Circuit board material as optical waveguide?
HollyMinkowski
Posts: 1,398
I have been messing around with optical fiber and it struck me that
perhaps ordinary circuit board material could act as a waveguide for
very short distances at 850nm.
If you place a bright visible light LED behind circuit board material you can see that
the visible light goes right through it (weakly). So I was wondering if anyone
had ever tried transmitting data a few inches or so at 850nm using this method?
I realize that the detector would need to be quite sensitive but that seems
like it should not be insurmountable. The upside is that you could fabricate
chips that could transfer enormous amounts of data without using copper
pathways or adding an optical waveguide layer. Perhaps simply aiming the
850nm emitter and detectors straight down from the bottom of ICs might work
since the light seems to spread out in all directions (at least at visible wavelengths)
BTW: It is interesting that in a dark environment you can see dim red light coming
out of optical cable ends that are fed data at 850nm.... must be some kind of red shift
phenomenon dropping the freq.
perhaps ordinary circuit board material could act as a waveguide for
very short distances at 850nm.
If you place a bright visible light LED behind circuit board material you can see that
the visible light goes right through it (weakly). So I was wondering if anyone
had ever tried transmitting data a few inches or so at 850nm using this method?
I realize that the detector would need to be quite sensitive but that seems
like it should not be insurmountable. The upside is that you could fabricate
chips that could transfer enormous amounts of data without using copper
pathways or adding an optical waveguide layer. Perhaps simply aiming the
850nm emitter and detectors straight down from the bottom of ICs might work
since the light seems to spread out in all directions (at least at visible wavelengths)
BTW: It is interesting that in a dark environment you can see dim red light coming
out of optical cable ends that are fed data at 850nm.... must be some kind of red shift
phenomenon dropping the freq.
Comments
is the speed of light in a vacuum, but light travels through optical fiber at a slower speed.
By "waveguide", do you mean a channel cut into the PCB substrate? Or are you talking about transmitting the light through the material itself? FR4 is fairly opaque. Unless the light is quite intense, it is going to dissipate below usability within fractions of an inch. Now, if you were doing a waveguide where a channel is cut into the material, that might work for a few inches as long as the path is straight. But then you essentially just have a line of sight between transmitter and receiver. The issue for anything other than straight path is low reflectivity.
That makes sense about the shifted 850nm to visible.
I did mean actually using the existing board material to try and transmit
850nm signals. I googled and discovered that waveguide layers have been
added to boards using a clear plastic that can withstand the heat of wave
soldering...polysiloxane, the same clear plastic used to encapsulate LEDs.
It would be best I think to just make the board entirely of something like
polysiloxane, perhaps doping the plastic with reflective particles could also
help propagate 850nm signals by bouncing them all around.
The arguments for using optical data comm on an ordinary pc board is similar to the
reasons people use wifi instead of wiring with copper ethernet cable. They do
this even though the signals require high power levels and the receivers must
be incredibly sensitive. I could determine the db loss per cm in common pc board
material at 850nm if I had the right equipment, I just wondered if anyone had
heard of others trying this idea. The amount of copper on circuit boards is just
getting so out of hand...so many layers carrying so much data that could
be better moved at 850nm. Perhaps a polysiloxane board carrying little more than power
on copper traces would be liberating.
Opticians know this method is entirely workable because a fiberglass epoxy substrate board can be fabricated to (more) readily transmit light. The entire substrate can be used as the medium and handle multi-channels of light transmissions.
The beams can move outward from the chips to corresponding locations called foci which have receptors or wavelength receivers. This is has some basis in the optical computer and optoelectronic circuits.
You can use directional optolectronics embedded in the substrate and/or feed directional light into the existing "off the shelf" substrate and/or use tiny lenses to achieve diversion, distance, intensification and focus.
If using the "entry" method of light to board, use of Snell's Law will help define the layout.
Snell's Law? Cool!
http://www.endmemo.com/physics/snells.php
I just ordered a nice looking book on optics.... gotta start somewhere when
you get the urge. And I now have some free time to delve into new stuff.
It is important to know the optical characteristics of the substrate which can be looked up in the Book of Physical Properties. The program could achieve total internal reflection, which is really cool, by adjusting the parameters.
In the case of the fiberglass printed circuit board, the beam could be made to do "internal reflections" off the "internal surface" to better define travel and positioning.
Internal cylinders or glass posts are sectional spherical in nature when hit with a beam slice and can serve as "lens" beams to new places in space and achieve locations internally.
If you have a bubble inside the substrate, you can get a diversion beam reflected to flood an internal area. So you may want to find boards intentionally with specific bubbles.
Circuit boards full of bubbles? .... I love it!
I was thinking about the fact that 850nm light shifts down to where
you can see a faint red glow after it passes through some optical cable.
SteveL said it was because it was slowed due to passing through a medium
more dense than a vacuum...that makes sense. But it struck me that light
passing through an enormous amount of space might also shift down as
even so called empty deep space might have a very slight amount of
matter in it? Could this be the reason that light from distant galaxies is
shifted toward the red? I saw a TV docu that said the shift was due to the
distant galaxies moving away from us at enormous speed...but could it
instead be because light has slowed because it has passed through such
an enormous distance to reach us and the vacuum of space is not a
perfect vacuum but contains a bit of something? If this was the reason for
the frequency shift I would think that the galaxies farthest from us would
have the greatest shift and those nearer would have a lesser amount. Not
due to the farthest object moving away faster but just because the light passed
through more space and the frequency thus shifted further. Hmmmm
It takes true confidence to judge a book by it's cover, and shouldn't be done generally, unless you are Holly... Bless You Holly, You are a Tinker Fairy of the best kind...
Not trying to HiJack your thread by anymeans, and I want to contribute more than just praise, it's just..
It went from fractions of Nano Meters, to, Light Years... I am going to try and keep up, I really am...:nerd:
Again Bless You Holly. I can't stop grinning...
I was all ready to ask if You had checked out the EdmundOptics web site? but then We blasted off into space...
And that's ok too...I just gotta be a little bit faster next time..
I can go a long way on compliments like those....maybe even all the way to my next manic phase... Lol :-)
Take a look at EdmundOptics.com, they have lots of optic stuff, the latest seems to be BioPhotonics..
They are from the old Edmund Scientific, if the name sounds familiar.
'
I know your looking at using a PCB core, But the fiberglass used is non directional for strength (often called Matt)
'
I have dealt with a lot of plastic fiber-optic cables and they suck. I have just pinched the plastic fiber and killed all the com.s on it.It acts like coax in a RF application, "Kink" the coax and you have a bunch of standing waves.
'
I have a nice new Fluke power meter that I test with.
'
embedding glass fiber-optic into a fiberglass PCB should make for a ballistic optical bus.
all be made so that they could communicate at high speed optically. Using the
circuit board itself as the optical waveguide would allow you to plunk the parts down
anywhere and know they were all able to talk to one another. Many parts would
only need to be connected to a source of power to function...2 pins instead of many
pins would make it easier to design a project.