CRTs, LCDs and image capture [was 1966 Mag: Build NE-2 Neon Bulb Computer - scan available]

Ethan Dicks ethan.dicks at
Wed Jul 18 10:59:28 CDT 2007

On 7/18/07, John Foust <jfoust at> wrote:
> It certainly is if you have equipment that likes a video scan mode
> that isn't supported by today's LCDs...
> A similar problem extends back into the RS-170 composite-mode
> era of home computers.  Devices from 20 years ago tossed out
> all sorts of approximations of "composite mono video" and
> older TVs might handle it better than newer TVs with
> less forgiving circuitry.

I have had that specific problem with RCA 1861 "Pixie" graphics - I
can get viewable images on 1970s-era CRTs and TVs with no problem, but
I could not, for example, get a stable picture with an Apple 5" CRT
(used with the Apple IIc line), or a DEC VR201 (+12VDC/RS170 mono
monitor for Rainbow, DECmate, Professional, etc.) nor any sort of
recently-manufactured TV or LCD.  Depending on the exact output device
in question, either I see pixels but can't get a frame to properly
lock, or I don't see anything at all.

> I daydreamed about a little gizmo the other day.  Plug it in
> as a pass-through with two VGA 15-pin connectors so I watch
> the video.  Press a button, capture the image as a bitmap for me.
> Has a USB plug so I can fetch it with another computer.

Hmm... interesting thought.

For my Pixie problem I had been looking at a 4066-based circuit from
Don Lancaster's "CMOS Cookbook" - from memory, you run the 4066 off of
+5V and -5V rails, then hook your Hsync, Vsync, and video outputs to
three of the switch enable inputs to the 4066.  The 4066, in effect,
synthesizes a "proper" RS170 signal from various appropriate voltage
levels rather than the standard Pixie technique of summing those
signals through resistors and letting the CRT sort it out.

One of the things I'm not sure about is if a device of this sort would
need an isolated +5V supply to work or not.  I can use my Spare Time
Gizmos STG1861 Pixie emulator ( ),
populated with modern parts, so I don't have to risk a real CDP1861 if
I screw something up.

At the very least, I think I could breadboard a 4066 as described in
the CMOS Cookbook, then get -5V from an ICL7660 (I have a few lying
around, and it's still available as a MAX7660).  I could also be
paranoid and use a 4050 as a buffer/level shifter between the Elf and
the 4066, but that might not be required.



More information about the cctalk mailing list