Searching \ for 'Mechanically Reading A Digital Display' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/index.htm?key=mechanically+reading
Search entire site for: 'Mechanically Reading A Digital Display'.

Truncated match.
PICList Thread
'Mechanically Reading A Digital Display'
1996\05\09@110112 by Martin McCormick

flavicon
face
       I am puzzling over the best method for electronically reading a LED
display without actually modifying the device containing the display.  On
the surface, it sounds simple and it would be if one could find the right
photo cell array.

       The closest thing I have seen to what would do the job is a TI
CCD chip that contains 64 photo cells in a line.  This would be perfect
except that one must mechanically scan at right angles to the line of cells
in order to see all parts of the display which greatly complicates the
mechanics of the system (cost).

       What I need is a photo cell array which is physically about the
size of the LED array I am reading.  I even entertained the idea of using
a similar LED array and using the individual segments as photo pickups, but
it might not be possible to read all segments since one would actually be
reading backwards.  If the photo arrays were cheap enough, one could construct
a mirror image of the display and not have to have any moving parts.  Another
thought was to have an addressable LCD matrix and cause a little spot to
open up or become opaque and read the display by scanning this spot, sort of
like the old experiments in television in which the scanning was done by a
spinning disk with holes in it.

       I realize that there are many differences in size and even shape
of the characters in LED displays so the best solution would be some sort of
matrix that had enough resolution that you could train your decoder as to
what a given display looked like.

       What I don't want to do is to try to build precision optics or
mechanics.  That's what we have PIC's and other microcontrollers for.:-)

       I have been kicking this around in my brain for quite some time so
this is not something that has to be done today.  I am mainly interested in
whether there is anything new that might get this project off and running.

Martin McCormick WB5AGZ  Stillwater, OK 36.7N97.4W
OSU Center for Computing and Information Services Data Communications Group

1996\05\09@131838 by Mark A. Corio

picon face
Why do you want to read the LEDs?  Is this going into a product of some kind?

Mark A. Corio
Rochester MicroSystems, Inc.
200 Buell Road, Suite 9
Rochester, NY  14624
Tel:  (716) 328-5850
Fax:  (716) 328-1144
e-mail:  spam_OUTMcorioTakeThisOuTspamaol.com

***** Designing Electronics For Research & Industry *****

1996\05\09@153512 by John Safrit

flavicon
face
       I've also be looking into doing this.  I'd like to increase our
    products automated test coverage to verify the state of our status
    LEDs while still in the chassis.  If anyone has done this I'd be
    interested in hearing about it.
                               John Safrit
                               .....safritKILLspamspam@spam@wg.com


______________________________ Reply Separator _________________________________
Subject: Mechanically Reading A Digital Display
Author:  pic microcontroller discussion list <PICLISTspamKILLspammitvma.mit.edu> at
WG-RAL-SMTP
Date:    5/9/96 11:18 AM


       I am puzzling over the best method for electronically reading a LED
display without actually modifying the device containing the display.  On
the surface, it sounds simple and it would be if one could find the right
photo cell array.

       The closest thing I have seen to what would do the job is a TI
CCD chip that contains 64 photo cells in a line.  This would be perfect
except that one must mechanically scan at right angles to the line of cells
in order to see all parts of the display which greatly complicates the
mechanics of the system (cost).

       What I need is a photo cell array which is physically about the
size of the LED array I am reading.  I even entertained the idea of using
a similar LED array and using the individual segments as photo pickups, but
it might not be possible to read all segments since one would actually be
reading backwards.  If the photo arrays were cheap enough, one could construct
a mirror image of the display and not have to have any moving parts.  Another
thought was to have an addressable LCD matrix and cause a little spot to
open up or become opaque and read the display by scanning this spot, sort of
like the old experiments in television in which the scanning was done by a
spinning disk with holes in it.

       I realize that there are many differences in size and even shape
of the characters in LED displays so the best solution would be some sort of
matrix that had enough resolution that you could train your decoder as to
what a given display looked like.

       What I don't want to do is to try to build precision optics or
mechanics.  That's what we have PIC's and other microcontrollers for.:-)

       I have been kicking this around in my brain for quite some time so
this is not something that has to be done today.  I am mainly interested in
whether there is anything new that might get this project off and running.

Martin McCormick WB5AGZ  Stillwater, OK 36.7N97.4W
OSU Center for Computing and Information Services Data Communications Group

1996\05\09@234730 by Bill Cornutt

flavicon
face
----------
>        I am puzzling over the best method for electronically reading a LED
>display without actually modifying the device containing the display.  On
>the surface, it sounds simple and it would be if one could find the right
>photo cell array.

While a LED is a light source, it can also be used as a light detector.
Place a led display (used as a detector) over the led display, and
see what detects.  It may be necessary to use a 5x7 matrix led
for each digit as a detector.

Bill C.

PS.  You may need mirrors.

1996\05\10@083035 by Mark A. Corio

picon face
Martin,

Do you intend to make a universal device that gets moved around alot?
If not, I would suggest using seven discrete photodiodes or phototransistors
for each segment mounted to allign with the center of each segment.  This
approach would require a different arrangement for any given LED readout
(size and shape) but should be much cheaper (per application) than a CCD
camera, etc as has be suggested.

Another question is how do you activate the read function?  Does a clock
announce the time at each change?  Do you push a button to get a reading?
Mark A. Corio
Rochester MicroSystems, Inc.
200 Buell Road, Suite 9
Rochester, NY  14624
Tel:  (716) 328-5850
Fax:  (716) 328-1144
e-mail:  .....McorioKILLspamspam.....aol.com

***** Designing Electronics For Research & Industry *****

1996\05\10@094756 by n/a

flavicon
picon face
Hi,

  If you want to keep it simple and low cost, why not construct a tool of
10 or so short lengths of fibre optic cable alligned in a straight line such
that the ends could be "wiped" across the led display. Each cable length
would have a phototransister attached to the other end, so that it would
only sense light immediately beneath it. Since the outer cables would not
detect any light at all (assuming that the display was smaller), the height
of the display could immediately be measured. Once the height of the display
was known, the individual segments of the seven segment display characters
could be sensed. Any upward/downward variation as the tool was wiped across
the display would be detected by the outer optic cables, so the processor
could keep track of its movement across the display.

         Geoff.

PS. Do I get a share of the royalties.

1996\05\10@103956 by Martin McCormick

flavicon
face
       First, I want to thank everybody who has provided such good ideas
on different ways to read an electronic display.  One person had what sounds
like a wonderful suggestion, that of using one of those relatively cheap
hand scanners which are usually connected to a PC.  If the hand unit contains
some type of shaft encoder to send horizontal movement, then somebody's already
done the dirty work.


       Has anybody analyzed the signals on the cable link between the hand
unit and the peripheral card?  My project idea would be for a portable unit
that would use PIC's and maybe a 68HC11 to handle the signal processing.

       Another good suggestion was to turn off the LED's in the hand scanner
which would normally be used to illuminate the paper.  If the scan head
has a separate line for the light source, then one could turn it on for
LCD's and off for LED's or back-lit LCD's.

       I would like to also contribute an idea to the companion thread
dealing with reading of utility meters.  The good point was made that the
present meter design is very robust.  I had a part-time job in graduate school
in a department that trained workers for jobs in the electrical power
distribution industry.  I got to examine a 1910-ventage electric meter and
found it to be a pint-sized version of the ones that are on most houses
today.  It had the same mechanical construction and principle of operation
of today's meters.  The only difference was that it was a single-phase meter
and it may have not had as many pointers, but the gear train was otherwise
the same.

       Why not simply put bar codes on one of the modern meters that has
wheels instead of pointers and electronically read them so that the reading
value is still mechanically stored and can not be zapped by a good lightening
hit or a bad customer.

Martin McCormick 405 744-7572   Stillwater, OK
OSU Center for Computing and Information services Data Communications Group

1996\05\10@120336 by Paul Haas

flavicon
face
On Fri, 10 May 1996, Martin McCormick wrote:

>         Why not simply put bar codes on one of the modern meters that has
> wheels instead of pointers and electronically read them so that the reading
> value is still mechanically stored and can not be zapped by a good lightening
> hit or a bad customer.

I think the part on the meter is called a "sending unit".

My house has one of those on the water meter.  It senses the the least
significant wheel and sends one pulse every 1000 cubic feet to a unit on the
outside of the house.  The meter readers walks up to the outside of the house
and enters in the reading to portable computer. Once a year they have me read
the inside meters to verify that it is still in sync with the outside meter.

I can't tell how it reads the wheel.  I've seen similar similar sending units
on electric and gas meters.  It looks like the units are built into
replacement plastic or glass faceplates.  To install one, they break the seal
on the meter, replace the old faceplate with the sending unit, run a cable
outside the house and install the readout module.

In a neighboring city they have 1/2 inch thick, inch diameter black plastic
targets on the outside of houses connected to the electric meters. I assume
they inductively couple to the meter readers computer.  In the city where I
work they have 20 prong male connecters on the outside of buildings.  I
assume they go to some internal meters.

Although I'm sure some of the units have PIC's in them, we are still way
off topic.

====

Back to display reading, the Connectix Quickcam costs less than $100 and has
a PIC as the interface to a PC parallel port.  This would be easy to
interface to another PIC or a 6811.  It is a greyscale camera providing up to
320x240 pixels and 16 or 64 shades of grey.

If you design an interface, remember that you don't have to use the whole
image.  If your algorithm wants 320x1 pixels or 32x32 or 80x60, the camera
can do that.   When people say that a CCD camera needs a big computer for
image processing, they are making the assumption that you have to use
every pixel.

--
Paul Haas
EraseMEpaulhspam_OUTspamTakeThisOuThamjudo.com

1996\05\10@150716 by John Payson

flavicon
face
> Back to display reading, the Connectix Quickcam costs less than $100 and has
> a PIC as the interface to a PC parallel port.  This would be easy to
> interface to another PIC or a 6811.  It is a greyscale camera providing up to
> 320x240 pixels and 16 or 64 shades of grey.

What do you know of this device?  Is it easy to process information from it?
Could a small micro such as a PIC do piece-wise data procesing based upon
a snapped picture?  Can the same picture be read out more than once?  I've
had a couple of ideas for PIC data acquisition, and a camera like that would
be great if it would work.

1996\05\10@182358 by William Chops Westfield

face picon face
   > Back to display reading, the Connectix Quickcam costs less than $100
   > and has a PIC as the interface to a PC parallel port.

   What do you know of this device?  Is it easy to process information
   from it?  Could a small micro such as a PIC do piece-wise data
   procesing based upon a snapped picture?

The interface protocol is proprietary.  Although claims say that connetix
has loosened up recently, giving specs to anyone who signs up for their
developer program and signs a non-disclosure agreement, I don't know how
much is really happening there.  I submitted an application for the
developer program, and haven't heard ANYTHING back in about a month (nor
to the email I sent two weeks ago asking what was up.)

Meanwhile, there are gnu-types trying to reverse-engineer the protocol
so that code can be distributed in source form and such without violating
the non-disclosure agreement.  I believe there's already a Linux driver,
for example.  A webcrawler search for "QuickCam" will turn up all sorts
of interesting stuff (including Connetix themselves.)

BillW

More... (looser matching)
- Last day of these posts
- In 1996 , 1997 only
- Today
- New search...