[sci.electronics] Optical interconnects

reddy@mips.com (T.S. Reddy) (03/08/91)

    Has anyone tried to make their own fiber optic interconnects? From
what I know, the optical input/output receptacles on audio/video products
for most brands are similar, so I would think that standardized parts for
fashioning such a beastie would be available. I am not familiar with
the techniques involved in cutting/splicing fiber optic cables, but
if anyone could enlighten me, I would be grateful.

Thanks in advance.
-- 
T.S.Reddy
e-mail: reddy@mips.com

lstowell@pyrnova.pyramid.com (Lon Stowell) (03/08/91)

In article <782@spim.mips.COM> reddy@mips.com (T.S. Reddy) writes:
>
>    Has anyone tried to make their own fiber optic interconnects? From
>what I know, the optical input/output receptacles on audio/video products
>for most brands are similar, so I would think that standardized parts for
>fashioning such a beastie would be available. I am not familiar with
>the techniques involved in cutting/splicing fiber optic cables, but
>if anyone could enlighten me, I would be grateful.
>

   This is an area in which professional fiber installers prefer
   not to fool around.  Most just buy preterminated fiber
   jumpers and use optical splices to connect these to the main
   fiber runs.

   You COULD terminate your own...there are kits available in
   all classes of cost and real value.  The good ones have
   optical epoxy, polishing materials....and most importantly, a
   good inspection microscope...some even have optical power
   meters (although that would be overkill).  

   Optical termination is an acquired skill...if you have any
   lapidary background it would help.  Expect to waste a lot of
   fiber and connectors until you gain the manual skills
   required  (at which point you have a salable skill....).

   Beginners would likely get the optical equivalents of cold
   solder joints, loose connections, etc....how much these would
   really affect sound quality I cannot say...although some old
   comparison tests between internal D/A convertors and HIGHER
   RESOLUTION external D/A's showed that typically there was
   significant impairment with the external convertors....I
   would have to blame poor interconnects.          

lamb@brahms.udel.edu (Richard E Lamb) (03/08/91)

Is the question here to use OPTICAL interconnections between "chips".

    Or "EMP hardened" communications system between missle silos?

erk@americ.UUCP (Erick Parsons) (03/08/91)

>In article <782@spim.mips.COM> reddy@mips.com (T.S. Reddy) writes:
>
>    Has anyone tried to make their own fiber optic interconnects? From
>what I know, the optical input/output receptacles on audio/video products
>for most brands are similar, so I would think that standardized parts for
>fashioning such a beastie would be available. I am not familiar with
>the techniques involved in cutting/splicing fiber optic cables, but
>if anyone could enlighten me, I would be grateful.

I've never fooled around with fiber optics myself but have heard horror
stories concerning splicing/repairing these beasties. I occasionally work
with the USA division (Underground Service Alert) where I work and talk
regularly with employees doing the same for the telephone company and
other utilities. There was reportedly a telephone Fiber Optic cable that had
been hit by a backhoe which required splicing, I'm a little fuzzy on
the details as it was about a year ago, but I do remember reeling when I
heard the cost to resplice this baby, 6 digits ! They (Whoever splices)
do the splicing inside a specially constructed truck that is for all
intents and purposes -air tight-. They also at some point in the splicing
have to use a microscope to align the fibers to avoid losses in the strand.

The only clue I have as to how many fibers might be in a bundle would
be from the conduit that I've seen and the fact that you have to splice
the fibers with a microscope. The conduit they used for the TransAmerica
(for lack of a better name) cable is about 8-10" in diameter. It is
divided inside into 4 sections internally which would give each section
an area of less the 12 sq. inches. probably about 8. Unfortunately I
did not see the cable being pulled in but given the expense of running
a conduit from San Fransisco to Washington D.C. I would guess that
they are using about 40 - 50 % of each section for fibers with about
5 - 10% of that being sheath. The fibers are about 3 mils in diameter
each. Worst case given these guesses would put the number of fibers in
each cable at about 1200 fibers. This particular line BTW is burried
in concrete, pipe cable and all.

1200 fibers as a guess for the cable mentioned in the first paragragh
at over 100,000 dollars to repair !

This should be a plausible way to make money ;-)

P.S. This is just a guess based on what I've heard from others and from
what I've seen, read. Any and All corrections gladly accepted.

>Thanks in advance.
>T.S.Reddy

For a good understanding of the basics and some applications you might
want to pick up the book Understanding Optronics at your local Radio
Shlack. These books are written for Texas Instruments and are dirt cheap
for what you get. That is if they are still available.

--
 -------------------------------------------------//-------------------------
 (ames att sun)!pacbell! ----> sactoh0!pacengr!americ!erk     Multitasking,
  ucbvax!ucdavis!csusac! --/			//	      Never leave
	     uunet!msac! -/   Erick Parsons \Sacramento Ca    Home Without it
 --------------------------------------------\X/-----------------------------

sherwin@cbnewsi.att.com (behzad.shahraray) (03/09/91)

In article <782@spim.mips.COM> reddy@mips.com (T.S. Reddy) writes:
>
>    Has anyone tried to make their own fiber optic interconnects? From
>what I know, the optical input/output receptacles on audio/video products
>for most brands are similar, so I would think that standardized parts for
>fashioning such a beastie would be available. I am not familiar with
>the techniques involved in cutting/splicing fiber optic cables, but
>if anyone could enlighten me, I would be grateful.
>
>Thanks in advance.
>-- 
>T.S.Reddy
>e-mail: reddy@mips.com

There are several types of connectors that can be used on an optical
fiber. Some involve the use of index matching gels at the interface
to minimize the loss. The type of connectors that are used in audio
are different and in order to minimize the loss the surface of the 
fiber has to be ground and polished. In the process on cutting, grinding
and polishing the fiber several things can go wrong and you can endup with
chips, scratches or cracks in the fiber. So you need to inspect the fiber
end under high magnification (microscope) to check for these defects
(I worked on computer vision algorithms to perform the inspection task
for a couple of  years).
In short, making your  own interconnects is not impossible if you
have the tools, but it is much harder than making your own
electrical interconnects.

----------------------------
Behzad Shahraray
Machine Perception Research Dept.
AT&T Bell Labs.
behzad@vax135.att.com

mh2f+@andrew.cmu.edu (Mark Hahn) (03/09/91)

behzad@vax135.att.com said that polishing and inspecting fiber
was a big pain.  of course, he was referring to glass, which
would be prefered for low loss in a long haul connection.
plastic fiber, on the other hand, can be 'finished' with a hot-knife 
or with some sort of epoxy.  Radio Schlock once carried a 5M
length of plasic with a LED and photodiode (or transistor, maybe).
Note also that a recent EDN has a (terribly vapid) article touching
on the available transcivers.  If anyone finds something good,
please post it, since lots of us would like to play with fiber.

regards, mark

grayt@Software.Mitel.COM (Tom Gray) (03/14/91)

In article <782@spim.mips.COM> reddy@mips.com (T.S. Reddy) writes:
>
>    Has anyone tried to make their own fiber optic interconnects? From
>what I know, the optical input/output receptacles on audio/video products
>for most brands are similar, so I would think that standardized parts for
>fashioning such a beastie would be available. I am not familiar with
>the techniques involved in cutting/splicing fiber optic cables, but
>if anyone could enlighten me, I would be grateful.
>

At the rates used in audio system, plastic cable is suitable. The splicing of
plastic cable is trivial. The cable can be cut with a pair of
scissors and butted into a plastic connector. The cable itself
is very large with 1 and 2mm diameters being typical.

The electronics for driving optical interconnects should cost
less than 20 dollars for both ends.

Amp, HP and Motorola all make a suitable selection of devices
for use in these circuits. AMP is the cheapest with HP probably
having a wider selection. An AMP transceiver for the 1megabit range
should cost about 5 dollars.

Video is another matter altogeher, the transceiver cost here would
be about $500 for both ends. You would have to by preterminated
cable since the equipmrent for splicing glass cable can cost
a few thousand dollars. 20 meters of terminated glass cable
should cost around $100.