TV horizontal scanning - why flyback?
 

Pondering the makeup of an old-style analog TV signal, I wondered:

Why was the line-scan done by scanning a line then doing a hard 'flyback' ? It would have made more sense to me if the spot scanned alternately left to right and then right to left, so no need for the dumping of vast amounts of magnetically-stored energy from the line-scan coils. Swinging a spot from side to side would surely have been more efficient than doing a sawtooth?

The only downside I can see for this 'alternate' scan would be that by the time the spot got back to the side of the screen it started from, the phosphor glow would have diminished more so it could cause alternate lines to have slightly different brightnesses?

Re: TV horizontal scanning - why flyback?
 

There would be no space for the line sync pulses which occur during flyback time, so a different method of synchronisation would need to be used. Also there'd be no 'back porch' and so no black level reference. So the overall picture black-levels would be all over the place!

Re: TV horizontal scanning - why flyback?
 

I would have thought that this Boustrophedon type of scanning method would also complicate the waveform to the scanning system, requiring sawtooths alternately of a rising and falling form.

Re: TV horizontal scanning - why flyback?
 

Apart from the sync and black level problems already mentioned the raster would consist of a series of zig zag lines rather than a series of parallel scans.

Re: TV horizontal scanning - why flyback?
 

I think we have to view this as to what was possible when the TV system was invented. To use a scan from side to side would have required the removal of interlaced scanning. This was introduced for two reasons, one flicker reduction and the other bandwidth reduction of the vision signal.

The bandwidth was a problem in that components for wide bandwidth and high gain were not easily available and would have raised the cost of transmission and reception equipment.

The side to side scanning would not reduce the problem of flicker, a higher repetition rate could have been used but that would complicate equipment design I.e more bandwidth and increase costs.

Just my thoughts, of course I could be completely wrong.

Re: TV horizontal scanning - why flyback?
 

Hadn't thought about the back-porch issue; must admit, I've been thinking about this having recently done battle with a dot-matrix printer where the printing was bi-directional rather than doing-it-like-a-flyback-TV, and it just seemed such an obvious way of delivering a 'raster' that I wondered why it wasn't done!

Perhaps the original TV-standard designers also made some underlying acknowledgement to the difficulties of doing anything 'intelligent' while still accomodating baird/Scophony-style mechanical scanning? A Boustrephodon spinning-mirror scanner or Jeffree cell would be an intriguing animal to have to design!

Another thought: do the line- and frame-scans have to be linear? I'm thinking here that since most 'action' in a TV (or movie) image tends to take place in the centre of the screen - which is where the viewer's eyes naturally settle - a 'smart' TV system could use non-linear scanning in order to present greater detail in the centre of the screen at the expense of that at the outer edges.

Re: TV horizontal scanning - why flyback?
 

I think it would over-complicate things and, again, there'd be synchronisation issues: you'd need defined points of resynchronisation or a sort of sync 'frequency-hopping' as the image parameters change.

And you'd have to keep stagecoach wagon-wheels in one part of the screen, especially when viewed from the corner of one's eye! Unless you used a varying gradation of phosphors on the tube.

Re: TV horizontal scanning - why flyback?
 

Incompatible with mechanical scanning systems?

Re: TV horizontal scanning - why flyback?
 

I suppose what you could do is an Apollo 13 excercise.

Only have what they had in 1936.

Then see what was possible.

And bear in mind valves were expensive so you couldn't throw unlimited bottles at the solution.

I suspect PJL's suggestion about mechanical scanning was a factor should be taken into account.

Cheers

Mike T

Re: TV horizontal scanning - why flyback?
 

Surely the main reason is that there would be no chance to get the "pixels" to line up vertically.

By sweeping each line in the same way, whatever the non-linearities of the scan it would at least be the same in each line.
By scanning alternate lines backwards the scan would have to be astonishingly linear and accurate otherwise large horizontal discrepancies would be very obvious between adjacent lines.

I am sure this is overwhelmingly the reason. I think it would even be hard to make a CRT system work that way properly with modern technology.

Re: TV horizontal scanning - why flyback?
 

Any delay in sync to the scan running shifts existing technlogy pictures a little sideways but doesn't do any damage.

with reciprocating scand, alternate lines get shifted right and left confusing the picture. MUch more noticeable.

David

Re: TV horizontal scanning - why flyback?
 

I tried to do this years ago on a new design of linear scan (power amplifier driven) flight simulator monitor. It was a very high resolution 26" delta-gun job with no expense spared. The snag was that a hybrid power amplifier driving the line scan very precisely still needed fast flyback to get it back to the start of the line within the blanking period.

There were three solutions. One was a high voltage asymmetric supply power amplifier capable of carrying out the flyback voltage excursion. The second was an 'injector' whereby a high voltage was switched in during flyback and the third was 'zig-zag' scan, as it became known, or Boustrophedon if you're posh.

The injected higher voltage needed for flyback wasn't too bad as the coils were low inductance although we were running 64 KHz line scan, but we still decided to try zig-zag scan. What a rabbit hole that turned out to be!

Firstly, you need a staircase field scan otherwise the lines cross in the middle. That took a while to organise but with digits it wasn't too bad. Then we thought that we would be in business by programming the video source (an IG - Image Generator) to display backwards/forwards video. Oh what a mess! The timing was all up the shoot and so the digital video had to be fed into a line store with adjustable timing. Almost there now, but it wouldn't line up left to right - you could have one or the other - a sort of linearity issue. That turned out to be a magnetics issue in the deflection coil but we could just about correct that with a switched linearity corrector. Boy, this was getting complicated, but it now looked good on green. Then we switched on blue and red and realised that although we had frigged the convergence waveforms for zig-zag, all our fancy zoned digital corrections didn't work anymore and the convergence was really complicated to sort out. Might have been OK on a PIL tube, but what a shebang on a delta gun. In the end we had to abandon it and go for the switched high voltage solution, but I didn't half learn a lot along the way!

Zig-zag? No thanks!

Best regards,

Paul M

Re: TV horizontal scanning - why flyback?
 

Not to mention the complications at the origination end - cameras etc etc - not a runner!

Re: TV horizontal scanning - why flyback?
 

Actually, Baird's 30-line system did just that, except that with vertical lines top to bottom he did it by increasing the width and spacing of the outer lines.

Re: TV horizontal scanning - why flyback?
 

There were proposals for 'each-way' scanning using vibrating mirrors. I believe Dénes Mihály had a go.

Re: TV horizontal scanning - why flyback?
 

It would have been a nightmare to align the end of one scan line with the beginning of the next.

At least if the electron beam is always moving in the same direction, any delay as the signal works its way through the circuit will shift the image in the same direction. If the beam reversed on each line, then "later" would be to the right when travelling left to right, but to the left when travelling right to left. The resulting distortion would be a nasty "tearing" effect.

A dot matrix printer can print text in both directions, but this is because your eyes are less sensitive to vertical misalignment of evenly spaced rows of horizontal text -- even if it's monospaced. But most dot-matrix printers switch to one-way mode for graphics to improve the vertical alignment between dot positions. I am sure my old Citizen used to let you force it to print graphics in two-way mode, but warned against this in case of unsatisfactory results.

(It also had a test mode in which it just printed rows and rows of the letter H, for testing the horizontal alignment. There was a potentiometer which presumably set the threshhold for an optical sensor, and thus the position within the beam width at which it knew to start printing on the backward pass.)

Re: TV horizontal scanning - why flyback?
 

I think the advantages of sawtooth raster scan are too great to improve on for analog representation of an image on a CRT.

In the case of the H scan, the flyback is so rapid that the beam energy delivered to the phosphor is very very low. Normally only the H scan line is seen, not the H retrace, though the H retrace is blanked in some professional monitors and computer monitors.

The V retrace is not fast enough to be effectively self blanked as it is slower and ideally the TV set has retrace blanking.

I think though the linear sweep for H & V, with a flyback, was simply a "logical and practical way" to evenly distribute the picture elements over time with even illumination on the CRT face.

Then it turned out that the H retrace interval and the L & C energy changes associated with it, had "huge utility value" in generating auxiliary voltages and EHT and in energy recovery scanning. The vertical scan & its flyback was less helpful in this respect. Also, in very early CRT TV's, a sawtooth wave was easy to generate.

So while it could "work" using a zig-zag scanning method for the H scan, for example a triangle wave with flat tops & bottoms to allow for sync and porch time, the utility value of the H flyback energy changes would be lost and nothing else obvious would be gained in its place to make it worth it. Also the time bases would be more difficult. It would still then require a vertical sawtooth scan to get the beam back to the start.

I have attached an interesting oscillogram (photo of CRT face) of how a raster scan is constructed...can anyone guess how this image was created ?

Re: TV horizontal scanning - why flyback?
 

Hmm... I'll take an intutive go at it..

Sampling the current in the deflection yoke on one axis of the scope, and the current in the vertical deflection yoke on the other?

Is it showing hysteresis?

Re: TV horizontal scanning - why flyback?
 

Vertical = frame scan current.
Horizontal = some mix of frame and vertical scan currents + some odd/even field component from somewhere ? Or, vertical scan added to a ramp oscillator running at field rate / 2 (if your scope allows modulation of x-axis then regular scope scan could be used)

Whatever, quite neat !

TTFN,
Jon

Re: TV horizontal scanning - why flyback?
 

Oops, forgot the important qualifying word!