Veni, Vidi Video, Vici

So far in “The Custom T/S,” I’ve shown you how to overcome problems with all three of the little inline jacks on the left side of your computer, with fixes that handle trouble with power input and tape-recorder functions. There’s only one other connection (other than the rear edge-connector) left to cause any grief; the TV output.

Actually, this one is the least likely of all four to be the cause for concern, so perhaps it’s just as well that left it for last. If you get a LOUSY picture, chances are that the problem is with your TY. If you try to use an older tube-type or hybrid (tubes + transistors) TV, you probably won’t be too happy with what you see. I’ve had fine results with the Sanyo (also sold under the Sears brand-name) 12″ solid-state model selling for around $80. This one and others like it, by Panasonic, Samsung (the same company Sinclair Ltd. is currently dealing with), Tatung, and their various name-brands, probably represent the best buys for a “monitor” that will also double as a TV.

For dedicated computer use though, you’d probably be happier with a nice green or amber-screen “computer monitor” since these can be found in about the same price range. You may have gotten a good TV picture with your basic machine/RAMpack only, but find that the picture gets dirtier as you add other peripherals, Unfortunately, there is no jack provided for a video monitor on the ZX81/TS1000. It won’t work if you try to use the TV jack.

In this installment, I’ll show you how to use a video monitor with your ZX/TS, and while we’re at it, we’ll clean up the signal with comparators and allow a switchable “reverse video: option (not included on most under-$100 monitors). The cost is about #10: an parts. This and the “clean-up” feature make the project worthwhile even if you want to keep using a TV (such as my small but very crisp Panasonic 3″ portable).

This should take care of any remaining complaints you might have about your computer’s video. (Except its speed in printing to the screen; which we can’t do much about until we build John Oliger’s video upgrade in the first two issues of Volume 2.)

But first, I’d like to present a “video primer” for beginning to intermediate ZX/TS computerists. So on the next pages you’ll find “everything you wanted to know about TY, but didn’t know who to ask.” Hopefully this will give you the “why’s” and motivate you to tackle the hardware project (the ‘hows.”) If you’re a software hacker, knowledge of the hardware of TV might help in your making sense of the ROM’s TV display routines. The discussion is limited to monochrome (B&W) systems to keep it simple, as you lucky TS2068 users already have all you need for connection to a color monitor.

TURNING A POINT INTO A LINE AND A LINE INTO A PLANE

A television set (or video monitor) is a serial device; this means that at any given time, only a single dot of light is illuminated (or not). The signal that reaches the TV is processed into a voltage called a “video signal,” which increases or decreases to vary the brightness of the dot as it travels across the screen. The dot starts at the top left corner of the screen, traces to the right, then flies back to the left and traces the next line. This happens 15,750 times a second. Meanwhile, the dot is swept from top to bottom at the much slower rate of 60 times a second. When the spot reaches the bottom of the screen, it flies back to the top to start over.

So given these facts, we should be able to figure out the total number of lines, right? So let’s divide 15750 by 60 and get… what? 262.5? That’s right; only 262 and half lines are drawn every 1/60 of a second. The “one half” means that the spot is already half-way through the next horizontal line when it gets back to the top. The result is that the subsequent set of 262 lines is drawn interleaving the first set. Note that while there are 60 “frames” per second, only 30 complete pictures (“fields”) are generated each second. This approach reduces flicker as compared to using a vertical retrace interval of 1/30 sec., since the eye can easily discern discrete “flashes” at that rate. So we have a total of 524 full lines, and a half-line at top and bottom, for a total of 525 lines, the quoted resolution for the NTSC (American standard) TV system.

This resolution figure assumes that the vertical return (retrace) happens immediately. Actually, this time is non-zero; the result is that some horizontal lines are lost during the retrace interval. Other lines are lost off the top and bottom to ensure that the borders will always be off-screen. Also, the top and bottom of the actual computer display is even less than the physical limit, for the same reason. The end result is that, in the ZX/TS system, only a total of 192 (24 X 8) lines are actually used.

Similarly, the horizontal retrace operation takes a non-zero time to complete, and again we aren’t using the full 1/15750’s of a second to actually convey the data in a line. Good thing, for us ZX/TS users, since it’s during these “dead times” that your poor overworked machine has to do its actual computing in SLOW mode. With the American TV standard the ratio of “on time” to “off time” (as far as the screen is concerned) is about 6:1, which is why SLOW mode is 1/6 the speed of FAST mode. Yes, I said 1/6, even though your manual says “1/4.” This much under-publicized fact was first brought out in SYNAPSE (the publication of the Central PA ZX/TS UG); those guys are really on the ball! The reason is that in the 50 Hz. PAL (British and European) TV system, there is more time to “do other things” (ie. compute) during the retrace intervals. That bug in the manual has persisted to the present TS1500 manual! While we’re on the subject, dear North American reader, please note that any references you see to “50 Hz.” or “50 times per second” in the manual or other books should be changed to “60.” (eg, anything having to do with PAUSE, FRAMES, etc).

But back to the topic. How does that little spot know exactly when to start a sweep? This is done by sending “synchronizing pulses” along with the stream of video information (light and dark). The width of these pulses determine the “dead-time” between adjacent lines (horizontally) and frames (vertically). These pulses are negative-going; i.e. “negative logic,” which has the same advantages, engineering-wise, as negative logic in digital circuits. The level of the signal is typically about +2 volts, and drops to close to 0 volts during retrace. A circuit in the TV/ monitor called a “syne separator” sorts out these pulses and uses them to “lock-step” the horizontal and vertical oscillators which control the position of the electron beam (spot on the screen). To see what happens when you don’t sync the video signal, mess up your horizontal or vertical “hold” controls. See how the sync signal “locks in” the picture over quite a wide range of adjustment. By the way, the horizontal hold works as a centering control on most sets.

Now we’ve got a lit, stable screen (“raster”), but it’s blank. So at the same time as this is going on, we have to vary the brightness of the spot to make a picture. This is done with the “video signal,” which is exactly synchronized at the sending end (computer or TV camera) with the syne pulses. When the signal is as high as it will go (typically 1 volt) the spot is bright, and when it is zero, the spot is completely darkened. Note that the video signal uses “positive logic” in the NTSC video system.

By now you may be wondering how TV deals with the three different signals involved; horizontal syne, vertical syne, and video. The solution is to add them together and send them all as one signal. During syne pulses (retrace) we couldn’t care about video, since we’re “off-screen” anyway. So we might as well superimpose the video on top of the sync pulses. The sync signals are timed such that the vertical pulses is much wider than the horizontal ones. The result is a monochrome “composite video” signal, as shown in Figure 1. Note that the negative-going sync pulses help insure that the screen is black during retrace; another reason for this approach, since it makes the “retrace blanking” circuitry quite simple indeed. At the receiving end, the syne separator can easily differentiate the vertical (wide) pulses from the horizontal (narrow) ones and from the video signal (time between sync pulses).

So far the discussion has been general enough to apply to both TV’s and monitors. What makes the difference between the two is that the monitor will accept a composite video signal, whereas the TV won’t. In order to get a signal to the TV without wires, good ole AM radio is called “into the picture.” That’s right, AM. The composite video signal is “modulated” onto a radio-frequency (RF) “carrier.” The composite signal simply varies “how big” (amplitude) the radio signal is at any given time, The “modulator” in your computer is actually a tiny TV transmitter, consisting of an oscillator (about 55-65 mHz. for N.A. channel 2-3 units) and the modulator itself which controls the radio signal’s amplitude, The signal then is attenuated (reduced) so it “looks like” something a TV would expect at its antenna terminals (about 10E-7 volts). At the TV end, the signal is amplified (RF amp), reduced to a standard, more manageable frequency (converter), amplified some more (IF amp), and only then converted back to a video signal (demodulator). With all that electronic “stuff” to go through, no wonder TV displays don’t look as good as video monitors, which accept the composite video signal directly.

At this point, I might mention that the sound portion of the NTSC standard signal is FM (frequency modulated), unlike the video portion, This is why it is difficult to implement “through-the-TV” sound on any computer; the ones that do (like game machines, CoCo, and others) usually don’t sound too good because they’re actually using a bastardized AM signal to try to “trick” the FM demodulator into a reasonable facsimile of the intended sound. Once again, you TS2068 users have the edge with your direct audio-output jack and built-in speaker.

So a TV is nothing more than a video monitor with a radio receiver (a rather complicated one, but still a radio) attached to its input. So how about tapping into the TV’s composite video line directly? Sounds good, but rotsa ruck.

On most TVs you’ll have trouble right off the bat because the TV is not isolated from the AC line with a transformer. At best (if your system is grounded) you’ll have blown fuses and sparks, at worst (if using a “floating ground” like most) you’ll have a serious shock hazard.

Well, yes, you can isolate the two systems using an opto-coupler or isolating power transformer. Even so, it might be hard to locate the composite video line; and if you do, it may be offset by some arbitrary voltage.

The goal of TV designers is to make a working set with the lowest parts count, and over the years they have come up with some remarkably tricky but effective ways to make a bunch of hardware do “more than it should,” The “cost” is that it makes things difficult for experimenters. Even on portables that are isolated, there may not be an accessible point to inject the composite video; that point exists, of course, but it might be buried inside some chip where you can’t get to. Trust me! You’ll be better off getting a monitor made for that purpose.

But if your budget doesn’t allow this additional piece of hardware, there is still something you can do to substantially improve your TV computer display.

The signals that race around in your computer are digital; ie, “square waves.” They have a rapid ON transition and a rapid OFF transition. If you’ve studied Fourier analysis, you know that rapid transitions in a waveform means a large high frequency content (harmonics) in the waveform. If you haven’t, don’t worry about it; take it as “Blind Faith Principle No. N+1.”

These harmonics extend well into the VHF range (especially into the “low VHF” area our computers work in). This is what causes those annoying moving Moire-pattern effects; the more hardware you add onto the rear-connector, the more your system acts as an antenna to radiate this garbage to your TV.

This type of interference can be greatly reduced by changing from a VHF (channel 2 or 3) to UHF (typically channel 33) modulator. UHF is so much higher in frequency that the harmonics from the computer have a negligible effect.

Such a modulator can be purchased from Computer Continuum and others, and simply replaces your existing modulator. Add our clean-up/reverse circuit, and you can get most of the advantages of a monitor using a good TV.

There is another good reason to switch to a monitor, dealing with the persistence of vision and the persistence of “phosphors.” The light-producing coating on the inside of a picture tube actually continues glowing for a while, just like “glow in the dark” substances (essentially the same thing).

On TV sets, the phosphors used have an intentionally fast “decay time” so that there is no or little “streaking” of fast-moving scenes, The price is that there can be noticeable flashes or flicker, especially when you’re shifting your glance from the screen to somewhere else (like a program listing) and back.

Computer monitors use a picture tube with a longer-persistence phosphor; the image glows for a longer time, and almost completely eliminates flicker or “strobing” with flourescent lights. This is great for programming, text, etc. but a mixed blessing in other ways.

For one thing, it causes noticeable “trails” on fast-moving graphics (particularly if the moving “object” is light on a dark background) which may or may not annoy you while playing space games. The other minor negative effect is that it makes light-pens more difficult to adjust, since they depend on receiving a sharp pulse when the light-dot passes the location of the pen. These points are almost trivial, though, when compared to the reduced optical annoyance when looking to the screen and away again for hours on end.

So what’s a good monitor to get? Probably the most common right now is a unit brand-named BMC which is currently appearing at stereo/video outlets. It’s not bad, and produces a bright, green picture with adequate horizontal resolution. It does have a minor “bug” when used with the ZX/TS which I personally find objectionable; there is a pronounced vertical bounce when the computer comes back from FAST mode or after a PAUSE. This unit (and a few others) is essentially a re-worked TV; early units even still have the speaker grille in the casing. A TV is not meant to have the video signal, including retrace pulses, disappear and re-appear.

I use a 9″ USI green-screen (excellent product) and have used the Zenith 12″ amber-screen (prettier case but more expensive) on the battery-evaluator project for G&M Power Products. Both provide near-instant stability on return from FAST, though the USI is just a tad quicker. USI also makes a 12″ version which includes a reversing switch, and both the 9″ and the 12″ have a DC restoration switch (discussed later).

Don’t pay too much for a monitor! Spend a few hours at a book stand looking through some magazines to find the bargains. You shouldn’t pay more than $100 for a decent green-screen, slightly more for amber. (I prefer the green, myself, but this is largely a matter of taste; personally, I seriously question the claim that amber causes less eye= strain.)

Now, let’s take a look at the question of “reverse video.” I’ve found that most beginning computerists prefer the black letters on white as provided by the ZX81 and other home machines. Perhaps this is because it is most like dark ink on white paper.

However, the longer you stare into the screen debugging programs or entering text, the more you’ll appreciate light letters against a black background; it’s simply easier to look at for long periods of time. You don’t have all that light in the background burning a roughly rectangular image into your retinas.

I also like what happens during FAST mode; instead of going grey, the screen simply disappears (goes black). The “flash” is not nearly as pronounced, and the effect is more like “I’m going away to think for a while” than “I’ve just lost my marbles.”

Another worthwhile feature is a high-impedance input option, provided as a switch on some models (incl, USI). It takes less power to drive a hi-Z input, so the “driver” you make for it can use a lower-power transistor (and takes less from the supply).

Finally, look for a “DC Restoration” switch. You may have noticed that night-scenes on TV don’t look black, they’re grey. This is because video amplifiers are usually AC-coupled, which means that the black-level varies with the ratio of light-to-dark across the screen. A circuit called a DC restorer is used in the better monitors to clamp the black-level at or near “true zero” no matter how much or little is printed on the screen.