A trinary signal?
Archiver
Posts: 46,084
I require some way to send trinary data between devices, I was
thinking about using stamps, but I have run into problems with
electronic forms because of the requirement of the common ground. I
was wondering if there would be a way to encode the signal in
multicolored light and send it through fiber to a light sensing cell
matrix, CDS, or phototransistor group. The application of this is
rather complicated and would take me a couple hours to explain, so
please don't ask me to. This is the last component I require for
possibly one of the most random ways to use a basic stamp ever
thought
of! Anyone who can help will be mentioned with the release of
schematics and documentation (hopefully within the next few months if
I can get some sponsors as well!). I have considered tri-color led's
because the signal would only have to pass a few inches, but the
led's would need to emit at an angle into the fiber so that any
signal coming behind them would not be stopped.
/``\
\tri\
\led\
[noparse][[/noparse]signal]--->
(Fiber Line)
I really do not want to have to detect the signal behind the emitter
and emit it again, because that would slow down too much. I am
guessing the signal could be emitted at the same rate it is detected
so there would be no need for slowing the signal down and it may
actually corrupt the data to do so.
I also need to be able to split a signal into a different direction
if it has a certain leading three digit code, so there would have to
be three switches in line that each detect the first bit and
retransmit the rest of it. these switches would need to stop the
signal so that the next switch could look at only the second bit,
which would have become the first bit after the first switch took the
first bit out and passed the signal. if the switches do not find what
they are looking for, however, they must reconstruct the signal and
send it to the next line. So you see the problem? The signal must
only get stopped when it reaches a destination, unless I use a stamp
to emulate all three switches, and just retransmit the signal if it
is not the correct one. This would be rather lengthy, wouldn't it be?
I need it to be optimized for speed. Another alternative may be a
chemical signal but I have found absolutely no consolation that
anything yet exists which could do what I require.
Any idea what exactly would work to sense the light color or
whatever you may have an idea about?
I am also currently on a limited budget, so this is purely
hypothetical, I would also like if someone could emulate this in
software so that I can finish making a prototype emulation of it on
my computer. I like to work in C because of the unlimited access to
resources, but any language would be helpful. I am well versed in
over 7 of the mainstream languages.
-Thanks a lot for any help one of you can give!
BTW: This design is patent pending so please dont try to claim it as
your own, or build it before I can gather resources.
thinking about using stamps, but I have run into problems with
electronic forms because of the requirement of the common ground. I
was wondering if there would be a way to encode the signal in
multicolored light and send it through fiber to a light sensing cell
matrix, CDS, or phototransistor group. The application of this is
rather complicated and would take me a couple hours to explain, so
please don't ask me to. This is the last component I require for
possibly one of the most random ways to use a basic stamp ever
thought
of! Anyone who can help will be mentioned with the release of
schematics and documentation (hopefully within the next few months if
I can get some sponsors as well!). I have considered tri-color led's
because the signal would only have to pass a few inches, but the
led's would need to emit at an angle into the fiber so that any
signal coming behind them would not be stopped.
/``\
\tri\
\led\
[noparse][[/noparse]signal]--->
(Fiber Line)
I really do not want to have to detect the signal behind the emitter
and emit it again, because that would slow down too much. I am
guessing the signal could be emitted at the same rate it is detected
so there would be no need for slowing the signal down and it may
actually corrupt the data to do so.
I also need to be able to split a signal into a different direction
if it has a certain leading three digit code, so there would have to
be three switches in line that each detect the first bit and
retransmit the rest of it. these switches would need to stop the
signal so that the next switch could look at only the second bit,
which would have become the first bit after the first switch took the
first bit out and passed the signal. if the switches do not find what
they are looking for, however, they must reconstruct the signal and
send it to the next line. So you see the problem? The signal must
only get stopped when it reaches a destination, unless I use a stamp
to emulate all three switches, and just retransmit the signal if it
is not the correct one. This would be rather lengthy, wouldn't it be?
I need it to be optimized for speed. Another alternative may be a
chemical signal but I have found absolutely no consolation that
anything yet exists which could do what I require.
Any idea what exactly would work to sense the light color or
whatever you may have an idea about?
I am also currently on a limited budget, so this is purely
hypothetical, I would also like if someone could emulate this in
software so that I can finish making a prototype emulation of it on
my computer. I like to work in C because of the unlimited access to
resources, but any language would be helpful. I am well versed in
over 7 of the mainstream languages.
-Thanks a lot for any help one of you can give!
BTW: This design is patent pending so please dont try to claim it as
your own, or build it before I can gather resources.
Comments
data along fiber line already implemented in the industry, but why not
use a similar method configured for use with copper connections? If
you only need half-duplex, then why not set specific analog "plateaus"
for your -1 0 1 trinary signal?
i.e. Transmit
R
+5V >--/\/\
o
|
5R |
+5V >--/\/\--o |
| | R
o-->|---o--/\/\--o
| |
Stamp I/O >--o o---> Out
| R |
o--|<---o--/\/\--o
|
R |
GND >
/\/\
o
Truth Table:
I/O pin Out
-1 (LOW) 1.67 Volts
0 (INPUT) 2.61 Volts
1 (HIGH) 3.33 Volts
On the receiving end, all you need are two comparitors...
i.e.
10K
+5V >----/\/\----> GND
|
|
Vref1
10K
+5V >----/\/\----> GND
|
|
Vref2
|\
Vref1 >
|-\
| \
>|---o
o--|+/ |
| |/ |
Input >--o o---> Out
| |\ |
o--|-\ |
| \
|<---o
Vref2 >
|+/
|/
Vref1 and Vref2 determine a voltage threshold in which the comparitor
turns on or off. The only thing is, is that you need to select a
comparitor capable of driving both rails, some do not.
>I require some way to send trinary data between devices, I was
>thinking about using stamps, but I have run into problems with
>electronic forms because of the requirement of the common ground. I
>was wondering if there would be a way to encode the signal in
>multicolored light and send it through fiber to a light sensing cell
>matrix, CDS, or phototransistor group. The application of this is
>rather complicated and would take me a couple hours to explain, so
>please don't ask me to. This is the last component I require for
>possibly one of the most random ways to use a basic stamp ever
>thought
>of! Anyone who can help will be mentioned with the release of
>schematics and documentation (hopefully within the next few months if
>I can get some sponsors as well!). I have considered tri-color led's
>because the signal would only have to pass a few inches, but the
>led's would need to emit at an angle into the fiber so that any
>signal coming behind them would not be stopped.
>
> /``\
> \tri\
> \led\
>
>[noparse][[/noparse]signal]--->
>
>(Fiber Line)
>
>I really do not want to have to detect the signal behind the emitter
>and emit it again, because that would slow down too much. I am
>guessing the signal could be emitted at the same rate it is detected
>so there would be no need for slowing the signal down and it may
>actually corrupt the data to do so.
>
>I also need to be able to split a signal into a different direction
>if it has a certain leading three digit code, so there would have to
>be three switches in line that each detect the first bit and
>retransmit the rest of it. these switches would need to stop the
>signal so that the next switch could look at only the second bit,
>which would have become the first bit after the first switch took the
>first bit out and passed the signal. if the switches do not find what
>they are looking for, however, they must reconstruct the signal and
>send it to the next line. So you see the problem? The signal must
>only get stopped when it reaches a destination, unless I use a stamp
>to emulate all three switches, and just retransmit the signal if it
>is not the correct one. This would be rather lengthy, wouldn't it be?
>I need it to be optimized for speed. Another alternative may be a
>chemical signal but I have found absolutely no consolation that
>anything yet exists which could do what I require.
>
>Any idea what exactly would work to sense the light color or
>whatever you may have an idea about?
>
>I am also currently on a limited budget, so this is purely
>hypothetical, I would also like if someone could emulate this in
>software so that I can finish making a prototype emulation of it on
>my computer. I like to work in C because of the unlimited access to
>resources, but any language would be helpful. I am well versed in
>over 7 of the mainstream languages.
>
> -Thanks a lot for any help one of you can give!
>
>BTW: This design is patent pending so please dont try to claim it as
>your own, or build it before I can gather resources.
I'm not exactly sure about that... And besides, you just made it public
domain, thus negating even your patent holding rights.... Have a great
day! [noparse]:o[/noparse])
Beau Schwabe Mask Designer IV - ATL
National Semiconductor Wired Communications Division
500 Pinnacle Court, Suite 525 Mail Stop GA1 Norcross, GA 30071