Idea: LED Wearable Display

O.K. an 8 by 8 array of .11 inch diameter 2k mcd LEDs would be under an inch square. It would be at least an inch deep with the drivers, 8 latches, and uC like a PIC with one semi high power 8-bit port (C) for the data and one for the 8 latch enables (B), 2 pins for serial (port A), 1 for the indexing pulse from the mirror. Mounted at a slight angle, it would produce 64 scan lines. Any bigger and you might as well mount a camcorder viewfinder.

That will be 20mA x 64 = 1.28 AMPs if fully driven (which they will never be) but in that compact arrangement, heat dissipation may be an issue for the LEDs and for the driving transistors (cause the PIC sure as hell won't handle it right?)

If we mount two mirrors (one on either side of the motor shaft) and tilt the back one so that it shifts the image up or down one half pixel, we can get 128 lines interlaced. That's about 16 lines of text. The interlacing will not affect brightness (right?) because the image wouldn't have been present through that part of the arc anyway. So... if we build a very well balanced "mirror box" with 4 slightly different angles on the sides, we get 256 lines and 8 sides gets us 512!

I don't have a clue how to build something that precise and balance it so well that the vibration won't drive the wearer nuts in a few hours. <GRIN> Also, the LEDs are now on more and more so they pull more and heat more. One of these days I've got to just get a pager motor and some little front surface mirrors and some glue and just see what type of useable visibility different types of LEDs produce vs. how much power/heat they give off.

James Newton,   



-----Original Message-----
From: Mark Willis [mailto:spammwillisspam at nwlink.com]
Sent: Tuesday, August 24, 1999 12:08 AM
Cc: Mark Willis
Subject: Re: Wearable displays


I try <G>

Newark has 2.0x1.25x0.8mm Orange or Amber SMD LED's, 65 mcd, $0.47 each.

If 65 millicandelas isn't enough, what do you want?  <G>

I STILL think, 'THE' way to do this is to use existing 2000 mcd LED's in
arrays, and mirror/Lens focus several of these onto a single "emitter"
array (i.e. focus 4 groups of 12 of these, onto 1 48-pixel emitter
array), giving you 48 pixels, scan that 10 times & that's 480 pixels. 
THAT would be pretty nice, methinks.  You'd just need some minor current
adjustments to make all pixels identical (Quite do-able automatically in
software, I think.)  It's a thought <G>

  Mark

James Newton wrote:
> 
> Hey! The SMD idea has merit. They can me mounted in a descending row (just
> like the MAN 2 with the slightly twisted mounting) to produce a very dense
> vertical array, and they could all be driven at once (not matrix) to
> increase the brightness. The only division is for the columns and for the
> scan (if using a rotating mirror) A quick search does not return any high
> mcd surface mount versions however. Do you know of any?
> 
> James Newton, webmaster http://get.to/techref
> jamesnewton@geocities.com <mailto:spamjamesnewtonspam at geocities.com>
> 1-619-652-0593 phone
> 
> -----Original Message-----
> From: Mark Willis [mailto:spammwillisspam at nwlink.com]
> Sent: Wednesday, August 18, 1999 12:19 PM
> To: eplus1@san.rr.com
> Cc: Mark Willis
> Subject: Re: Wearable displays
> 
> Well, the 300 DPI LED array was an idea, anyways <G>  I know you can
> find SMD LED's for pretty reasonable numbers, make up your own array at
> a little wider spacing, maybe?  I know about over-driving LED's, it's
> just plain easier to do it a different way IMO, go for what works for
> you though <G>
> 
> If I wanted a REALLY bright display, I'd get some of those 8000 mcd Red
> LED's & set those up to be lens-focussed onto a row of small spots -
> then use those small spots as light sources, I think I'd have a REALLY
> bright display <G>  I think you could make something that'd be quite
> visible in bright sunlight, if you over-drove those monsters...  Also
> could use a laser diode onto one (or more) spots & then scan with what
> reflects/transmits through those...
> 
>   Mark
> 
> James Newton wrote:
> >
> > Right, the Okidata LED array printers... Much nicer, but... I used to work
> > for a company that OEM'd one of those printers and... the array is a solid
> > unit, so.... you would have to cut it (Not a big problem I guess) and... I
> > tried to beg, buy or steal one and I couldn't get them for anything under
> > the price of the printer itself! (approx. $300, $180 in quantity). The
> > explanation I got was that Okidata had financed the development and had an
> > exclusive contract with the supplier. They don't want other (small)
> > companies competing after they spent the money to develop the display. There
> > just aren't that many applications so no one else is developing this. Please
> > let me know I you have found a better source. Its still less than the
> > cheapest commercial LCD or etc... display of the same resolution.
> >
> > But the MAN 2's are about $2.
> >
> > Want to know the real killer on either the MAN 2's or the printer Array?
> > Intensity. Since they are matrix connected, only 7 can be powered at any
> > instant. You have to scan the other 4 sets of 7. This divides the effective
> > intensity by 5. Then you have to scan the scan to generate the columns on
> > the final display. This divides the intensity by the number of columns. If
> > you want a 4 by 20 display, you need 20 times 6 (min) or 120 columns. Now
> > you have divided the intensity by 600. Its better with the printer array
> > because you don't have the divide by 5, but its worse because the printer
> > array is no where near as intense as the MAN 2. Or at least it is not driven
> > intensely in the printers. And this assumes that your mirror or orifice or
> > what ever is only scanning the area of the final display. The best way to
> > make a mirror scan an area *without vibration* (on your eyeglass frame) is
> > to spin it 360 degrees on a pager motor (very smooth with the little off
> > center weight removed). Now we have to add a division by what part of the
> > full circle the display area covers. Divide by four at least. 1 in 2400 the
> > original intensity of the LED.
> >
> > I have found, however that you can drive MAN 2's at 12 volts (direct... no
> > wimpy resistors for me <GRIN>) for a few seconds before they slag <EVIL
> > GRIN> and (oh boy!) are they bright! May be able to do the same with a
> > printer array. P junction cooling?
> >
> > A laser diode could put out the required intensity, but then you have to
> > scan in 2d and getting two little mirrors to sync at those speeds is, I
> > guess, damn hard <GRIN>. Anybody know of an array of laser diodes?
> >
> > ....small enough to actually mount comfortably on my eyeglass frame.
> >
> > Interesting to think about out loud.
> >
> > James Newton, webmaster http://get.to/techref
> > jamesnewton@geocities.com <mailto:spamjamesnewtonspam at geocities.com>
> > 1-619-652-0593 phone
> >
> > -----Original Message-----
> > From: Mark Willis [mailto:spammwillisspam at nwlink.com]
> > Sent: Tuesday, August 17, 1999 2:17 PM
> > To: James Newton
> > Cc: Mark Willis
> > Subject: Re: Wearable displays
> >
> > Sounds like the HARD way to do it <G>, but do-able.  I suppose you could
> > use shift registers to clock the data in to the LEDs, over time, if you
> > could figure out the timing, too.  I still think the easy way to do that
> > is a 50-LED linear array off an LED printer (at 300 DPI), using 48 of
> > those LED's (with 10 scanning orifices), I'll have a nice mono VGA green
> > display that's an effective 1.5" tall by 2.5" wide or so, which would be
> > fun <G>
> >
> >   Mark
> >
> > James Newton wrote:
> > >
> > > Mark, I overheard that you were looking at adapting a heads up display for
> > > wearable computing. I've had this (other) idea for wearable displays for a
> > > while and just though I would ask you what you think.
> > >
> > > Remember the MAN 2 LED arrays? 5x7 little points of light LED's in a single
> > > case? How about this: Mount the display slightly off center so that as a
> > > mirror scans past it the right most LED in the first row appears just above
> > > the left most LED in the next row down. Let me try that another way. Draw a
> > > perfectly horizontal line through each LED in the display. Now rotate the
> > > display just a titch to the right so that the horizontal lines are equally
> > > spaced. Makes 35 scan lines. 4 or 5 lines by many (??) columns for text.
> > >
> > > Now, using a high speed processor (Scenix 50Mhz or a fast PIC) turn the
> > > LED's on and off as the mirror spins past the display so that the left most
> > > LED doesn't come on for the first pixel in the first row of the picture we
> > > are trying to display until it reaches the position that the right most LED
> > > in the first row was just below when it came on to display the first pixel
> > > in the second row of the picture. This is hard to explain. Basically, time
> > > delay the turn on of the left columns until they are aligned with where the
> > > right columns WERE when they displayed a given column of pixels.
> > >
> > > The MAN 2's are connected in a matrix so an external driver chip may be
> > > necessary to scan the rows as we are changing the display. Don't use
> > > resistors so that each of the LED's is WAY overdriven when its on since we
> > > will only be turning it on for a very brief period (God help us if the
> > > display sticks on)
> > >
> > > James Newton, webmaster http://get.to/techref
> > > jamesnewton@geocities.com <mailto:spamjamesnewtonspam at geocities.com>
> > > 1-619-652-0593 phone

Omnidirectional Scanning

http://www.hut.fi/~then/mytexts/lasercontrol.html

  .