Build a link budget and decide go or no-go

applied · 40 min · Objective 1.5

Task

Calculate whether a proposed fibre run will actually work, using the loss budget method: transmit power, minus every source of loss along the path, against the receiver's sensitivity. This is the calculation that decides whether a link is a purchase order or a support ticket.

Steps

  1. Take the link: 1000BASE-LX single-mode, 8 km, with two patch panels at each end and one fusion splice in the middle.
  2. Write down the budget inputs. Typical 1000BASE-LX: transmit power minus 9.5 dBm minimum, receiver sensitivity minus 20 dBm. That gives you 10.5 dB to spend.
  3. Count the losses. Single-mode fibre attenuates about 0.4 dB/km at 1310 nm. Each mated connector pair costs about 0.75 dB, and you have four. A fusion splice costs about 0.1 dB.
  4. Total the losses, subtract from the budget, and state the margin. Anything under about 3 dB of margin is a link that works on day one and fails after somebody cleans a connector badly.
  5. Re-run it for 20 km and state what you would change. Then re-run it for a 50 m link using the same long-reach optic and identify the opposite problem.

Verify

python3 -c "
tx, rx = -9.5, -20.0
budget = tx - rx
for km, conns, splices, label in ((8,4,1,'8 km as specified'), (20,4,1,'20 km'), (0.05,4,0,'50 m')):
    loss = km*0.4 + conns*0.75 + splices*0.1
    margin = budget - loss
    verdict = 'OK' if margin >= 3 else 'MARGINAL or FAIL'
    print(label, 'loss', round(loss,2), 'dB   margin', round(margin,2), 'dB  ', verdict)
"

The 8 km link should come out around 3.4 dB of margin — tight but workable. The 20 km link fails, and the answer is an ER optic, not more patching. The 50 m link passes the loss test and fails the opposite one.

Notes

That last case is the one field engineers actually hit: a long-reach optic on a short link overdrives the receiver. The symptom is errors rather than a dead link, which sends people hunting for a cable fault that does not exist. The fix is a fixed attenuator, and lesson 40's signal-strength readings are how you spot it — received power above the maximum rather than below the minimum.

Real installers use a certifier's measured loss rather than a table of typicals. The table is how you decide what to buy; the certifier is how you prove what you installed.