Conference 7.286::fddi

    The DECconnect Fiber Optic Network Design Guide on page 5-27 describes
    the maximum lengths and losses for fiber cable runs:
    
    62.5 fiber, 1300nm wavelength, 2000 meters max distance, -11.0dB max loss
    
    The FDDI standard (to which Digital adheres) specifically states that
    the maximum distance between two stations must be limited to 2Km, using
    62.5 fiber.  This distance is ultimately going to be shorter onsite,
    because of "next station in bypass mode" difficulties, and the
    likelihood of passive hubs further reduces your overall distance. 
    There are many nice Video-tape courses and DVN tapes showing the FDDI
    configuration details nicely.
    
    The new DEBAM-RF (LB200 with 1300nm option) is supported over dual
    window 62.5/125um fiber optic cable at distances of up to 10km.  There
    is talk of some optional boards for our un-annouced products that might
    support other types of fiber too, but it's all rumor, and for selling
    NOW, your best bet is the DEBAM-RF.
    
    As far as "different optics, and single mode fiber", what I thought
    that I heard at NetU was that "...single mode laser sources would be 
    far too expensive for a "station" application, and would probably only
    be offered for inter-concentrator long haul links..." and the distances
    thrown about were huge (35Km between single mode stations), but I don't
    beleive that Digital will build one soon.  
    
    For now, if you need to go a few Km, I would plan little enclosures
    along the way, or stop in buildings at 2Km intervals, and place a
    repeater or a Star, or "future FDDI ring product" in each site.  The
    Total FDDI ring can be quite long (100Km I beleive is the pair limit) 
    
    Contact your local NWSS Sales Support Specialist, who can help you to
    configure an F.D.D.I. compatible fiber optic cable plant, or even sell
    a design for one to your customer, (complete with blueprints, and
    optical loss budgets etc.)
    
    JR

  I've heard that the 2Km limit for multi-mode fiber is a dispersion
limitation.

  The current PID indicates that in the near future there may be support
for up to 5Km on multi-mode fiber.

  How do we plan to support such distances if the limitation is dispersion
related?

There is no such thing as 5 km on multi-mode fiber.  Where did you see that?

On the other hand, single mode would indeed support this.  Rather than
speculate about when or whether Digital would provide such products, you
should discuss this with product management.

	paul

I can't really tell whether .2 was referring to Ethernet (10 Mbit/s) or
FDDI (100 Mbit/s).  My answer is for FDDI.  .-1 talks about Ethernet,
where you certainly have a lot of margin at 1300 nm.  For FDDI you do NOT
have that sort of margin and 2 km really is the limit.

	paul

    re .5
    
    Paul, correct me if I'm wrong but I thought the whole discussion
    was just meaningless : the FDDI spec reads 2km max, period. Any product
    that would deviate from this would be non-standard ie would probably
    negate the interoperability of the whole thing.
    
    As you said Ethernet (DEBAM) is a different issue.

More precisely, the FDDI PMD (i.e., multi-mode fiber) standard says 2 km.
However, the single mode (SMF-PMD) standard allows for more.  So you will
have a conforming system even at > 2 km station distances, but only with
single mode fiber, not with multi mode.

	paul

RE .2
    
    The 2Km limit is not merely a "dispersion limitation", but a limitation
    from the cumulative affects of Fiber length loss, launch loss,
    connector loss, splice loss, future maintenance allowance, and a
    dispersion penalty.
    
ETHERNET ON FIBER:
        
    For a typical fiber optic cable at 820nm (digital's current products)
    
    Average of two ST connectors (.7 dB each)		-1.4dB
    Maintenance Loss (two splices at .3dB each)		-0.6dB
    Splice Loss (average of one at .3 dB each)		-0.3dB
    Coupling Loss (Launch Loss, included in HW budget)	-0.0dB
    Dispersion Penalty (Usually specified as Fiber Bandwidth.Km)
    Length Loss (typically 4.3dB/Km) for 2Km		-8.6dB
    
    Total System Loss at 850nm				-10.6dB
    
    That Number is very close to the Margin limit for Digital's (-11.0dB)
    devices.  Notice that Dispersion Penalty is not used as part of the
    calculation.  The typical method of determining the dispersion (both
    chromatic and Modal) of the fiber is by cross referencing both factors
    into a bandwidth specification for the fiber.   The trait of different
    wavelenghs of light to be refracted differently in a medium is called
    chromatic dispersion (and can be easily verified by a prism), the
    actual "straightness" of the path the light follows (either directly
    through the center of the fiber, or bouncing off the internal walls) is
    called modal dispersion.  Both dispersion types cause a small part of
    the light appear at the far end of the fiber both before and after the
    majority of the signal, causing pulse broadening, and eventually data
    loss. 
    
    FDDI (Digital's FUTURE PRODUCTS)
    
    Here loss is vital to the budget, but Dispersion is now a major player. 
    Again, based on Bandwidth of the fiber, where the dispersion is
    un-calculable to an average person.  The Bandwidth of a Fiber optic
    cable is listed in MHz.Km (which is MegaHertz times Kilometers).  If
    you have a 500MHz.Km cable, than you can send 500Mhz data for a
    Kilometer, or 1GHz data for 500meters, or more importantly, 250MHz data
    for 2Km.  This gives us plenty of headroom for FDDI data at the
    supported distances.  The calculation for longer distances shows:
    Bandwidth drops off to 166Mhz for 3Km, which is really tight, and 4Km
    at only 125MHz which leaves no spare B/W at all.
    
    If your ring has a mechanical bypass in some of the stations (let's say a
    third party PC with an optical bypass relay) that is in "bypass" mode,
    it would insert an extra 4 or 5 dB of loss in the system between the
    two nodes.  This loss could push you over your -11dB budget, if you
    have any patch panels or faceplates.
        
    Digital has specified minimum spectral bandwidths for both windows, in
    order to negate the requirement of calculating chromatic or modal
    dispersion in the field.  AS long as you use fiber optic cable that
    meets the 160MHz.Km at 850nm, and 500MHz.Km at 1300nm, your loss
    calculations needn't worry about dispersion, below 2Km.    
   
    Your comment is correct about supporting longer distances on multimode,
    at 1300nm but only at Ethernet Speeds where these distances won't
    affect the loss budget, BUT NOT AT FDDI SPEEDS.  The new DEBAM-RF in
    fact uses this characteristic of the second window to support up to
    10Km of multimode fiber.   The FDDI products cannot exceed 2Km on
    Multimode without a significant B/W increase over the industry standard
    fiber.  The B/W of single mode fiber is immense, and therefore allows
    much greater distances.  
    
    Sorry for leaving out the 10M vs. 100M from my previous reply....
    
    JR
    
    
    
    
    
    

    John:
    
    Where do you learn all this?  I'd love to have that kind of info in my
    head.  Do I have to pick it up on the streets, of is there DEC training
    with this kind of detail?
    
    Patrick