29

Green was traditionally the most common color for computer monitors; it combines strain-free readability with low cost.

Given this, it's surprising that the first versions of the Commodore PET and TRS-80 came with bluish-white monitors, though mainframe terminals had been using green screens for many years. The PET at least, later switched to green.

Had just one of them used bluish-white, random chance might seem a likely explanation, but both at the same time suggests a systematic reason. So why did they not use green?

  • 8
    Why do you think green was a sign of something good? I had a Hercules green screen for a while and have distinct memories of the resulting yellow blob in my vision for hours after using it. – Criggie Jun 21 '18 at 13:04
  • bluish-white sounds like a normal CRT without any coating – Thorbjørn Ravn Andersen Jun 21 '18 at 18:28
37

The reason was cost, since neither the original PET nor the TRS-80 required the extra high resolution and finer dot pitch found in more expensive computer monitors.

Virtually all of the computer terminals of the time paired a CRTC with a high-resolution green or amber screen to generate an 80x24 text display, as that was the early standard. This required a relatively expensive monitor that could resolve the full horizontal resolution required for readable 80-column text (normally in the range of 560 to 640 pixels) without blurring adjacent pixels. The PET used 40 column text, and the TRS-80 used 64 column text. Therefore, neither required a high quality computer monitor to achieve their more modest resolution needs.

Radio Shack, in particular, rather famously repurposed a CRT from a B&W television that the company also sold. They removed the unnecessary tuner parts and added a monochrome input to turn it into a monitor. This no doubt greatly reduced their costs in comparison to using a more "legit" monitor. Commodore was likely motivated by similar economics and chose a lower resolution TV CRT as well.

  • 2
    @rwallace Right. Vertical resolution was not the main limiting factor. – Brian H Jun 21 '18 at 0:22
  • 5
    @rwallace In my experience the 80x24 were far more common early on. So much so that a lot of software had 80x24 hardcoded so when 80x25 became more common a lot of terminals mostly (sometimes "only") used the 25th line for function key labels or similar stuff - sometimes under program control but separate from the main 24-line section. – manassehkatz Jun 21 '18 at 2:26
  • 2
    You can even see this as late as the original Amiga OS in versions up to 1.3. It was really designed to only work properly with a RGB monitor, but the white text and blue background is still there by default, and it includes a 60-column setting in the OS to maintain readability on a cheap colour CRT TV via the onboard RF output on the A1000 and inexpensive external RF adapter on the later A500. – mnem Jun 21 '18 at 5:38
  • 2
    @mnem Pity that I can't find the link. One of the Workbench developers said that they took their prototype to an eletronic store, attached it to a TV set and tried different colors. White on blue was the most readable one. – trunklop Jun 21 '18 at 13:20
  • 2
    @mickeyf - but what would have happened if you ran kermit on it? Then kermit would have been just as green, surely? – Jules Jun 21 '18 at 14:21
23

The answer, as always, was cost. 'White' CRTs were cheaper because they were used in B/W TV sets. The color itself has no impact on resolution, but TV tubes didn't need to be as sharp so they could be made cheaper.

So why were 'high resolution' monitors green or amber? Because these phosphors have a longer persistence so the image doesn't flicker as much, especially at the lower refresh rates commonly used to get higher vertical resolution. Paper-white is nicer, but requires a mix of different colored phosphors which is difficult to make with matched persistances (without which the image would change color as it faded).

The down-side of a long-persistance screen is horrible smearing of moving images - hopeless for action games or TV shows, and not good for scrolling text either.

  • 8
    The ghosting due to high phosphor persistence that green screens were famous for is famously imitated in the "Matrix" movies. – Robert Columbia Jun 21 '18 at 14:28
  • 1
    The Compaq portable computers used CGA-style display hardware, but displayed higher vertical resolution by reducing the vertical scan rate by almost half. The screens had a very long persistence green phosphor, however, so flicker wasn't noticeable at all. On the other hand, I suspect on many computers the issue wasn't so much "flicker" as update noise; many display designs would momentarily blank the screen any time the CPU was writing it. On a low-persistence screen, this would often cause random black streaks to appear superimposed on the screen. – supercat Jun 21 '18 at 15:00
  • @supercat: that was a problem with memory-mapped displays on personal computers, which would suffer from displayus interruptus if programs weren't careful about when they wrote to display memory, but I don't recall seeing that on standalone terminals such as the DEC VT-100 & family. – Bob Jarvis Jun 22 '18 at 2:34
  • @BobJarvis: For machines with long-persistence phosphors I don't think it was a case of being careful. I think there was simply an assumption that the faint dark streaks would be less objectionable than slower display output. – supercat Jun 22 '18 at 5:30
6

A few points to expand on previous answers:

In those days, resolution wasn't limited by phosphor, but by video bandwidth (and video memory). A standard TV had a bandwidth around 2-3 MHz, enough to support a PET or APPLE ][ - style 40x24 text display, or about 320x240 dots. The TRS-80 Model I monitor had slight modifications to the video circuitry to allow a 64x16 text display, about 384 horizontal dots (characters were 5x7 in a cell that was 8 by... shoot, I don't remember how many vertical pixels per character).

Higher-quality monitors had higher bandwidth, and could support more horizontal resolution. (Vertical resolution was tied to NTSC/PAL video standards, which limited the number of scan lines; non-NTSC monitors got really expensive really quickly.)

As others pointed out, long-persistence phosphors could let you get away with slower scans, but they were also less flickery at standard rates. Thus, they commanded a premium.

I remember more of this than I should, because I used Don Lancaster's TV Typewriter Cookbook to build an outboard video card for my TRS-80, and I had to learn all about video timing and dot-clocks. It ended up displaying 100x30 characters on an 800x240-pixel raster, on an amber monitor that's still out in the garage somewhere. Pretty luxurious, for the time...

  • 1
    A great book. Most of Don's books are available for free download on his website: Free ebooks. Unfortunately, CMOS cookbook is not available for free download, as he co-authored that work :-( – Greenonline Jun 22 '18 at 5:06
  • Very cool! Thanks for the link (and the edits)... – jeffB Jun 22 '18 at 13:10
3

The biggest single problem with long-persistence green monitors is that they were extremely vulnerable to burn-in unless you kept the intensity relatively dim. Amber monitors were usually a lot brighter. Computers like the TRS-80 Model I and Commodore PET didn't use long-persistence grayscale... they used picture tubes literally manufactured for use in normal black & white TVs to cut costs.

From what I recall, long-persistence white phosphors existed, and weren't outrageously expensive at the manufacturing level... but they had even bigger problems with burn-in than LP green phosphors did... partly, due to differences in the way people USED long-persistence black & white monitors vs the way they used green or amber. With green and amber, people generally used colored text on a black background. With long-persistence WHITE, people wanted black text on white backgrounds, which caused the monitor to visibly burn in even faster. White backgrounds on 60hz monitors with normal-persistence phosphors were kind of distracting... they didn't necessarily flicker while you were looking straight at them (though they did), but if you tried reading from a book in your lap while using one, you'd REALLY notice the flicker in your peripheral vision. 50hz was even worse... it flickered even if you looked directly at it, and almost felt like a strobe light if you tried reading a book in your lap with the screen visible in your peripheral vision (at least, on a Palbooted American Amiga with a color monitor in a room with 60hz incandescent lights, which might have made the strobing effect worse than it would have appeared to someone in a country where the room lights were flickering at 50hz as well).

From what I recall, green (and probably long-persistence monochrome monitors in general) had TWO intensity controls... a recessed one on the rear (that usually required a screwdriver to turn), and the "official" one near the front. The rear one established the maximum-allowed brightness... the front one allowed users to adjust the brightness between "totally black" and "the limit established by the rear control". In most cases, the maximum-permitted intensity was WAY lower than users would have really preferred, especially in brightly-lit rooms. And despite their relative dimness compared to amber, green monitors STILL ended up with visible burn-in artifacts long before amber monitors of comparable age.

When I was in college (circa 1992), the University had a mix of green, amber, and color monitors in the computer labs. The amber ones were generally less-preferred than the color monitors (unless it came down to, "non-IBM clone with color monitor and mushy keyboard vs IBM VT-102 terminal with silicone-greased clicky-keyboard and amber monitor), but EVERYONE hated the green ones. They were dim, had burn-in artifacts, and just generally weren't nice to use.

  • I wonder why terminals that used green monitors weren't designed to slightly shift the position of the display over the course of a few hours as a means of mitigating burn-in effects. Most displays had a row or two of border at top and bottom, and 2-3 characters on the sides, so shifting by +/- half a character or so would have yielded a much less concentrated burn-in pattern. – supercat Oct 6 '18 at 23:02
  • 1
    "IBM VT-102" - no such thing. Probably DEC VT102. – manassehkatz Oct 7 '18 at 1:12
  • Hmmm... now that you mention it, they were probably IBM 3151 terminals. – Bitbang3r Oct 7 '18 at 5:49
2

Early research found that green or amber was easiest on the eyes of people looking at CRT monitors all day. Actually, I preferred amber, myself.

  • 1
    That explains why they were green. The OP is asking why they were often NOT green even after green was established as easier on the eyes. – manassehkatz Jun 22 '18 at 17:50

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.