Gedeon's FaithFiber install and repair

Reference

Colors, loss, and wavelengths.

Field numbers to keep next to the meter. They are typical, not a cable spec and not a warranty. If the datasheet disagrees, the datasheet wins.

Fiber

OM3 Aqua · 50/125 µm

10GBASE-SR is comfortable to about 300 m. Past that, check the optic, not just the loss.

  • 850 nm · 3.00 dB/km
  • 1300 nm · 1.00 dB/km

OM4 Aqua · 50/125 µm

Same loss ballpark as OM3, more bandwidth. 10GBASE-SR is commonly specified to 400 m.

  • 850 nm · 3.00 dB/km
  • 1300 nm · 1.00 dB/km

OM5 Lime · 50/125 µm

Wideband multimode for short-wave WDM. Treat 850 nm loss like OM4 unless the cable sheet says otherwise.

  • 850 nm · 3.00 dB/km
  • 1300 nm · 1.00 dB/km

OS2 Yellow · 9/125 µm

Premises jumpers and outside plant. PON drops are almost always this, and almost always green APC.

  • 1270 nm · 0.38 dB/km
  • 1310 nm · 0.35 dB/km
  • 1490 nm · 0.24 dB/km
  • 1550 nm · 0.22 dB/km
  • 1577 nm · 0.23 dB/km
  • 1625 nm · 0.25 dB/km

Connector color

  • Blue — UPC, single-mode.Return loss often around −50 dB or better. Mate only with UPC.
  • Green — APC, single-mode, 8°.Return loss often −60 dB or better. PON, many splitters, and a lot of outside plant. Never into blue.
  • Aqua — OM3 or OM4. Lime — OM5.Orange jackets are older multimode. Beige sometimes too.
  • Yellow jacket — single-mode.The boot color, not the jacket, tells you the polish.

Clean sequence

  1. Inspect if you have a scope. If you do not, still clean.
  2. Dry first for dust. One click, one cassette swipe, or one stick. One pass, not a scrub.
  3. Wet-dry for film or oil: a proper fiber solvent, then a dry finish. No shirt, no compressed air from a dirty nozzle.
  4. Clean the bulkhead with a stick sized for that sleeve. Then inspect again.
  5. Cap it the moment it is open. A new-from-the-bag end is still a suspect.

Loss you can expect

  • Clean mated pair, field

    0.2–0.5 dB

    What the budget defaults toward

  • Connector pair, test limit

    0.75 dB

    Common certification ceiling, not a goal

  • Fusion splice, decent

    0.05–0.15 dB

    Budget default is 0.10

  • Splice, test limit

    0.3 dB

    Do not spend this on every joint

  • PLC 1×32, typical

    about 17 dB

    The splitter, not the fiber, dominates a PON drop

  • OS2 at 1310 nm

    0.35 dB/km

    Premises cable is often specified a bit higher

  • OS2 at 1550 nm

    0.22 dB/km

    Bends show up here first

  • OM3/OM4 at 850 nm

    3.0 dB/km

    TIA maximums are looser, near 3.5

Wavelengths that get mixed up

  • 850 nm

    Multimode SR optics. Do not launch this into yellow single-mode and expect a link.

  • 1300 nm

    Older multimode. Still on the meter. Easy to leave the meter here by mistake.

  • 1270 / 1310 nm

    Single-mode uplinks, and the upstream side of common PON.

  • 1490 nm

    GPON downstream. Budget it. Do not assume it matches 1550.

  • 1550 nm

    LR and longer single-mode. More sensitive to macrobends than 1310.

  • 1577 nm

    XGS-PON downstream. Same glass, slightly different loss than 1490.

  • 1625 nm

    Often an OTDR or monitoring wavelength, out of band for the traffic.

dBm to power

Meters speak dBm. Datasheets sometimes speak milliwatts. Same quantity. 0 dBm is 1 mW. −20 dBm is 10 µW. −30 dBm is 1 µW.

Linear power

10.00 µW

mW = 10^(dBm / 10)

Bend and safety

Standard G.652 single-mode hates a tight loop, especially at 1550 nm. G.657.A1 and A2 patch cords forgive more — roughly a 10 mm or 7.5 mm radius — and still fail if you crease them. Follow the radius printed on the cable. Do not look into a port. Infrared will not look bright, and it will still hurt you. Use the meter.