bob@jacobs.CS.ORST.EDU (robert s. richardson) (08/02/89)
Update on Flicker-Free Monochrome Amiga Monitor
I am sending this to all those who mailed me regarding my previous post
on this subject, and am posting to the net for others who may be
interested.
BACKGROUND
A couple weeks ago I wrote that I had found a source for a "monochrome
composite monitor with a high-persistence phosphor" that would probably
work with the Amiga. I ordered a sample to test for myself.
Well, it arrived last week, and after a week of interfacing woes, I
finally came up with a cable that works (see below.) What follows is a
brief review of the monitor and how to order.
AUTHOR'S NOTE
I am a student who also runs a very small software consulting company
dealing with desktop video and hazardous waste (no kidding!). I
ordered the monitor with my company name. I am selling it to net users
at a modest profit. I need to pay my tuition somehow (and justify the
purchase of one of these babys for myself!) On with the review.
DETAILS
This is a Kraft brand (not the dairy people) monitor that is currently
being liquidated. It features a 14" low glare flat screen with a
paper-white long persistence phosphor. It was originally designed to
work on IBM-PC's outfitted with a CGA digital RGB adaptor. It accepts
an RGBI input with separate horizontal and vertical sync.
The case is tapered to the back (it follows the outline of the picture
tube rather than being a big square box), is off-white with grey trim
around the screen, and comes with two bases. The first is simply a bar
that folds down to tilt the monitor up. The second replaces the bar
with a complete tilt/swivel base, which raises the monitor about 2"
higher than the bar.
The resolution of the IBM CGA was 640x200 non-interlaced. There is a
switch on the back of the monitor that super-imposes an interlace
signal onto the video, causing the gaps between scanlines to disappear.
There is apparently some interesting circuitry inside this thing.
Unless it gets exactly the signals it wants, it does not turn on the
electron beam, but keeps the tube warm. This means that if you turn
off your computer, the monitor also shuts half-way down. It also made
it very hard to see if my cable was working or not while I built it,
because the monitor would not turn on. I was seriously wondering if it
worked at all until I took it to an IBM store and hooked it up to a
CGA.
THE GOOD NEWS:
When I finally got a working cable built, I booted my 640x200 workbench
and began to experiment. I turned on the interlace switch in the back
and the picture was BEAUTIFUL! Very little or no flicker at all
depending on how bright I turned up the picture. (And you can turn it
up blindingly bright.)
I tried it with Handshake and smooth scrolling and I swore I was using
a crisp, authentic VT100 display.
Playing around in preferences I determined the monitor displays 5
levels of gray and 2 levels of black. (One black is fuzzier on the
edges than the other, causing the pixels that are black to appear
bigger.) The brightness is determined primarily by the red and green
level settings. The blue level is more of an intensity. ODD levels of
blue are more intense than EVEN levels. So heavy black would be RED=0
GRN=0 BLU=1 and normal black would be R=0 G=0 B=0. When blue goes in
the range of 8 to 15, more intensity is added than in the range of 0 to
7. It takes some getting used to, but some nice combinations can be
arrived at.
THE BAD NEWS
Although this monitor is very adept at spicing up a 200 line signal, it
is not quite sure what to do when interlace is sent from the computer
itself. (Remember, IBM's simply don't interlace, so the monitor was
never designed to support an external interlace.)
When I brought up a 640x400 workbench, strange things happened. First,
if you have the interlace switch OFF, then every two scan lines are
displayed as one with gaps in between. (ie, lines 1 and 2 from the
computer display right on top of each other, then a gap, then lines 3
and 4 and so on.).
Aha, I thought, all I need to do is turn the interlace switch on,
right? Sort of. There's a 50/50 chance it will catch on the wrong
field display. This means that even and odd numbered lines display out
of sequence. (Line 2 appears above line 1, line 4 above line 3, and so
on.)
So you're saying, Okay, simply keep flipping the switch until it
catches the right field? Well, that indeed does work, partially. You
still get gaps every two lines that aren't there when the monitor does
it's own interlacing. And if you drag an interlaced screen around over
a non-interlaced one, you can sometimes get the reversed scan-line
syndrome again. A few flips of the switch
or re-drags corrects this.
A solution to all this is some sort of hardware or software hack to the
Amiga that enables interlace mode without modifying the vertical sync
singal. (Send out alternating frames, but keep lo-resolution sync.)
Any ideas, fellow netters?
There still is very little flicker, though. Its better interlacing
(when it works) than a 1084 as far as I can see, and the 640x200 mode
is outstanding.
MORE BAD NEWS
Okay, you say, you'll order it anyway and can probably live with the
difficulties of getting interlace up and running. But since this is
liquidated the warranty is no longer valid. My source will replace
defective units only as long as there is remaining inventory, which
might be two weeks or two years, nobody's shure.
I called the manufacturer and they said they will still provide
schematics and service manuals to anyone who wants one. They sent me
mine for free, but they haven't arrived yet. I don't know if they will
keep sending them for free if they get a lot of
calls.
MORE GOOD NEWS
The monitor looks fantastic on an old CGA/PC if you have one lying
around. It may also work on a C128 if it provides separate horizontal
and vertical sync on its RGBI connector. (I can't remember.)
THE CABLE
My problem was, as I discovered, that the sync from the Amiga was
active low and the monitor wanted exactly the opposite. So I bought a
$.99 hex inverter at Radio Shack and hooked it between the sync lines.
Still nothing. Hmmmm, as it turns out my +5v user supply is out for
some reason. Time to get my 2000 fixed. So I used the +12 supply with
a 330ohm resistor instead and ta da, it worked like a charm. Total
adaptor cost, about $5 in chip and connectors (a DB-9 and DB-23
(actually a DB-25 which I filed down.))
If you don't want to build a cable, Redmond Cable sells a DB-23 to DB-9
with hex inverter for about $30. I will supply a scematic with each
monitor for those who do like to mess with a soldering iron.
THE NITTY GRITTY
Okay, so you want the monitor, how do you get it?
Well, I'm asking $149 for the monitor and $15 for shipping (UPS
ground). I will test each monitor before it gets shipped. Please call
me anytime between noon and 10pm if you have any questions about this
monitor.
SUMMARY
I'm keeping mine. Its beautiful enough to overcome its deficiencies.
I guess the best I can say about it is that I can't part with it now
that I have it, in spite of the problems.
SIG, ETC.
Thanks for wading through this, I hope I sounded more informative than
commercial.
| Bob Richardson (or, for you UNIX buffs: bob@jacobs.cs.orst.edu) |
| 218 NW 21st #2 Corvallis, OR 97330 503-758-5018 |
| "They have taken my Jesus and locked him up in a church!" - Unknown |