Wobbly Spots.

[Steve Anderson]

Thought I'd start a thread on spot-wobble aside from the CRT thread.

I found the following on Johnz Valve Page..

http://www.thevalvepage.com/

An Experimental Spot-Wobble Circuit

The following is cribbed from a pamphlet entitled "Mullard Valve Notes No.1".

The line structure of a television picture is an unpleasant feature, and is emphasized to the point at which it may become intolerable as the focus becomes sharper. The high horizontal definition achieved by good focus can be retained and at the same time the line structure removed by elongating the spot vertically. The most convenient way of doing this is by deflecting the spot vertically at a fequency which is high compared with the line frequency. This is called "spot wobble".

The best circuit arrangement for producing spot wobble depends upon the picture tube used, its associated components and the layout of the receiver. A good starting point for experiments in this field is provided by the circuit described below. It should be appreciated, however, that if the interlace is not good, or if the spot is astigmatic, or if there is appreciable deflection defocussing, spot wobble is unlikely to effect any considerable improvements. These points should therefore be looked into before attempting to apply spot wobble.

The diagram shows a simple oscillator in which the spot-wobble deflector coils form part of the oscillatory circuit. An EF80 pentode is employed, connected as a triode, and the total drain on a 180 volt H.T. line is about 12mA.

The amplitude control R1 is used to adjust the elongation of the spot to the condition in which the lines just merge. The switch SW1 permits the spot wobble to be switched off while the normal focus is being adjusted. The spot-wobble deflector coils consist of a pair of saddle windings similar to conventional deflector coils. There is, however, no yoke, and the windings are much smaller, each coil consisting of five turns of 0.018" diameter (26 S.W.G.) enamelled copper wire. These should be wound on a rectangular former 1 1/8" x 1 3/4". When removed from the former the flat winding is applied to the tube by folding the longer sides round the neck, i.e., with the 1 1/8" wide parallel to the axis of the tube. The two coils are mounted on opposite sides of the tube neck directly behind the normal deflection coils and are connected in series in such a way that their magnetic fields assist each other.

PRACTICAL HINTS

As explained overleaf, the spot-wobble coils should be mounted on the tube immediately behind the normal deflector coils. This means that part of the spot-wobble coils will have to be inside the focus and centering assembly, and the method of mounting them must permit this.

A simple temporary arrangement is to make a sliding a former from two or three turns of Empire cloth 1½ inch wide to fit the tube neck, and to bend the spot-wobble winding over this while it is on the tube, retaining the windings in position with two turns of 1½ inch wide P.V.C. adhesive tape.

The leads from the spot-wobble coils to the oscillator should not be longer that 6 inches.
CHOICE OF OSCILLATOR FREQUENCY

The spot-wobble frequency should be greater than 8 Mc/s, otherwise, with a finely focused spot, the sinusoidal pattern becomes apparent. The frequency should ideally be chosen so that neither this frequency nor its harmonics fall within the I.F. or R.F. channel of the receiver. If the whole of band 1 and all the I.F. channels that ar in use are taken in consideration, the lowest ideal frequency available for the spot-wobble circuit is 40 Mc/s. The losses at this frequency are prohibitively great, and a nominal frequency of 12.5 Mc/s has been selected. Provision has been made for adjustment of this frequency by means of C1 so that harmonics can be moved out of the receiver channels.

The original artical appeared on the page immediately following an artical entitled "Important trends in valve design" describing the EF80 valve. Was this artical therfore just Mullard trying hard to flog a few more EF80's ?

In the March 1952 edition of Practical Television magazine, Ekco announced a new 15 inch table model employing this feature but I have not seen any other manufacturer adopt this.

The following excert is from the 1960/1961 Television Servicing book

The larger the screen, the more noticeable becomes the line structure of the British 405-line picture. The 'spot wobble' system of reducing the effects of line structure is not found in any current models: not because the system is ineffective, but simply because many viewers tended to equate clearly visible lines with a good picture and were often reluctant to use the H.F. oscillator. A new method of reducing line structure is under development on the Continent, consisting of a special transparent plastics panel arranged to have the optical effect of elongating the spot so that it appears ellipsoidal with successive lines just touching".

I hope it's of use.

Steve A.

[dominicbeesley]

Thanks for that article Steve and thanks for starting a seperate thread, the other was getting a bit cluttered...

From what I've heard from other sources the spot wobbles were pretty unpopular when fitted to 405 line sets and often disabled or turned off...I wonder was anything like this done on any of the huge NTSC CRT sets in the states?

I think I might well follow the article's lead and add extra spot wobble windings to the tube rather than try and get the field amplifier and coils to pass high frequencies....

Using a sine wave gives a line doubling effect as the sine wave spends most of its time at the extremeties of its travel. Also, if the lines overlap this leads to a set of very bright thin lines on top of grey or if they don't touch very dark thin lines between grey. Both are however an improvement but not by a huge amount except where the focus is pin sharp and the gaps between lines is large.

Any non-linearity in the frame drive means that you get a combination of the two problems in different parts of the picture - see attached photos... here the scan is fairly linear but the response to the HF is compromised at the start / end of frame due to the way the feedback works.

In the photos the spot wobble is a fairly distorted sine wave - more like a gentle ramp, this fills in between the lines nicely but gives a sharp error if there is any over/underlap. The middle picture has it more or less spot-on in the middle but wrong at the edges. [Frm left to right: too much overlap; about correct in the centre of the picture; too little]

In my CRT experiments so far the sine/ramp wave wobble gives a pretty good improvement over no wobble on a tightly focused picture filling in the gaps between the lines whilst preserving the along the line resolution.

However if a non-sine wave can be generated that gives a linear falling off in intensity can be found then the wobble can be used not just to fill in the gaps but to smooth between lines. This is shown in the simulation picture below. Once I've got the CRT working a lot bigger, better and without nearly so much smoke and stink from the transistors I hope to get something like the simulation.

Subjectively the smoothed picture looks blurred, but this is just because the jagged transitions between lines are removed (which impart no reliable information). The smoothed picture shows just as much detail and in longer viewing sessions is far easier on the eye - or at least I think so!

Cheers

Dom

[dominicbeesley]

Well after much messing with the circuitry I can get a full height more or less linear drive to both sets of coils giving a reasonable picture.

Unfortunately I'm stimied when it comes to driving the spot wobble. The inductance of the coils (about 5mH with about 3R series resistance) is such that I need a large drive at low MHz. To make much impact I need a fairly high voltage and this totally overloads everything in the feedback loop!

My only option seems to be to go for a much larger supply voltage to give me more headroom, at which point I'd prefer to move to valves but I don't think I'd manage to get them to deliver such low impedence drive without a huge (expensive) transformer.

I tried the idea of auxilliary coils as described in the Mullard article above but no joy, they just don't seem to make any inroads on the tube at all, I'm guessing that the ferrite former and quite hefty windings on this little tube give out a lot more oomph than the lower deflection angles in the days of the article?

There's not really anywhere to put them on the tiny tube neck either and dismantling the yoke and adding more windings looks like it would be a fiddle too far even for me!

Any ideas? One idea was to try winding something up on the little ferrite E-cores from a computer PSU driver transformer but no idea whether this would work at the low MHz needed or how well this would affect the beam over its whole scan...or where to start!

It would be a shame to give up now after all this work. I've attached the latest version of my V-drive circuit, now transistor only. This could still do with better linearity but its getting there! [I've certainly learned quite a bit about NFB and amplifier design in the past few days, though I'm sure I'm still doing it all wrong!]

Dom