Shop OBEX P1 Docs P2 Docs Learn Events
Communication via fiber optics? — Parallax Forums

Communication via fiber optics?

DavidNymanDavidNyman Posts: 52
edited 2010-11-19 08:30 in Accessories
Has anyone had any luck getting two BS2 stamps to communicate via a fiber optic cable?
I've read about the IR Buddy that Parallax used to sell and that seems adaptable to my needs, they just don't make them anymore. :-(
I'd like to start a new ROV project and control it over a fiber link.
Any suggestions would really be appreciated.
Thanks.

Comments

  • Peter KG6LSEPeter KG6LSE Posts: 1,383
    edited 2010-10-17 10:43
    I Have have had luck with one way data 115Kbaud Hyperterm ,( I had no need to do 2way DATA .. and got a BS2 up to 4800 Baud .

    http://forums.parallax.com/showthread.php?t=120132

    this is for voice but the curcuit is modifyable to use DATA too with a few tweeks ..

    The Key point of my designed system was the fibre to LED( or photo diode ) interface using a ST coupler .....

    as you see in my Circuit the Laser is OOK'ed ( on off keyed ) with a MOSFET.
    there are a few reasons I use the FET . one is input Resistance is uber high so its allmost a non issue ( NOTE not including the XsubC) .

    Ill bet a IGBT with a FET as its input characteristics would work too..

    now If you use a LASER you have 2 options .

    1) use a module and OOK the whole module >> You may kill the diode driver . the onboard driver wants to AT ALL COSTS keep that diode fed with tight current and voltage . and if its a switching driver it may glitch IF you key it at just the right FREQ . linear regs I have found do not have this issue . My Green 525nM LASER is OOK'ed right at the input to the home made driver .

    2)now if you can put the active switching element FET or BJT or optoisolator right at the Neg side of the diode .. You can get a Much more clean modulation then the OOK of the whole module . mind most of the LASERS you may be useing are 5mW and those small ( on the whole) LASERS are VERY picky with current and voltage . unlike a LED one look at a LASER diode can blow it in picoseconds .

    I used a Brute force reg system for the PhotoPhones . I use a LM7805 to keep the Bias on the MosFet tight regardless of the 9V powering it . AND to create the Ref to feed the LASER Diode . the Cheap($5) LASERS I used were from wally world . they used one resistor ( like a LED) . VERY RISKY . one over healthy batt Cell and there goes the diode .

    They relied on the batt's DC resistance to be a active part of the REG system .

    so In my circuit I use a Ammeter to read the diode in its virgin state and made sure that under no circumstances can it go over .

    and to make the diode happy on the voltage side I dropped voltage with some 1N4007s to chop that 5V off the LM7805 to I think was 3.8 or so volts .
    ( the pointer used 3 buttons in series .. Silver I think.. I know thy were not Lithium .. )

    I assume cause you are going from Stamp to Stamp youre not needing gigbit data rates . I would try OOK'ing a module . 4800 Baud to a Module I think is nothing .
    I have no hard data on where there is a FREQ cutoff as its on a per LASER module basis .

    Now on the subject of LASERS . Wavelength is a must . not power . there are SO many losses useing non matched Fibre .
    808 nM is way better then 650 or so red or 535 green .
    that Hair thin Fibre Likes IR light more then visible


    eg I used a IR-LED for my Voice TX and it was VERY good at long DX ..
    Mind at this point Iam assuming your useing that hair thin fibre not the 1mm think Audio plastic stuff . plastic or glass for that matter does not care as much of the wavelength at that huge diamater .

    On the RX side is also a challenge

    . use a comparator to set a 1/0 set point then feed it in to a buffer to make the output TTL to feed the stamp.
  • plenty_otooleplenty_otoole Posts: 2
    edited 2010-10-31 19:36
    David,

    Don't know how much time or money you want to spend, but the least expensive off the shelf solution I can think of would be to sidestep using serial entirely and head straight for a pair of UTP to Fiber Optic converters. They sell surplus for about twenty bucks a piece with ST terminations.

    Canoga Perkins makes the smallest package units I've seen so far. There are modules that are smaller but they tend to be rather pricey.

    The big plus is you'd have plenty of bandwidth for the inevitable "feature creep" and more time to work on the application than the interface.
  • stamptrolstamptrol Posts: 1,731
    edited 2010-11-03 12:37
    Go off-the-shelf. See b&B Electronics for fibre-optic to serial converters.

    Save your design effort for the rest of the ROV.

    Cheers,
  • AldoAldo Posts: 9
    edited 2010-11-11 13:39
    http://www.avagotech.com/docs/AV02-1501EN

    try this, if the range is not ok, try another one wich couples to glass fiber instead of plastic fiber.

    its cheap and can communicate with a propeller on ttl logic level.
  • AldoAldo Posts: 9
    edited 2010-11-19 00:33
    Here :

    Propeller_Optical_Fiber.pdf

    Component values taken right from the datasheet.

    You just need a way to convert from 5V to 3.3V (The DataIn pin accepts 3.3V but the DataOut is a straight 5V output).
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2010-11-19 03:14
    Although I have done some single-mode fiber for long distance serial communications (40km/node) your request begs the question: Why do you think you need fiber? What distance are you talking about anyway?

    Since you are using Stamps you obviously are not talking about high-speed serial at all, so what are your requirements then? Perhaps all you need is something basic like RS485 which at low speeds should be more than adequate for ROVs.
  • AldoAldo Posts: 9
    edited 2010-11-19 08:30
    currently, plastic fibers are not much costlier than copper...
Sign In or Register to comment.