Wavy horizontals

[Klaas Robers]

Albert, I just read this thread almost from the beginning, as we were in Germany memorating WWII.

Now you have ensured that the waviness is not coming from the disc, I have a completely different suggestion: LISTEN to the separated sync signal and LISTEN to the signal coming from the sync pick-up of the disc. You can do so using a cristal in-ear receiver, or with some form of LF amplifier and a loudspeaker or head phone.

You should hear a humming sound, the same that you hear from a black video signal (NBTV CD No 1 track 2, 3, 4). Listen to the video channel (Left) and get used to the sound.

Now listen to the separated sync in the same way. Assuming that you use the circuit from page 13 of the new bue Handbook, listen at the signal "sync out" of the LED-driver circuit. You should hear the same sound, which should be absolutely independant from the video content on all tracks of the CD. Be sure that you will hear all irregularities in the humming sound. If you have some potentiometer to adjust the sync level, try that and get the best humming sound you can get.

Then listen at the sync as picked up off the disc. That should be at pin 3 of the 4046. That should give you a comparable sound, however that is not really true. Unhappily you don't have access to the real signal as this is only available inside the 4046.

I hope this will help you.

[Viewmaster]

Thanks very much for that Klaas.
I do have a small oscilloscope so I can see the sync on both pins of the 4046. (3 and 14). Although, as you mentioned, I am using the sync seperator on page 13 of new handbook I am using your single transistor LED driver on page 18

The sync from the CD on pin 14 is sharp and steady as one would expect.
I derive the disc sync from a light and volt detector, not an opto fork. This is then sharpened and shortened with a 4098 multi vib, and again when the picture is steady the pulses are sharp and steady on pin 3.

I have rebuilt the sync detector and 4046 circuits but no improvement.......

BUT, concentrating on my new multi CD flywheel, the more the inertia the better the picture.
So, I put a very heavy 1920s 12 inch 78RPM gramophone record on the shaft to act as a flywheel and the picture is even better. (The time from switch off power to rest is 5 seconds with the gram record so the inertia is now improved from the original set up at about 2 seconds)

So I will now get a smaller flywheel, with most of the mass in the outer diameter, in order to decrease the flywheel from a huge 12 inches but keeping the same inertia.

The trouble with the Nipper is, that in itself, it has near zero inertia so a flywheel seems essential, but much larger than I first thought would be OK.

Thanks again Klaas.

[gary]

Albert, have you had a look at the waveform at the base of the motor driver transistor? One would think that to get that variation in motor speed that there must be a varying (12.5 Hz) signal there (ideally it should be a DC value). If so maybe all you really need to do is tweak the low pass filter.

[Klaas Robers]

Gary,

isn't it that a huge innertia of the disc shouldn't be needed if:

1. the sync pulses of the video signal (from CD) are properly separated? and

2. the reference pulses from the optical detector are stable and without jitter?

I think adding innertia to the disc might decrease the amount of waviness, but is not the problem, nor the solution.

However there is another problem, which is part of the frame sync. The Peter Smith circuit likes to synchronise as well with an offset of one line. This includes a kind of hic-up during the missing sync puls. It is first missing in one of both signals and then in the other. The motor cannot follow that but gives a certain accelleration + decelleration to the disc. This can cause the waviness, however I never saw this in the MUTR monitor. That has also a low mass disc.

[Viewmaster]

The DC volts at the base of the motor driver transistor is 3.5v
It has a small 400hZ sawtooth component on top of approx 30mV.

This can be removed by increasing the 100nF to 300nF (see sketch)
BUT it makes no difference to the picture.
Likewise, increasing the 47uF to 90 makes no difference either.

The shots from my scope show the waveform at the base.
You can see the 400hZ sawtooth. Those peaks at each frame freq are what is giving the motor a kick. They do vary in amplitude from picture to picture.
So the motor speed increases and then drops back again, thus giving the waviness.
A heavy flywheel helps to mitigate the kick effect I reckon, but doesn't explain what is causing the 4046 to kick the motor every frame.

edit......the sync wheel has the missing 31/32 frame pulse missing as required. I might change the sync wheel pattern just in case there is an small irregularity there which I cannot see on the 'scope..

[Viewmaster]

Although this will not cure the reason for the problem, I wondered if a 400kZ pass only filter would remove that 'kick' in the waveform and so let the motor run smoothly ?? Is there any circuit that removes kicks?

[gary]

Yes it does seem that it's the inability of the circuit to achieve "true" frame lock that is at the root of the problem, however it's really surprising that the inertia of the motor/disk/flywheel combination isn't greater that the 80ms of the 12.5 Hz pulse caused by the "out-by-one-line" phenomenon.

I suggest that it may be worth trying increasing the 100nF cap to, say, 1uF. That should remove most of the 12.5 hz component.

[Viewmaster]

Prior to my reading your latest, Gary, I decided to do something drastic this afternoon.....no, not hitting it with a hammer !!

I thought that the frame 'kick' might be linked to the missing 32 line pulse
although I haven't the faintest idea why.

So, I reinstated the missing No 32 line pulse on the sync wheel, so there is now 32 pulses per frame into pin 3 of the 4046
(although pin 13 from the sync seperator still has the missing pulse from the CD)

The waviness has nearly ceased on different pictures.

The previous large positive going pulse, I showed you on the 'scope, at each frame at the transistor base, has disappeared.

There is now only a very small negative pulse at each frame.
So the motor is no longer getting that 'kick' at each frame.

ALSO, and I am surprised at this, the picture locks OK with 32 pulses per frame from the sync wheel.....

....I just had track 44 running for 3 minutes with continue frame lock, and using my original flywheel.
As Klaas said, a big flywheel should not be necessary.

Tomorrow, I shall try that 1uF idea too, although as I mentioned, a 300nF C does remove the 400hZ.

[Klaas Robers]

Indeed Gary, it is the "not in frame-sync" that causes the waviness. If I see Picture 2 180 then I see that the disc has a frame offset of 4 lines!! And I see the waviness also in the gate voltage after the pedestal. The motor is powefull enough to follow that.

What Albert could do is to disconnect the synchronisation for a short moment, e.g. by a push button and wait for a new synchronisation until there is no pedestal any more in the gate voltage. At that moment the missing sync pulses coincide and the waviness should be absent. This is a trial an error method, but simple.

I once have been thinking if there is an easy way to include or delete just one sync pulse from the video sync, e.g. by two push buttons, a + and a -. Then you do frame synchronisation more or less by hand but it is more controllable. If the circuit is unable to do that automatically (and it is), then we should use other means, e.g. the two push buttons.

In my disc monitor I obtained frame sync by one separate pulse of the disc per revolution, which should coincide with the missing sync pulse. If that is not the case the sync-loop is opened and the disc runs on its own. Yes, that is quite some extra circuitry, but it works, with zero lines of offset.

[Viewmaster]

As my NBTV picture is shown on a monitor, I could have a smaller size of picture and have a moving frame on a rack and pinion which could be raised or lowered to fit the picture on the monitor...... Just like the mechanical framing on film projectors of old, where the film gate could be moved up/down to frame the picture.

But I still have 2 things to try yet. Gary's 1uF to replace the 100nF and Klaas's push button just refered to.

Also the 200k pot (replaces the 2 x 100k fixed Rs) does allow quite a bit of up/down framing before sync is lost

But at least I know now how to get rid of the waviness,
albeit rather cheating the 4046.

[gary]

I have to admit that I actually cheat and use software to advance or retard a line in order to frame up. In the rare cases where my source coming from CD or similar I use the well tried finger technique.

[Viewmaster]

I have put the push to break switch in the o/p of the 4046 as Klaas suggeted. I put the push switch in a seperate remote box for ease of use.
(The long unscreened leads do not seem to bother the system working.)

..........Just fancy, the Nipper now has a remote control.

The picture can now be quickly shifted sideways to frame correctly.
This might be very unorthodox but it works for me.

So I am putting it all back together (once again!). Will video it all working and give the link on the main Nipper thread in due course.

Thanks to all for useful suggestions etc during this saga, especially Gary, for hanging in on there for me since the beginning of this thread.