[Coco] RGB2VGA

[Gene Heskett]

On Friday, February 13, 2015 10:17:20 AM Mark Schoenberger via Coco wrote:
> I have this board hooked up to my Amiga 2000.آ  Works very well. I
> haven't made an adapter cable for the Coco, yet, but I am sure it will
> work. 
> آ
Yikes!  Teach ymail some manners Mark, you just included in your reply 32 
kilobytes of a daily summary, encoded as a base 64 format.  Which I have 
deleted.

Short answer is yes.

The longer answer is that you will have to add a chip in the coco to 
combine the H&V syncs and invert the mix to use this card.  Otherwise it 
won't sync to the upside down signals the 10 pin socket provides.

I've since become the owner of one of Roy's 2nd generation cards which 
works much better IMO. The jamma cards output is rather low, and in spite 
of that, its inputs are also fairly easy to saturate, rendering the hf 
components of the coco3's output to slow to even let the vertical bars of 
an 80 column character render on screen, so certain colors of screens 
(blue in particular) are not usable at all.  90% of the fault lies in the 
gime itself, its own video outputs are too slow to do it well, but to have 
character detail just disappear can leave one guessing as to whether you 
typed an n or an m in many cases.  Since they are adjacent on the 
keyboard, its too easy to do anyway, without a poor display that does not 
clearly show the difference adding to the error.  The gime is so puny in 
its output drive that adding the 11 pf of one of my 10x scope probes so I 
can measure its rise time is easily seen in the further smearing of the 
output image in the color I was tapped onto. The rise time? over 400ns at 
the output of the buffer transistor, where the scope probes influence 
cannot be discerned on screen.  For accurate renditions, at NTSC speeds, 
this needs to be about 120ns or less, 120 being the limit for NTSC when 
its kept within a 4.1Mhz bandwidth when we transmitted it from 1946 to 
2008.  Yeah, me retired tv engineer. The gime's output transistors simply 
are not big enough.  The designers may have had a power budget that did 
not allow that, but no one here is privy to that info.

I have had, in the back of my mind, the idea to replace those 3 buffer 
transistors with a modern video speed op-amp, with enough hf boost built 
in to alleviate that problem, but seem to have miss-laid my round tuit.

That and I'm getting lazy in my dotage. ;-)

I do not think it would be that big a problem to do, and it should, if 
close enough to the gime, be able to do it without emphasizing any other 
hf clocking noises the gime may produce.

Adequate supply rail bypassing to reduce that noise at the source, 
bypassing thats lacking in the coco3, might be required, as would an 
additional volt of supply. That I don't believe would heat it to a degree 
to effect its dependability, but I was amazed at its improved performance 
in my coco3 when I bypassed a 10 ohm resistor in its 5 volt line 1/2" 
south of the socket where some d__ned bean counter had replaced a ferrite 
bead used by the designer, as shown on the published schematics, used for 
noise control with the only 10% as costly resistor when it was converted 
to a Bill of Materials for production.  That bypassing gained only about 
80 millivolts of supply at the chip. And made a huge difference in the 
quality of the video.

I have also entertained the thought of removing said resistor and 
replacing it with a teeny voltage booster that I could adjust upward from 
unity to perhaps another volt, keeping a close watch on its temperature 
with an IR thermometer.  It seems to be very very sensitive. 1987 version 
in this machine.  Since  I have external power in mine, with power to 
throw away, there is also the possibility of setting up an lm324 running 
from the 12 volt line to feed it 5.5 or more volts. Feed it into the pcb 
hole where the far end of that 10 ohm is now.

I've also toyed with the idea of bypassing the gimes video outputs by 
grabbing the data when the gime reads it, and doing my own 3 channel D to 
A. But that isn't something the stock coco has power supply enough to get 
away with. That of course is something that would have to be done on a pcb 
under the main pcb as there isn't room on top for anything like that.

I'll get me coat now...



Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page <http://geneslinuxbox.net:6309/gene>
US V Castleman, SCOTUS, Mar 2014 is grounds for Impeaching SCOTUS