[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference 7.286::fddi

Title:FDDI - The Next Generation
Moderator:NETCAD::STEFANI
Created:Thu Apr 27 1989
Last Modified:Thu Jun 05 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:2259
Total number of notes:8590

924.0. "FDDI on 50/125 fiber?" by KXOVAX::LOVE () Mon Apr 12 1993 09:14

	I have a customer that has a backbone fiber Ethernet by FiberComm
and they use 50/125 cable.  Could this same cable be used to build a
FDDI backbone?

Norm
T.RTitleUserPersonal
Name
DateLines
924.1KONING::KONINGPaul Koning, A-13683Mon Apr 12 1993 10:3810
Quite possibly, but you need to know more about the cable than just its
diameter.  In particular, you need to know the specifications (attenuation
and "modal bandwidth") at 1300 nm wavelength.  If the cable is "dual window"
cable designed to support that wavelength, you're probably in good shape.
You'd need to check the various parameters to make sure the links will work,
since you get some additional losses when using 50/125 fiber.  But that fiber,
IF it's the FDDI kind, usually has quite low losses, so it's usually not hard
to stay within the FDDI link specifications.

	paul
924.2More informationLEVERS::KIRKtis a gift to be simpleMon Apr 12 1993 10:5817
    Back in 1989, Chris Baldwin published a memo regarding supportable
    distances on alternate optical fibers. This memo has been used to
    generate the supported distance tables that can be found in the
    Networks Buyers Guide. Since Chris is no longer with the company, I'll
    take the risk of posting the memo without his permission.
    
    	As Paul indicated in his response, you must know the attenuation
    and modal bandwidth of the 50/125 micron fiber. If the fiber vendor is
    able to provide information regarding the chromatic dispersion values,
    then you may be able to achieve the maximum distance. 
    
    	You can contact me if you need more information.
    
    Dick Kirk
    Networks Engineering
    LEVERS::KIRK
    DTN 226-7048
924.3Supported Distances MemoLEVERS::KIRKtis a gift to be simpleMon Apr 12 1993 10:59290
From:	GIGA::BALDWIN "WYSIWYG....no refunds  28-Jul-1989 1127" 28-JUL-1989 11:30:32.20
To:	@DIST.
CC:	BALDWIN
Subj:	Supportable Transmission Distances with Alternate Fiber Sizes for FDDI




+---------------------------+
!   !   !   !   !   !   !   !           INTEROFFICE MEMORANDUM
! d ! i ! g ! i ! t ! a ! l !
!   !   !   !   !   !   !   !
+---------------------------+


TO:Distribution:                        DATE:7/28/89
                                        FROM:Christopher Baldwin  
                                        DEPT:CSE Technology: Fiber Optics  
                                        EXT:227-3637
                                        LOC:TAY2-2/N15
                                        NODE:GIGA::BALDWIN  


   SUBJECT:Supportable  Transmission  Distances  for   FDDI   Physical
     Connections Using Alternate Fiber Sizes.


   EXECUTIVE SUMMARY:
   ==================
     The supported transmission distances and loss budgets  for  DEC's
     FDDI  products  transmitting  through  any of the alternate fiber
     types (50/125, 100/140 or 85/125)  or  un-characterized  62.5/125
     cableplant is detailed in this memo.


     1.  50/125 um cableplants will have only a 6 dB  loss  budget  to
         operate  with.   Those complying with the chromatic and modal
         bandwidth characteristics in the PMD Chapter 7 can support  2
         km transmission distances with this loss budget.

     2.  ANY 50 um cableplant of 500 meters or less with 6 dB or  less
         of  loss  can  be  supported  provided that the fiber's modal
         bandwidth at 1300 nm was controlled by the manufacturer.

     3.  ANY 100/140 or 85/125 um cableplant of  500  meters  or  less
         with 11 dB or less of loss can be supported provided that the
         fiber's modal bandwidth at 1300 nm was controlled.

     4.  Up to 1.6 km lengths of 50,62.5,85/125 or 100/140 fiber  that
         was   not  characterized  for  chromatic  dispersion  can  be
         supported provided that the installed cableplant has a  modal
         bandwidth of 200 MHz at 1300 nm over its full length.

     5.  A PRIORI TRANSMISSION DISTANCE GUARANTEES CANNOT BE GIVEN FOR
         ANY CABLE TYPE THAT WAS NOT CHARACTERIZED FOR MODAL BANDWIDTH
         AT 1300 NM OPERATION BY THE MANUFACTURER.

     The experiments performed to derive these results were  conducted
     on  both  the  Hewlett-Packard  and Sumitomo devices that will be
     used in the DEC product set.  The results apply equally to either
     vendor's devices.

                                                                Page 2


   BACKGROUND
   ==========
     The FDDI Physical Channel is optimized for use with  62.5/125  um
     GI  fiber.   DEC  can  support the alternate fiber sizes (50/125,
     100/140  &  85/125)  but  with  decreased  loss  budgets   and/or
     transmission  distances  due to the power penalties and decreased
     bandwidth associated with use of the alternate fiber types.

     The requirements in PMD Chapter 7 describe  the  requirements  on
     the  cableplant  (irrespective  of  fiber type) to support a 2 km
     transmission span.  However, installed cableplants may  not  have
     been completely characterized for use with the 1300 nm technology
     employed in  the  FDDI  Physical  Layer  and  it  is  accordingly
     difficult  to unequivocally state whether a given cableplant will
     function  with  an  FDDI  implementation.   This  memo   combines
     empirical  data  along  with  published  analysis  of  cableplant
     characteristics to arrive at rules for the  use  of  non-standard
     installed cableplants.


   Analysis:
   =========
     An alternate fiber cable must be guaranteed to adequately control
     end   to   end   loss   and  bandwidth  when  used  with  a  FDDI
     implementation.   Bandwidth  is  composed  of  three  components:
     electrical,  modal  and  chromatic  - the latter two of which may
     depend on the fiber type being used.  The loss  a  of  cableplant
     consists  of  the  bulk loss of the fiber, its connectors/splices
     and additional penalties that may be  incurred  by  the  mismatch
     between   the   cable's  core  size  and  the  station's  optical
     components which are optimized for use with 62.5/125 cable.  Loss
     and   bandwidth   for  the  alternate  fiber  sizes  are  treated
     separately below.


     Loss Analysis:
     -------------
       At the transmitter end, the use of alternate cableplants  can
       cause  more  or  less power to be launched due to the size of
       the alternate cable's core relative to 62.5 um cable.  At the
       receiver  end, either the focusing optics or the PiN detector
       can vignette the image of a larger than 62.5 um core  causing
       an effective receiver penalty.

       1.  50/125 um Cableplants:

           13   Hewlett-Packard   Transceivers   were   tested   for
           transmitter   optical  launch  power  penalty.   Sumitomo
           transmitters were also tested  and  were  found  to  have
           slightly less degradation; according the results from the
           HP devices will be used.
           Both NA = 0.20 and NA  =  0.22  fiber  was  tested.   The
           transmitter  was found to suffer a 5.0 dB penalty maximum
           when used in conjunction with  NA  =  0.20  fiber  (worst
           case).   No receiver vignetting will occur (50 um core is

                                                                Page 3


           smaller than the 62.5 design fiber.)

           LOSS BUDGET = 11 dB - 5 dB penalty = 6.0 dB  with  50/125
           cableplants

       2.  100/140 Cableplants:

           3 HP Transceivers and Sumitomo transmitters  were  tested
           for transmitter optical launch power increase.  The NA of
           the fiber was 0.29.  A 1.5 dB increase was found.

           The receiver was  found  to  have  a  0.5  dB  vignetting
           penalty  when  used  with 100 um core fiber.  The penalty
           was confirmed for worst case mode scrambled conditions  (
           mode  scramble  exit  conditions ensure that the light is
           maximally distributed throughout the core of the fiber so
           that  maximum  vignetting  will occur.) HP explained that
           the source of  the  vignetting  occurs  at  the  detector
           because  the NA and pupil aperture of the focusing optics
           is more that sufficient for  a  100  um  core  1:1  image
           relay.   The  PiN  detector  is  only  90 um in diameter.
           While only a 0.5 dB vignetting penalty was found, 1.5  dB
           will be reserved for this penalty to account for possible
           misalignment between optical relay and PiN detector which
           would  exacerbate the penalty.  The Sumitomo devices were
           found to have a 0.3 dB Rx vignetting penalty;  since  the
           difference  between the HP and Sumitomo results are minor
           the 1.5 dB penalty will be used.

           LOSS BUDGET = 11 dB + 1.5 dB Tx gain - 1.5 dB Rx  penalty
           = 11 dB

           NB:  The  use  of  100/140  um  patch  cords  to  connect
           stations is NOT RECOMMENDED.  There is a possibility that
           the Tx could saturate a Rx when using a short  length  of
           100/140  fiber.   This is currently not a serious concern
           as FDDI Connectorized 100/140 cables are a special  order
           item.

       3.  85/125 um Cableplants:

           The same  analysis  applied  to  100/140  um  cableplants
           applies to 85/125 cableplants.

           LOSS BUDGET = 11.0 dB.



     Bandwidth Analysis:
     ==================
       A fundamental requirement that an installed  cableplant  MUST
       satisfy is that the manufacturer controlled the 1300 nm Modal
       Bandwidth (MBW) of the fiber.  If the  manufacturer  did  not
       measure the 1300 nm performance of his product then it is not
       possible to predict its performance now.   The  bandwidth  of

                                                                Page 4


       the  specific  cableplant  in question need not be known now,
       but it is required that the manufacturer ascertained that his
       process met a minimum bandwidth at 1300 nm.  Corning, ATT and
       Spectran currently all measure 1300 nm MBW of  all  of  their
       fiber  types  and  have  been doing so for some (unspecified)
       time.  They all have a minimum requirement of 100 MHz*km.

       The bandwidth of a FDDI cable is determined by  the  MBW  and
       the  chromatic characteristics of the cable; the requirements
       are treated separately.

          If a cable can be shown to have a MBW of 200 MHz over  the
       length  of  the  fiber  ( MBW distance product divided by the
       length of the cable), then it has sufficient MBW for  a  FDDI
       link.  Cable manufacturers state that customers almost always
       specify (thus know) the MBW of their cable.   Accordingly  it
       is  likely  that  an  installed  cableplant will have a paper
       trail that can divulge its MBW.  In a case were the MBW of  a
       particular  piece  of cable is not known it is still possible
       to define a maximum supportable length:  the minimum MBW  for
       all  fiber  types  from  all  manufacturers  is  100  MHz*km.
       Clearly then a 0.5 km link will have sufficient MBW over  its
       length.   Alternately  the  installed cable could be measured
       for modal bandwidth in the field.

          The  optical  system  also  requires  that  the  chromatic
       dispersion characteristics of the fiber fit within the curves
       of figure 7-1 of the PMD.  The chromatic  characteristics  of
       MM  GI  fibers  are  very  difficult  to  measure  even  in a
       controlled laboratory environment and impossible to do in the
       field.  In two separate papers (1), ATT and Corning data show
       data that  demonstrates  that  fiber  NA  and  its  chromatic
       characteristics  have  some  correlation.   In the paper they
       measure the chromatic characteristics of a  large  number  of
       spools   of  fiber  (greater  than  30  each)  and  the  data
       demonstrates that the maximum slope of the  dispersion  curve
       never  exceeds  0.11 ps/nm*nm*km and that the maximum minimum
       dispersion wavelength does  not  exceed  1365.   However  the
       curves in the PMD do not allow this particular combination of
       worst case conditions.  To guarantee that such a  worst  case
       condition  will work, the maximum length of a cableplant that
       has not been characterized for chromatic dispersion needs  to
       be derated from the 2 km maximum.
          Since the slope of the dispersion curve and  cable  length
       have  the  same  mathematical  function  and weighting in the
       model for chromatic bandwidth,  the  maximum  length  of  the
       cableplant is derated by the same percentage that the maximum
       slope of the dispersion curve exceeds  the  PMD  requirements
       (0.093  ps/nm*nm*km);  accordingly the maximum length for the
       alternate fiber sizes that cannot demonstrate  compliance  to
       PMD figure 7-1 is 18% less than the 2 km limit or 1.6 km.



       (1)  "The  Importance  and  Application  of   Dispersion   of

                                                                Page 5


       Multimode  Fiber  in  LAN's  and  Its  Relation  to NA." M.J.
       Hackert CGW
       "Determining the Zero Dispersion  Wavelength  and  Dispersion
       Slope of Fibers Used for the Fiber Distributed Data Interface
       (FDDI)" J.J.  Refi & H.  Shang.


       DISTRIBUTION:
       NM%LEVERS::NEWMAN
       NM%LEVERS::THOMPSON
       NM%LEVERS::WARTER
       NM%LEVERS::GINZBURG
       NM%LEVERS::RIEGER
       NM%LEVERS::ROLLA
       NM%DELNI::BOYCE
       NM%DELNI::DUBE
       NM%DELNI::COOLIDGE
       NM%DELNI::CARBONELL
       NM%DELNI::ARIVERA
       NM%KELVIN::KOLTE
       NM%KELVIN::PAZARIS
       NM%POLAR::KAHLIL
       NM%POLAR::THOMSON
       NM%POLAR::NEWMANN
       NM%LEVERS::KIRK
       NM%LEVERS::ERWIN
       NM%ERLANG::SCHOFIELD
       NM%DONKEY::KNUDSON
       NM%DELNI::KRUEGER
       NM%DELNI::S_LEE
       NM%LEVERS::ERTEL
       NM%LEVERS::TIFFANY
       NM%LEVERS::PLOUFF
       NM%LEVERS::KOCHEM
       NM%LEVERS::E_ROGERS
       NM%LEVERS::P_CHOW
       NM%DELNI::SWAN
       NM%DELNI::PARIKH
       NM%DELNI::KUNZEL
       NM%JRDV01::KAWAMURA
       NM%BAGELS::LEVY
       NM%NACMIS::IAMARTINO
       NM%DELNI::PIEPER
       NM%Levers::M_FORD
       NM%Levers::S_MACARTHUR
       NM%Delni::B_GEARING
       NM%LEVERS::SHUDA
    
924.4ThanksMSDOA::LOVEDo it with DNAThu Apr 15 1993 15:513
    Thanks for the information.
    
    Norm
924.5STRWRS::KOCH_PIt never hurts to ask...Thu Aug 03 1995 00:2516
    
    Well, I have a customer site which seems to have 100/140 fiber.
    However, this was used for interconnecting fiber LANbridge 100s. Now
    they want to use it for FDDI. According to the spec sheet from the
    testing company from May 1987 it was tested at 850nm. The question is,
    is it possible to test this fiber for FDDI suitability with test
    devices? If so, what would those test devices be?
    
    The length of this cable is only about 200 FEET between 2 buildings.
    Also, if this cable does test OK, what kind of patch cords would I use
    from the patch panel to the DECconcentrator 900MX the customer is
    considering buying? A BN24D-xx or something else?
    
    I'm going to try and ask them for the attenuation and modal specs, but
    given this cable is over 8 years old, I don't think they have it. So,
    given this, what would be a good approach?