Convert a B&W TV to display 1080i?

[Whirled One]

From the pie-in-the-sky department:

Today I happened to be thinking about that new Coby HD ATSC tuner that I posted about over in the 'Solid State' forum (http://www.audiokarma.org/forums/sho...d.php?t=237955), and it occured to me-- I wonder if it might be feasible to convert an old B&W TV to display in HD, using one 'channel' of the analog component outputs on an HD ATSC tuner (or other HD video device, such as a Blue-Ray player) as a video source. If an old TV could be converted, a B&W set would be better suited than a color set for a variety of reasons. Among other things-- (a) the CRT resolution isn't limited by a shadow mask and phosphor pattern (as it is on a color CRT), and (b) the 2nd anode voltage is less "fussy" on a B&W CRT. Plus, you don't have to worry about converting colors from standard component outputs to RGB, or do any other goofing around with color decoding.

The vertical sweep is no problem, since it's still 60 Hz at 1080i or 720p. The horizontal sweep frequency would have to be more than doubled, however, for displaying 1080i (and even more so for 720p). Even if you could get the horiz oscillator running at the correct frequency, I wonder how badly the HV output would suffer, since that's generated off of the horiz oscillator. Would a totally different flyback be necessary? Of course, you could always either add in a separate high voltage power supply just to provide the 2nd anode voltage to the CRT, or leave the entire original horizontal section alone (and just use it for producing the HV supply), and add a new, separate, horizontal osciallator circuit to provide the actual video sweep.

Another thought-- I'd imagine this would require a rather complex scan converter, but what if a 1080i signal could be converted to a sort of "quad" interlace format? In other words, leave the field rate at 60 Hz, but put 4 interlaced fields per frame instead of 2? You'd have to add in some time to allow for the extra vertical retrace periods though. If this could be done though, that would cut the displayable horizontal lines per field down to 540, so even if you add in the non-displayed vertical retrace period, you're still almost certainly going to be well within the range that a standard 525-line NTSC TV can produce (and lock-in) just with the service controls.

Another thing-- what's the usual video bandwidth on a 1080i analog component signal? It'd be one thing to just get the scan rates right, but how much vertical resolution can one really eeke out of a typical B&W TV, even when bypassing the RF/IF stages?

[wa2ise]

Back in 1979, I worked at a company that made devices to convert video signal from hospital CAT scanners to film hard copy. Basically a lens system projecting video from a CRT faceplate onto film. 'Twas a standard issue job at a typical company that made a boring product. Most were 525i black and white, but we did a few special jobs for 525p and maybe higher (it's been a while....). We were able to hack most of the horiz deflection to run at the higher frequency (mostly RC time constants in the Conracs we used). The flyback period tended to be an issue, in that the horiz interval would be shorter for 525p than 525i, but the capacitance loading of the high voltage winding on the flyback would show down how fast the deflection could get back to the other side of the CRT screen. This would result in overscan, as the left and right edges of the image would be clipped. The duty cycle of the flyback pulse vs the electron beam writing the video on the CRT screen would change some, but that wasn't too big a problem.

[peverett]

Before working on this, a check of the CRT bandwidth is in order. I would not bet on an old standard TV B&W CRT being able to display 1080i.

[zenith2134]

Yeah, fat chance . . . by the way, this confirms my suspicions that wa2ise is a genius of the hobby....wow, man. That is pretty complex experimentation!

[jeyurkon]

wa2ise is an invaluable source. His web pages have been very helpful.

I don't think the bandwidth of the CRT is as much of an issue as trying to get the CRT focused well enough. The bandwidth is limited by the grid capacitance. The higher frequency would require a change in the video output circuit. It needs to have a lower output impedance to drive it at the higher frequency.

John

[old_tv_nut]

vidoe bandwidth needs ot be about 35 MHz, but you cna cheat it down to about 20 MHz without too much notieeable effect. The H sweep is more of a problem. playing with a 5-inch monitor that doesn't stress the HOT too much is not as hard as a large screen where components are run close to their limit.

Some HD systems were proposed that used spot-wobble to resolve more "pixels" on the screen. Dot intelace the fields, then wobble the beam up a quarter line on one dot and down a quarter line on the next dot. [EDIT - you would reverse the phase of the wobble in the next frame, to get all the spots filled in in 2 frames] Of course, you would need all the video memory to create the dot-interlaced fields. this does not give full HD resolution on diagonals, but does on vertical and horizontal lines. There was even a system proposed used a 6-field sequence if I recall correctly. I think some private demos were made, but it was never brought out as hardware for proof of concept, and I suspect it had motion and/or dot-crawl problems especially when played back on a non-wobbled legacyTV.

[miniman82]

My brother wants to convert my Philco to 640x480 computer usage just for kicks. I'll see if I can get him to comment on the HD thing. My guess is he'll say a complete redesign of the horizontal section would be required, including custom flyback (or maybe one hacked from another source).

[peverett]

Remember that most of the old B&W crts were designed for a 4 to 6 MHZ video signal, not anywhere near 20 MHZ and up.

[electroking]

Originally Posted by peverett

Remember that most of the old B&W crts were designed for a 4 to 6 MHZ video signal, not anywhere near 20 MHZ and up.

Is that limitation actually caused by the CRT itself, or by the conventional
video amplifier topologies, as well as the fact that the audio carrier
is 'only' 4.5 MHz above the video carrier? French 819-line TVs,
with their 14-MHz wide channels, probably used conventional CRTs.

[miniman82]

Tim says:

"Yeah, I don't know if I'd want to do that with the existing flyback, if even the yoke. Might be possible to use newer parts removed from a high resolution monitor (nice if the yoke fits!). To make things easier, I would put in triggered oscillators (instead of free-running locked oscillators that don't have as much range) and current amplifiers to drive the yokes. HV is a seperate supply.

It would be a bit of a mess, but easier to work on than a cheapified TV circuit, and also lets you use it as a slow X-Y display. (Add a high speed sampler circuit and you can make a 1GHz+ sampling oscilloscope.)

Tim"


I had a feeling that's what he'd say. It does make more sense to use newer technology, instead of hacking the original stuff. Now let's see someone do it!

[peverett]

France is France and the US is the US. I am sure that the makers of the US TV grade B&W CRTs did not go to any extra cost to increase TV CRT bandwidth beyond what was necessary.

Check the specs before wasting your time!

[wa2ise]

Many sets drove the CRT cathode, and one way to work around the capacitance from the cathode to the heater would be to introduce some inductance in the heater supply.

I suppose, as mentioned in a post above, you could use the chassis (circuit board) from a computer VGA monitor along with its deflection yoke. The HV may be too high for a B&W tube from the 50's. 20KV vs 10 to 15 KV? You might be able to kludge up a 6BK4 HV shunt regulator and a dropping resistor (one that can take a few KV across it!) to get the HV to something the B&W tube would run on. Just be very careful!

You'd use one of the RGB video channel amplifiers to drive the B&W tube.

The electron spot size in a 1950's B&W CRT may be too big to make the 1080i picture look any better than just the old fashioned 525i from a converter box. You could still make a sharper picture by injecting, at the video detector diode point in the IF circuit, the luma signal from a converter box that has luma and chroma separate, like the Channel Master CM7000. Most B&W sets that were made after NTSC color came out (1954) had their video IF strip's response have a depressed amplitude around the chroma subcarrier frequency. You still need to pass the sound subcarrier though. If you have TV alignment equipment, you might be able to tweak that dip out.

As for the sound amplifier, many TVs of the era used a quadrature FM detector, which yielded audio of enough amplitude to drive the audio output tube (via a volume control) directly. You'd need to build a audio driver stage to take line level audio from the converter box to make it strong enough to drive the audio output tube.

[jeyurkon]

Originally Posted by wa2ise

Many sets drove the CRT cathode, and one way to work around the capacitance from the cathode to the heater would be to introduce some inductance in the heater supply...

Oops, you're right. Nearly all sets do. Why do I keep forgetting that?

I also keep forgetting that the grid is a fairly large massive structure. The G1 to other electrodes capacitance for an 12LP4 is 6 pF. For the cathode it's 5 pF. I wonder if they don't include the heater in that rating?

At 20 MHZ that's a reactive load of about 1,600 ohms. Not too hard to drive.

John

[old_tv_nut]

Originally Posted by miniman82

Tim says:

"Yeah, I don't know if I'd want to do that with the existing flyback, if even the yoke. Might be possible to use newer parts removed from a high resolution monitor (nice if the yoke fits!). To make things easier, I would put in triggered oscillators (instead of free-running locked oscillators that don't have as much range) and current amplifiers to drive the yokes. HV is a seperate supply.

It would be a bit of a mess, but easier to work on than a cheapified TV circuit, and also lets you use it as a slow X-Y display. (Add a high speed sampler circuit and you can make a 1GHz+ sampling oscilloscope.)

Tim"


I had a feeling that's what he'd say. It does make more sense to use newer technology, instead of hacking the original stuff. Now let's see someone do it!

Flyback horizontal sweep/HV supplies can be fried (excess current) if the oscillator frequency gets too low or stops. Does anyone recall the early IBM PC color monitors that would die if they were run with the input disconnected?
Usng a limited-range oscillator has its purposes.

[jeyurkon]

Originally Posted by old_tv_nut

Flyback horizontal sweep/HV supplies can be fried (excess current) if the oscillator frequency gets too low or stops. Does anyone recall the early IBM PC color monitors that would die if they were run with the input disconnected?
Usng a limited-range oscillator has its purposes.

Yes, you could even kill them with a program to lower the sweep sync.

John