Construction Diary -- Part 2, Spinning the Nipkow Disc

[Steve Anderson]

Andrew,

I think Klaas and I need an update on how exactly the circuit you are using is configured, can you re-do the drawing so we can get a handle on it?

Steve A.

Klaas, I believe you're the designer of this circuit, so perhaps it's better that I hand over to you here. The opto-fork output does look somewhat grotty to me, might be an opto-mechanical problem rather than electronic.

Back to Andrew...

When you rotate the disc slowly by hand does point 'A' go at least close to 12V?

[Klaas Robers]

Steve, I am not the designer of the circuit. Pete Smith is the designer. I also never built the circuit, in my monitor I made one of my own, which is not really stable, I still have some slow changes in vertical position of the picture. But it has automatic frame synchronisation in a different way compared to this circuit. I can't understand how this circuit obtains frame sync, but Pete says it does within + or - one line (it can sync one line shifted).

I think Andrew is still working on the circuit below, which is the circuit as published on the NBTVA Handbook-site.

Andrew, with the 47k pot you should be able to control the voltage on pin 3 of the IC to cross the "half supply voltage" with clear margins when the sync holes are passing the opto fork. See that you have equal deviations above and below the 6 volt. So say from 2 volt to 10 volt with a supply voltage of 12 volt. If this is obtained there is no need to change that setting any more. It should not be critical.

It might be a valuable idea to connect a small capacitor from pin 3 to "ground", say 1 nF, when the voltage on pin 3 is somewhat ragged because there are long wires from pin 3 going to the opto fork.

[Andrew Davie]

The above is the one I have been using, with slight modifications (two 22K resistors in series with the 47K pot). It is interesting to compare the above with the original, from Peter Yanczer's site.

[Klaas Robers]

Andrew, I think the C3 value of 100pF is better. This is an unused output and I don't know why Pete connected it with a capacitor to ground. May be it gave some form of spikes.

You can change the value of the 47k pot to 100k or even 220k. Then the range is larger, you can get higher "bumps" on pin 3. The value depends on the type of opto fork and the positioning on both sides of the sync holes. If your voltages are too low a higher resistance is easier.

Do you know in the mean time how much volage your motor needs to spin the disc to 12,5 rps?

In Newsletter Vol. 32 No. 2 Garth Porter writes about his experiences with the speedcontroller of Peter Smith. I hope you have that issue. As far as I understand it even Garth is having troubles to get the disc locked. This is much easier for the 30 line situation, where no missing sync pulses are present. Pete Smith is mainly working with that situations.

I think this is a sensible way out for you:

1. Insert a hand operable switch in the line from IC pin 13 going to the 100k resistor.
2. With the switch opened you should be able to control the disc speed using the 10k pot.
3. Use here a potentiometer with knob. If you can get a multiturn (helipot) this is better, because you have better "grip" on the speed.
4. Adjust the this pot to the optimal speed, so a still standing frame in the correct position in the viewing window.

I think it is best that you measure then about 6 volts on the gate of the IRF in respect to the "ground". This is again half the supply voltage of the 4046 IC. In that case the 4046 can pull the voltage up and down by the same amount. This is advantageous for the control function of the 4046.

5. Then close the switch. The disc should keep this position or move slightly up or down. If that happens replace the opto fork slightly to a better position.

Unhappily you won't have automatic frame sync and I can't predict what happens if you close the switch earlier. Some members say their disc is automatically finding frame sync. Garth suggests that he will make additional circuits for it, but hasn't given yet any diagrams. I did it in a different way.

[Klaas Robers]

Andrew, I made a new circuitdiagram for you. This includes an update of Pete Smith from Newsletter 26 / 1 and a suggestion of Garth Porter for an adjustable damping of the control loop with a trimming potentiometer of 220k in stead of the two resistors of 100k.

Proceed as follows:

1. Open the jumper or even better the switch.
2. Adjust the motor to the correct speed on video signal or stroboscope,
3. Place the slider of the 220k pot in the most left position (to 10k pot)
4. Wait until the indicator LED extinguishes, indicating frame sync
5. Close the jumper or switch.
6. Turn the slider of the 220k pot stepwise to the direction of the 4046
7. The disc should lock to the video.
8. Touch the disc with a finger to slow it down somewhat
9. Look to the resynchronisation of the picture.
10. Readjust the 220k pot for optimal damping: The picture should refind its correct position in the viewing window with almost no overshoot.

Once the 220k pot has been set correctly only steps 1 to 5 are needed to get obtain synchronisation.

[Tex]

Hello – I suppose the first thing I should do is to introduce myself. I'm a 61 year old retired electrical engineer living in the Dallas, Texas, USA area. Until a couple of weeks ago I had never heard of NBTV, but Andrew posted a problem he needed solved on a slide rule forum of which I am a member (I collect slide rules). I was able to solve his problem, but he also posted a link to this forum and I started lurking here and have learned a bit about NBTV. It's been very interesting, though I don't have the time at present to construct my own NB TV set. For the past week or so, I have been trying to help Andrew with his motor synchronization problem, unfortunately without much success. He invited me to join this forum, but I initially declined, as I had no intention of building my own set; however, in the interest of helping to solve his problem, I concluded that it would be helpful to get the private discussions I've been having with him out into the open on this forum so that others may comment.
It's been many years since I've done any analog circuit design, but it seems to me that with the circuit he is using, he should be able to get the motor to run with the 4046 IC totally disconnected, and should be able to vary the speed with the 10K pot. I have even gone so far as to model the right hand side of the circuit (from the 100n capacitor rightward) using LTSPICE (and thanks to whoever mentioned that – I've been having some fun playing with it). I've even come up with a way to model a DC motor, though of course I can't model Andrew's motor, since I don't know any of the motor parameters. Using parameters from a small DC servo motor for which I do know the parameters, and taking an educated guess at the load moment of inertia and dynamic friction (and assuming I converted correctly from oz.-inches to Nm ), LTSPICE gives results which seem reasonable to me and indicate that the motor should run without the 4046. I really don't understand why adjusting the 10K pot doesn't seem to affect anything. The only thing I can suggest at this point is that perhaps using a higher voltage for the motor supply would help. Also, I'm not totally sure about his power supply scheme – what kind of regulator is being used to supply 12 volts for the motor and what is the current capacity?
Anyway, I am about out of ideas – maybe someone else can point to a new direction toward the solution of this problem?

Steve Treadwell

[Steve Anderson]

Anyway, I am about out of ideas – maybe someone else can point to a new direction toward the solution of this problem?

Welcome (another) Steve,

It's quite simple in my mind, I just wouldn't even have considered a DC brushed motor. OK, there are those who have had a moderate amount of success with them, but inherently they are not easy to control in an application such as this, plus the interference they generate requires careful consideration.

Additionally with the (usual) reduction of RPM by belts and pulleys results in even more errors creeping in.

The use of a bycycle dynamo 'motor' has been suggested by others, and I know that this works. I published an item in the newsletter a few years back about using a VCR head-drum motor. The advantage of both of these is that no sync holes need to be drilled in the disc, no opto-fork is required, they follow the instructions given to them (the driving waveform).

Sadly few have even considered these options as they are looking for 'simple' circuits. 'Simple' circuits yield 'simple' results. You end up 'simply' pulling your hair out.

Steve A.

[Tex]

Hi Steve,
Thanks for the welcome. I haven't had a lot of experience with controlling motors, but given that what Andrew has is a DC motor, maybe it's worth a little more effort to try to make it work before junking it. I have been trying to think of why in the world the 10K pot. has no effect, while the 47K (or whatever it is now) does. One crazy thing that occurs to me is maybe the 47 uF capacitor is really leaky and is bleeding off enough current to keep the gate voltage too low to turn on the IRF540 unless the 4046 is outputting high pulses. Maybe it's worth trying to swap out the 47 uF cap?

Steve Treadwell

[Klaas Robers]

Hi Tex,

welcome here. Firstly I would like to know from Andrew what the voltage is that his motor needs to run the correct speed of 12,5 rps. Then he can select his supply voltage.

When the jumper is opened the circuit is a voltage feed-back circuit. With the 10k pot in the center position there should be 12 volt at the drain of the IRF. So 12 volt is too low a voltage for the motor supply in this case. With the slider in the center position he needs a stabilised voltage of 12 volt plus the voltage that the motor needs.

Indeed the speed then should be settable with the 10k pot. I fear that Andrew made a misconnection, because the 10k pot doesn't do anything. He should first make the right part of the circuit working properly, before he can connect the 4046.

DC-motors (having a permanent magnetic field, they are called here PM-motors) run an almost constant speed when the applied voltage is constant. The speed is also proportional with the voltage. That is, when they are in the idle state, so no heavy disc on the axle. In fact it should be such that the speed should be almost the same with and without disc mounted, when the same voltage is applied. This is a thing that Andrew should check first.

The real problem is that the disc should not only run at the correct speed, but also the position of the holes at a certain moment should be correct. So also the "phase" of rotation should be correct. There should be made a Phase locked loop, not only a frequency locked loop, or spped control.

The problem of synchronous motors, like bike dynamo's, is that, although they run the speed of the frequency applied, their "phase" is not stable, there is always some hunting visible in the displayed picture. This hunting is undamped, unless you can influence the phase of the driving AC in anti-phase with the hunting, in order to damp this. So anyway you need a "position" pick-up in the form of holes in the disc and an opto-fork or some other means and a differentiating feedback.

Not so easy.

[Andrew Davie]

Andrew, I made a new circuitdiagram for you.

Many thanks for the effort involved in this redesign, Klaas. I bought the components today, and wired it all up. The results are very interesting.

Firstly, the 10K pot now spins the motor faster and slower. I could not, however, get anything reasonable to a good speed lock -- it was either too fast, or too slow.... adjusting even a tinyest bit sent it into wild gyrations in the other direction.

After a while I started to think about why this was happening. Firstly, I suspected the 10K pot. It does seem to have a position which is discontinuous -- right at the point where the motor would be spinning at exactly the right speed.

But then I had a look at the rubber belt driving the disk from the motor. And it was in bad shape. Rips and tears on most of its surface, and this is entirely due, I am very sure, to the figure-eight loop that it is running to spin the disk in the correct direction. In fact, I think although this method is very clever, it is really NOT the right thing to do. The added friction (which was variable, because the surface of the band was all ripped/uneven) made it next to impossible for the disk to be spun at a constant speed.

So, I removed that belt and replaced with a rubber band -- with no figure-eight loop. Now my images are upside down -- that is OK for now; I will rotate the motor 180 degrees to fix that up later.

After finding that problem, I now find that it is much easier to control the speed of the motor. I can get it so that it is very very slowly 'rolling' the image. That is to say, I still can't get it stationary... but I can get it to the point where it rolls (say) one frame every 20 seconds.

The LED is installed as you have instructed, and it does indeed "go out" but only for a very short time, when the roll happens to be in the right place. We're talking half a second-short. Otherwise it appears to be flickering at high frequency, reasonably brightly. There's a definite 'dark/off' moment, though, as I mentioned.

Since I could not stabilise the picture manually, I had to chance following the rest of the instructions to see what happened. And, in short, nothing happened. There was no real effect, though once in a while I do think I saw a 'kick' in the image where it appeared to rapidly skip ahead a frame... but that could just be my imagination.

I could not find a 220K variable resistor with multiple turns, so instead I placed a 100K resistor in series with a 100K variable resistor. I think this will have been OK? There appears to be little effect when all is connected and I rotate the screw on this. I'm curious about the replacement of the two 100K resistors with the single 220K, because the wiring for the original circuit had the 100nF capacitor going to their connection. I assume the new 220nF capacitor is equivalent when connected as shown.

At this point I'm wondering if my motor is just so inaccurately controlled (that is, it responds so dramatically to voltage changes either because of the motor, or the rubber band pulley arrangement, or even because the end-bit on the motor around which the band wraps is off-center) that this is the reason that the sync circuit just doesn't get a chance.

One other thing, there was not enough voltage when driving the motor with 12V -- the motor would not run fast enough. I powered the motor instead from the 20V supply that is also driving the LEDs. This does not have ripple removed. The motor now spins fast enough (and as an interesting side-effect, the voltage regulator on my 12V power supply no longer gets super-hot). I'd be very happy to continue with this current power configuration -- a separate 12V circuit powering the motor driver circuit, and the 20V that powers the LEDs also powering the motor itself.

So, that's where things are at now. Obviously a lot better because I'm able to control the motor speed with the correct pot -- and to some degree able to get a steady picture. BUt still, no signs of synchronisation.

I did check the synch pulse and the IR pulse early on (but this was when I had two 22K resistors in series with the 47K (well, 50K) pot. At that point, I was essentially seeing 12V on each. After I removed the resistors so that I was compliant with the circuit diagram, I now only see 3.8V as teh IR receiver signal. I understand this is not good enough, but even so -- I don't get synch when they are there.

Once the 220k pot has been set correctly only steps 1 to 5 are needed to get obtain synchronisation.

Are you saying that I will need to do steps 1 to 5 every time I turn on the monitor? I thought this was a circuit that would do the synchrnoisation automatically... once it was working, it was just a matter of turning on the monitor, and the picture would synch...?

[Tex]

Klaas - thanks for the welcome.

Andrew and others - there is something I see on the CD4046B data sheet which concerns me. Under "Electical Characteristics at 25 C, there is the line: "Input Rise or Fall Times, tr, ts, Comparator Input, Term. 3", followed by maximum values, depending on the supply voltage:
Supply = 5v - tr, ts = 50 us max.
Supply = 10v - tr, ts = 1 us max.
Supply = 15v - tr, ts = 0.3 us max.

If I'm interpreting this correctly, it says to me that for the 4046 to be guaranteed to work properly, the input at pin 3 must have a rise or fall time less than the above. It's understandable there would be such a restriction, since phase comparator 2 is edge-triggered. From looking at Andrew's oscilloscope traces, it appears his signal from the phototransistor does not meet these requirements. Could this be the problem in obtaining sync lock?

It is also interesting that I couldn't find a similar restriction in the datasheet for the MC14046B, the Motorola equivalent to the CD4046B. I don't know if that means the 14046 doesn't have the restriction or if they just left it off the datasheet.

Steve Treadwell

[Klaas Robers]

Andrew,

good to see that things are reacting anyway now in the right way. Have you measured the voltage on the drain of the transistor or at the centre tap of the 10k pot? It should be about 6V when the motor is running the correct speed.

I think that you may also use a 100k pot at the position of the 220k pot in stead. To begin with place it in the centre position for the first experiments.

Yes, it is likely that the LED goes out for a short period when the disc is in the correct position. At that moment you should flip the switch on.

I am unsure whether the disc will synchronise completely automatically. The different people in the Newsletter don't agree about it. It is not impossible, but nobody gives any form of guarantee. Try it and you will know it. Anyway you should give the disc the correct speed and then switch in the lock circuit. I have to do that too with my monitor.

Tex made a good observation. It might indeed be usefull to place a double inverter in the line from the optofork to the 4046. Andrew, do you have a spare 4001 or 4011 lying around? Then I will make you a add-on circuit with this extra gate IC. The point is that it might be that the 4046 doesn't like the "hump"-like slow voltage as an input, but needs a fast and sharp pulse form. This can be made from the "hump" with the aid of such a circuit. Very simple and checkable with your oscilloscope.

[Roland]

The use of a bycycle dynamo 'motor' has been suggested by others, and I know that this works. I published an item in the newsletter a few years back about using a VCR head-drum motor. The advantage of both of these is that no sync holes need to be drilled in the disc, no opto-fork is required, they follow the instructions given to them (the driving waveform).

Sadly few have even considered these options as they are looking for 'simple' circuits. 'Simple' circuits yield 'simple' results. You end up 'simply' pulling your hair out.

I managed to aquire an old bicycle dynamo at the weekend for just this experiment. I have a spare 12" club disk - but I was just wondering if you had any hints/tips for getting it working?

I'm not currently concerned with synchronisation - just getting the disk to keep to a fixed speed.

Thanks

Roland.

[Steve Anderson]

Roland,

As I recall (and this is going back a long time) these dynamos generated about six Volts AC at a 'normal' bicycle speed. To use them 'in reverse' you'll need to supply it with about the same, perhaps a little less.

I would start with a mains transformer with several secondary taps on it, say 3V, 4.5V and 6V rated at about 1A. Try each tap and find where the best voltage is. The lowest might result in it not running at all, the highest might make the dynamo run very hot.

If you have a Variac this would be better, used on the primary of the transformer.

Like car electrics they usually used the frame of the bike as a ground return so it might appear that it only has one terminal, the other being the case. You'll need to spin the shaft by hand to get it running, but once going it should continue to run.

Once running, it would be useful to know the applied Voltage (i.e. measure it) and AC current. Also the DC resistance value, this will probably be very low, well under one ohm.

If you have an old audio amplifier kicking around, earmark it for use as the driver of the dynamo in the future. If not, a simple audio amp can be built up, it doesn't need to be Hi-Fi.

Steve A.

Postscript...I've done a bit of 'Googling' and it appears that some more modern dynamos use internal Zener diodes to regulate their output voltage to stop those that pedal fast from blowing their bulbs on the bike. If your dynamo has these, they need to be removed.

[Andrew Davie]

Andrew,

good to see that things are reacting anyway now in the right way. Have you measured the voltage on the drain of the transistor or at the centre tap of the 10k pot? It should be about 6V when the motor is running the correct speed.

When I measured the drain of the transistor recently, it was 6V.

Andrew, do you have a spare 4001 or 4011 lying around?

I do now! I'm eagerly awaiting further instruction