, so I am assuming they are 'blown'.

What I see, though, is rather strange. Firstly, the angular position of the sensors is rather sensitive. When I put them right up against each other, though, here's what I see...

When detection is possible, measuring across the receiver, I see 0V. If I wiggle the transmitter slightly so that they do not directly point at each other, then this raises slightly -- going all the way up to 7V or so when they're quite out of alignment and far apart. Going back to directly aligned, and insert an opaque card between transmitter and receiver... then the voltage across the receiver is now 10V. So I effectively have 0V when IR is being detected, 10V when not detected or no transmitter connected.

So in a fashion, I'm confident that my last surviving IR xmitter/receiver pair are functional. Just not quite what I expected (I had assumed there would be voltage when detected, no voltage when not detected.

One odd thing I have noticed is a periodic 'glitch' in the signal when IR is being detected. Every 1ms, the received voltage plunges alternately from 10V down to (rougly) 0V or 7V over a 0.02ms period. That is, I see voltage at 10V across receiver... and 1ms later it briefly plummets to 0, then returns to 10V (in 0.02ms)... then 1ms later it briefly plummets to 7V or so (again, in 0.02ms). So on the scope I see a line on 10V, with two downward spikes, one big one about one third the size. I have no idea where these are coming from or what causes them. Should I be concerned about them?

In any case, I guess it's time to start putting the circuit back together again. By the way, not having a 680 ohm resistor... I built my own by placing a few of lesser value in series, ending up with a 674 ohm resistance.

[Andrew Davie]

I just can't seem to get this sync circuit to work. Mind you, I don't have the IR sender/receiver bracketing the Nipkow disk. My assumption is that with the "always on" with sender always seeing receiver, I would expect the motor to spin. In any case, when I block the signal, or jiggle it on/off, the motor doesn't spin either.

So, I really do need some more suggestions!

[Steve Anderson]

When detection is possible, measuring across the receiver, I see 0V. If I wiggle the transmitter slightly so that they do not directly point at each other, then this raises slightly -- going all the way up to 7V or so when they're quite out of alignment and far apart. Going back to directly aligned, and insert an opaque card between transmitter and receiver... then the voltage across the receiver is now 10V. So I effectively have 0V when IR is being detected, 10V when not detected or no transmitter connected.

Andrew,

OK, that is indeed progress! The current through the phototransistor is proportional to the amout of IR it receives, so as a generalization they should be in optical alignement and as close together as practicable. LEDs and phototransistors/photodiodes (both visible and IR) are made with different 'beamwidths', wide for a TV remote control (TX and RX) narrow for position detection. And in shades of grey in between.

In this application we need it to switch between 0V and (almost) 12V. With no obstruction the LED illuminates the phototransistor turning it on supplying most of the 12V supply to the following resistors. So what you're looking for is (almost) 12V with no obstruction and 0V with. This is with reference to 0V, see following...

A habit I suggest you get out of is measuring 'across' things. This is occasionally useful, but confusing to those reading text. Basically weld the negative lead of your multimeter to ground/earth/0V, just as you should be doing with the ground on your 'scope probe. (I hope). Always measure from ground to the specified point. Where a circuit has voltages on it, in one corner somewhere you'll see a note...'All voltages with reference to...(some reference, usually 0V).

One odd thing I have noticed is a periodic 'glitch' in the signal when IR is being detected. Every 1ms, the received voltage plunges alternately from 10V down to (rougly) 0V or 7V over a 0.02ms period. That is, I see voltage at 10V across receiver... and 1ms later it briefly plummets to 0, then returns to 10V (in 0.02ms)... then 1ms later it briefly plummets to 7V or so (again, in 0.02ms). So on the scope I see a line on 10V, with two downward spikes, one big one about one third the size. I have no idea where these are coming from or what causes them. Should I be concerned about them?

Yes, you should be concerned about them! Have you got the rest of the circuit connected? Are you looking at point 'A' on my simple diagram?

In any case, I guess it's time to start putting the circuit back together again. By the way, not having a 680 ohm resistor... I built my own by placing a few of lesser value in series, ending up with a 674 ohm resistance.

Any value between 470 ohm and 1k should be OK, you just need to pass a few mA through the LED, it's not that critical. But 680k...no way!

Steve A.

P.S. If you know you have a dud component, or even just 'suspect', throw it away! It'll come back to bite you on the bum in a few months when you will have forgotten it's dead. You'll end up wasting even more time.

[Steve Anderson]

Andrew,

I get the impression you've built up the whole thing again hoping now that it'll work...error, still most unlikely.

I know I said we'll work from left-to-right, but in this case I think it's better to get the motor control sorted, even just a manual version for now. We can add the feedback loop/PPL later.

First get your LED/phototransistor setup on the disc, make sure you're getting pulses of about 10V at point 'A' as you spin the disc by hand. Do this now, otherwise you're wasting your time. You'll have no idea of the discs' RPM otherwise.

Attached is the sub-circuit of the motor control. Build this up (only), I also attach a PDF of the IRF540. Pay paticular attention to the picture in the PDF as to the connections of the MOSFET. I have noted the pin numbers on the circuit. You'll notice the mounting tab is offset...get this correct!

This should give you manual control over the motor speed, it might be the full range or just part of it. It also might be quite 'cramped' in the middle of the pot, just a small tweak results in it going from stationary to full speed. Also, be aware there are (generally) two types of pot (for the panel-mounted versions). Linear and logarithmic. The linear ones are used almost exclusively everywhere, except in volume controls which are log. The circuit will still work with either sort but the log version will not be as 'smooth' in this application.

They are usually denoted by a letter after the value, e.g. 10kA is a log pot, a 10kB is linear. There are others but not often come across. Presets are always linear.

I also suggest you bolt the MOSFET to a piece of insulated metal, to perform its function correctly it needs to dissipate heat, Silicon doesn't like heat.

Let me know how you get on.

Steve A.

[Steve Anderson]

Andrew,

One other thing I cannot stress enough is for clean and stable power supplies. The supply for the LED/phototransistor and the motor should ideally be seperate. Brushed DC motors generate a lot of interference, if this gets coupled into power supply rail for the logic and phototransistor, you're asking for trouble. Once you get this going have a look at the drain of the MOSFET with a 'scope...you'll see a real mess!

What are your power supply arrangements?

Steve A.

[Andrew Davie]

What are your power supply arrangements?

I have a single transformer. From the AC off that (15V) I split to two separate rectifying circuits. One generates roughly 19V with ripple -- and this is used for powering the LED display and the NBTV circuits. Since the sync circuit (from memory) has a 12V thingy on it, I figured the power sharing would be OK here. The other generates 11.8V ripple-'free', and this is used to power the motor and the ground of the motor synchronisation circuit. So as far as I'm aware, there is separation between the various circuit sections.

I have today (after building the above circuit section) had a look at the power in just about all bits of the circuit (the motor one), and it looks pancake flat 12V pretty much everywhere. No sign of any interference.

[Andrew Davie]

I get the impression you've built up the whole thing again hoping now that it'll work...error, still most unlikely.

...
... no sir, no illicit circuits here....
... (long pause).... no-sir-eeee ...

Many thanks for your patient assistance.

First get your LED/phototransistor setup on the disc, make sure you're getting pulses of about 10V at point 'A' as you spin the disc by hand. Do this now, otherwise you're wasting your time. You'll have no idea of the discs' RPM otherwise.

I skipped this bit because I see that I can directly control the motor with the circuit you give. I'll come back to the above (I promise). The lure of getting some sort of control is too great...

Attached is the sub-circuit of the motor control. Build this up (only), I also attach a PDF of the IRF540. Pay paticular attention to the picture in the PDF as to the connections of the MOSFET. I have noted the pin numbers on the circuit. You'll notice the mounting tab is offset...get this correct!

I have built up this circuit as shown. I am absolutely totally sure I have the IRF540 the correct way, with the correct pins identified. Tab flat on the desk, pins towards me, numbering from left to right 1, 2, 3.

I am also absolutely totally sure that the motor just does not spin with this circuit as I have implemented it. I used a 10KB pot, and observe the voltage through this ranging from 0V to 12V, roughly, from one extreme to the other. It's possible I have the pot's leads connected incorrectly. I connected the center lead of the pot to R13, effectively. The other two I assume don't matter which way around.

There are some really strange things I don't understand. I'm powering this with 12V rectified and ripple-free. On the scope, just looking at the input power only, I see a very flat 12V 'signal'. If I connect one motor lead to the 12V positive, and the other to the 12V negative/gnd, the motor spins. So I know that the power supply bit to the motor is OK.

I also see pretty much 12V (varying in some places, if I turn the pot) on most places in the circuit, too. If I measure from gnd to the output of the circuit (just before it connects to the motor), I see 12V (when motor is connected). Hook the motor up to the output of the circuit (where D(2) and R12 meet), though, and it doesn't spin. Even though I 'see' 12V with the oscilloscope. What is happening here????

I note that the only way this circuit is seeing 12V at all is 'through' the motor -- it is that which is connected to the 12V... this circuit is only connected to GND of the same supply. I'm confused.

I also suggest you bolt the MOSFET to a piece of insulated metal, to perform its function correctly it needs to dissipate heat, Silicon doesn't like heat.

OK, I haven't done the above bit. I don't have the correct bits yet... I'll go out and buy a heatsink and bolt. But I suspect there's another problem, because I would expect the motor to at least try to spin, heatsink or no.

[Steve Anderson]

Andrew,

If you have wired it correctly, yet the motor doesn't run, measure the voltage with the multimeter at the gate of the MOSFET. (The junction of R13/C2). As you rotate the pot this should vary from about 1V to 11V. Do this again right on the pin (1) of the device. Check pin 3 is really connected to 0V (Ohm-meter).

Check there is 12V on the tab of the MOSFET (connected to pin 2, the drain). If this is the case the MOSFET is dead.

Alternatively you might find the wiper of the pot goes from 1V to 11V, (junction of C1/R13) but the gate doesn't. (Junction of R13/C2). Remove the MOSFET, check the junction of R13/C2 again. If this now goes from 1V to 11V then there's a short inside the MOSFET, pin 1 to 3. Dead again.

Remember, all voltages are referenced to 0V.

Steve A.

P.S. I've just had a look at the Jaycar site, they have the BUZ71 device listed, this is probably a better choice than the IRF540 which is a little too brutal, but it should still work.

[Andrew Davie]

If this is the case the MOSFET is dead.

OK, after reading your advice I took the simplest route and just assumed the MOSFET was dead. Just to be sure, I bought two more of them and when I plugged the new one in, voila! the circuit works! So, now I know that dead components can cause hair loss. As you said

There's a problem, though -- at full power, the motor is now not spinning fast enough. The 11.8V that my rectified supply was putting out was barely enough to run the motor fast enough -- and now that we're losing some voltage through the speed circuit, it's not enough.

I considered the options...

1) fiddling with the resistors in the circuit -- tried, and nothing really seemed to make much difference.

2) changing the diameter of the wheel on the end of the motor to a larger one -- tricky, because I don't have an easy replacement.

3) using a different power source.

So, keeping in mind the interference that will come from using the same power supply as is powering the NBTV circuits... I hooked up to that anyway, just to see. And it all works. Very well.

I can rotate the pot and control the image quite well -- never getting it completely stopped, but it rolls at a very slow rate -- perhaps 1 frame every 10 seconds or so, when I have it adjusted nicely. So, the bottom line is that this speed control part of the circuit is working.

I've been getting excellent pictures using this setup. I tend to judge the picture quality by what I can see in certain clips from the club CD (#1). Those are clip #44, Karen, where I see if I can see the face of the guy behind her right shoulder (on our left) just before the words THIS IS KAREN appears. If I can make out a face, which I can, I consider the picture to be quite good. Secondly, I use track #46 which has the teddy bear. If I can see teddy's fur texture, and a pattern on his bowtie (just), then I also consider it's a goood picture.

Now, using the single power supply (19V or so)... I get very good pictures, and excellent speed on my motor. I suspect that I'm not seeing problems with the circuit/interference because the 19V feeds firstly into the 12V rectifier on the NBTV board... and that's filtering out any signal problems...?

As to the availability of components at Jaycar -- they do not carry all that they list. You have to order some of them in -- doesn't take long, apparently, but when I'm there I typically want to walk out with something, not order something and come back.

The images attached here are pretty much for record/historical reasons, just so I can come back to this someday and remember what it looked like These images are all taken with the newly working manual speed control circuit, so image quality is indicative of what I'm seeing with the shared power source. They are all unenhanced, fairly indicative of the images I see with the eye.

[Klaas Robers]

Well done Andrew. This is comparable to what I see on my monitor. I still should try to make photographs of the picture for your reference. You happen to have taken that hurdle.