LASERS! && Freemont Street LED array (was Re: Cray J932SE (was Re: Straight 8 up on Ebay just now))
markwgreen at rogers.com
Tue Jul 19 16:23:05 CDT 2016
In my day job I work on computational holography and other forms of esoteric
3D displays, so I can give you some insight in how these things work.
Remember these are vector displays and not raster displays, so the
computational side is not an issue. You are basically looking at a pair of
D/A convertors that are driven by a pair of parallel ports. The circuit is
probably a bit more complicated than that, but you get the idea. This can
be done interactively with no problem. With a modern CPU you are probably
looking at less than 1% of the CPU time.
The complication occurs with the lasers and the optics. For an outdoor
display you need a very high power laser, which will literally melt standard
optics devices. There are special lens and mirrors that are used with high
power lasers, look at Edmund Optics. The deflection range is relatively
small, around 1 degree. The limiting factor is how fast you can move the
mirror, which depends on mass and inertia. With these small deflections you
can get pretty high rates.
Indoors with low light you can get away with much lower power, 10mW is more
than enough. With this power level you can use standard optics, and the
lasers are quite cheap < $30 as long as you like red. I've heard of people
using TI DMDs to deflect laser beams. Even the low end DMDs can display 1
bit raster images at 4000Hz.
One of the problems with this technology is it's hard to modulate the laser
intensity, which greatly restricts the range of colours you can produce.
From: cctalk [mailto:cctalk-bounces at classiccmp.org] On Behalf Of Mouse
Sent: July 19, 2016 4:47 PM
To: cctalk at classiccmp.org
Subject: Re: LASERS! && Freemont Street LED array (was Re: Cray J932SE (was
Re: Straight 8 up on Ebay just now))
>> You'd probably know, then - what's the fastest way to deflect a laser
>> beam? In particular, I'm wondering how practical it might be to take
>> a laser and turn it into a vector display on a handy blank wall [...]
> What bandwidth (deflection rate) do you need? Full scale in a
> microsecond? In 10 microseconds?
Well, if it takes longer than 100ms to replot the display, it will flicker
visibly, and the more under 100ms the better. In that time I'd like to draw
at least a couple hundred lines, though most of them will be short (line
length maybe 1-15% of corner-to-corner distance). What kind of
radians/second deflection rates this means depends on how far from the wall
you put the projector.
But, in terms of the bandwidth on the X and Y axis signals? If we say
200 lines at 25 ms replot (I get 20ms frame rate out of the cg6 for displays
significantly more complex than that - ie, with the cg6 the actual
limitation is the video signal vertical frequency), that's 125us/line.
Turning sharp corners is the hard part with mechanical deflectors like
mirrors, as it means very high acceleration of the mechanical parts. I
haven't done the math to be sure, but, until/unless taught otherwise by
testing, I'd feel dubious about clipping the X and Y signal bandwidths at
anything lower than ~1MHz.
/~\ The ASCII Mouse
\ / Ribbon Campaign
X Against HTML mouse at rodents-montreal.org
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
This email has been checked for viruses by Avast antivirus software.
More information about the cctalk