I've only handled one system that had a light pen on a raster display. My
experience with a lightpen is actually limited to the vector displays on CDC
systems back in the early-mid '60's. They were quite precise in their placement
on the rather large screeen, and didn't make first-approximations and let you
correct as you went as one might do on a raster display. Since putting a pixel
under the light pen was tantamount to making the entire screen white and then
erasing/redrawing it, which took quite a while, and that's not what they did,
what do you suppose they did on those storage tubes?
BTW, with a small-enough aperture at the tip of the light pen, they could have
achieved single-pixel resolution, though it might not have been terribly
helpful, since the amps used back in the '80's when the 6845 weren't generally
fast enough to resolve a single pixel at, say, 16 MHz, which was a pretty slow
dot clock, but adequate for the popular rasters of that time, e.g. 512-768
pixels per line. Some makers did manage it, however.
more questions/comments below ...
Dick
----- Original Message -----
From: "Tony Duell" <ard_at_p850ug1.demon.co.uk>
To: <classiccmp_at_classiccmp.org>
Sent: Monday, July 16, 2001 6:31 PM
Subject: Re: Light Pens ...
> > Ok, my mind is still blown. How does a light-pen drawing program work
> > then? (Or is a light-pen drawing program not possible?)
>
> Simple : You make sure there's something displayed where the light pen
> is, like a cursor.
>
> Typically, when you put the pen on the screen, the machine moves a block
> or crosshairs about until it finds the light pen (sometimes it just lights
> up the whole screen). Once it knows where the lightpen is, it displays a
> 'cursor' at that location. The important thing about the cursor is that
> it's larger than the resolution of the light pen (say a 3*3 array of
> lightpen 'pixels). Then, when the user moved the lightpen, the machine
> detects that the lightpen is no longer over the centre block of the
> cursor, and moves the cursor so that it is. The cursor thus follows the
> lightpen around the screen, that always providing something for the
> lightpen to detect.
>
What do you suppose they did to keep the cursor centered? Did the integrate the
waveform from the light pen? How would they handle the vertical centering? Do
you suppose they might have used two integrators with one for each axis? I can
envision such a scheme allowing them to acquire the pulse from the light pen,
then detect the peak and resolve to the single-pixel resolution with a pixel
counter. The scheme would work for either axis, though the pixel counter would
only be involved in the horizontal scan.
The Motorola system on which I saw the demo of a "full" implementation of the
light pen together with a 6845 was able to do the job all right, but I'd be
surprised to learn they didn't have a dedicated processor just for diddling with
the cursor location to track the light pen. I can see that the button on the
light pen was used to invoke a "find the pen" command, and, probably, while the
button was pressed, tracked the position of the pen on the CRT. That would
explain the presence of a 68121 (68121 = 6803+6834 on a single device) video
board, since that was possibly where the intelligence for the light pen, plus
the 2-ported semaphores + RAM for sharing data between the two processors
(68K+68121) would reside. It made for a pretty complex system, though. I doubt
this sort of thing would have popped up in a system costing less than $10K at
the time.
Received on Tue Jul 17 2001 - 09:50:59 BST