Optical Power Budget — Where a Link Lives or Dies

Why the Numbers Do Not Hold Up in the Field

The design document says 24 dB of link loss against a 28 dB equipment budget. Four decibels of headroom, which looks fine. Then you measure on site and get 27.5 dB. The link is technically alive, but one cold snap or one additional splice will kill it.

This happens not because the arithmetic was wrong, but because something was left out of the arithmetic.

First, Separate dB from dBm

This is where most confusion starts.

• dBm is an absolute power level. "The optical power at this point is -20 dBm."
• dB is a relative difference. "This section cost us 3 dB."

dBm is a position; dB is a movement. In temperature terms, dBm is "it is 20 degrees" and dB is "it dropped 5 degrees."

Which gives you the working equation:

Received power (dBm) = Transmit power (dBm) - Total loss (dB)

Transmit at +3 dBm through 25 dB of loss and you arrive at -22 dBm.

What 3 dB Actually Means

Decibels are logarithmic, which makes them hard to feel. These three are enough:

• 3 dB loss = half the power
• 10 dB loss = one tenth
• 20 dB loss = one hundredth

That puts the 15 dB split loss of a 1:32 splitter in perspective: roughly 3 percent of the original light survives. Communication still works only because the receiver is that sensitive.

Where the Loss Comes From

Total link loss is the sum of these terms.

1. Fiber attenuation

Proportional to distance, and wavelength dependent.

• 1310 nm, about 0.35 dB/km
• 1550 nm, about 0.22 dB/km

Ten kilometers at 1310 nm is 3.5 dB. Since FTTx access spans are usually a few kilometers at most, this term is smaller than people expect.

2. Splitter loss

Covered in the previous lesson, and typically more than half of the total budget. Around 15 to 17 dB for 1:32.

3. Fusion splice loss

A well-made splice runs 0.05 to 0.1 dB. Individually negligible; across twenty splices, not negligible at all.

4. Connector loss

0.3 to 0.5 dB when healthy. This is the term that degrades most readily. A contaminated connector will happily read 1 dB or 2 dB.

5. Bend loss

Light escaping where the fiber has been bent too sharply, most often where slack was forced into an enclosure.

A Worked Example

A 1:32 split, 3 km from office to subscriber, 8 splices, 4 connectors.

Against a 28 dB equipment budget, that leaves 7.15 dB. This is the number that ends up in the design document.

What the Design Document Leaves Out

In reality you also have to account for:

• Ageing — cable and joints degrade over time. Reserve 1 to 3 dB.
• Future splices — every repair adds a splice. Budget for several over the life of the plant.
• Temperature — characteristics shift slightly with the seasons.
• Measurement uncertainty — half a decibel of spread between instruments and operators is routine.

The working rule is to keep at least 3 dB of margin. The 7.15 dB above is healthy. Had it come out at 1 dB, the design would pass on paper and fail on site.

When Measurement Disagrees with Calculation

If the measured value is 3 dB or more worse than calculated, the cause is usually one of a short list.

1. Dirty connector — the most common by a wide margin. Clean it and measure again first.
2. Incompletely mated connector — seated short.
3. Sharp bend — check the slack management inside enclosures.
4. Bad splice — needs an OTDR to localize.

The order matters. Start with the most common and the cheapest to check. Spending one minute cleaning a connector beats opening an enclosure on a splice hunch.

Summary

• dBm is absolute power; dB is loss.
• 3 dB is half the power; 10 dB is a tenth.
• More than half the budget is usually splitter loss.
• Leave ageing, future splices, and measurement spread out of the math and the design fails in the field. Hold at least 3 dB of margin.

The next lesson moves from calculation to localization: finding out where the problem actually is, with an OTDR.