On Sun, 28 Dec 1997 22:09:03 +0100 =?ISO-8859-1?Q?Thomas_M=F8rch?=
<oz2afo@POST8.TELE.DK> writes:
>It
>could be fun to make a rotating (digitaly) clock display (like the
>opengl
>screensavers that come with WinNT 4.0), that was blown up on the
>wall...
>
> I was planning on using some galvanometers (if i can find some cheap)
>(Someone here mentioned that there was some in a videodisc player???)
>And a
>laser diode, for the light thing..

> 2. Where do one get hold of galvanometers?

Look for "swing motors" instead.  The pen in some types of chart
recorders (anyone remenber chart recorders) such as those used for
medical analysis was driven by one.  Most newer hard drives use one (and
a lot of feedback) to move the heads.  The hard drive motor would need a
spring added to give it some direction of where it should stay.  Also
I've heard of removing the paper cone from a loudspeaker and attaching a
mirror instead.  Driving these devices with current rather than voltage
may work better.

> 3. How fast can one turn a laser led on/off?

Very fast.  Several MHz are possible.  But, using a standard module with
the power control built in it may not operate faster than a few Hz
without modifying something.
>
>I know that it must be some kind of vectorised graphics engine in a
>DSP (i
>don't think a PIC is fast enough for this one). But besides this i
>haven't
>got any clues as to how i should attack this little challenge...
>
Early video arcade games used a vector scan monitor and obtained effects
like this using the 8-bit Von Neuman processors (Z80, 6502) that were the
style of the time.  So the AD demonstration may actually be using little
processing power, but still making an impressive demonstration.

The process generally involves computing a matrix based on the angle
position of the display, then multiplying the x,y coordinates of each
point to be displayed by it to get the screen coordinates of the rotated
image.  So a fair amount of multiplication and addition (DSP chips are
good at that) but very little trigonometry.

After the screen coordinates of each point have been computed, then it's
just a matter of computing the x and y slopes between the ones that are
to be connected with lines, and driving the display.

Rather than tackling the laser end right away, try getting it running on
an oscilloscope first.  Or, rather than trying to compute a rotating
display, try making a simple straight one first.