Full optical power budget calculation for PON/FTTH networks. Enter your OLT transmit power, ONT sensitivity, fiber lengths, splitter configuration, and connectors for an instant pass/fail result.
Unbalanced splitters divide power unevenly between legs (e.g. a "10/90" splitter) rather than an even ratio — pick the specific leg feeding this ONT, not the other leg. Common in cascaded/daisy-chain PON architectures to extend reach past one subscriber before the rest of the power continues downstream.
In a Passive Optical Network (PON), the power budget is the total optical loss allowance between the OLT (Optical Line Terminal) at the central office and the ONT/ONU (Optical Network Terminal) at the subscriber premises.
The budget is defined as:
Budget = OLT Tx power − ONT Rx sensitivity
All losses from fiber, splitters, connectors, and splices must fit within the budget minus the safety margin. If total loss exceeds budget − margin, the link fails.
| Split Ratio | Typical Loss | Max Subscribers |
|---|---|---|
| 1:2 | 3.5 dB | 2 |
| 1:4 | 7.0 dB | 4 |
| 1:8 | 10.5 dB | 8 |
| 1:16 | 13.5 dB | 16 |
| 1:32 | 16.5 dB | 32 |
| 1:64 | 20.0 dB | 64 |
| 1:128 (2-stage) | 27.5 dB | 128 |
An unbalanced splitter divides power unevenly between its two legs instead of an even split — used to reach one subscriber off the low-loss leg while most of the power continues downstream on the high-loss leg for further splitting. Values below are calculated (−10×log₁₀ of the power fraction) plus ~0.5 dB typical excess/insertion loss, the same real-world padding convention as the balanced table above.
| Split ratio | Low leg | High leg |
|---|---|---|
| 5/95 | 13.5 dB | 0.7 dB |
| 10/90 | 10.5 dB | 1.0 dB |
| 15/85 | 8.7 dB | 1.2 dB |
| 20/80 | 7.5 dB | 1.5 dB |
| 30/70 | 5.7 dB | 2.0 dB |
| 40/60 | 4.5 dB | 2.7 dB |
Typical/reference figures — real unbalanced splitters vary by manufacturer. Use your splitter's datasheet value if you have it.
The three major PON standards differ in downstream/upstream speed and power budget class:
All three standards operate on single-mode fiber at 1310 nm upstream and 1490 nm or 1577 nm downstream, using WDM to carry both directions on the same fiber.
Take a GPON deployment on Class B+ optics (28 dB power budget): a 1:32 splitter (16.5 dB per the table above), 20 km of single-mode fiber at a typical 0.35 dB/km, and 4 connectors at 0.5 dB each.
Fiber loss: 20 km × 0.35 dB/km = 7.0 dB. Connector loss: 4 × 0.5 dB = 2.0 dB. Total loss: 16.5 + 7.0 + 2.0 = 25.5 dB.
Against a 28 dB budget, that leaves 2.5 dB of margin — under the 3 dB safety margin generally recommended for GPON designs. This design would come back marginal, not a clean pass: it'll likely work on day one, but leaves little room for splice degradation, connector wear, or fiber repairs over the life of the plant. The fix is usually one of: drop to a 1:16 split (13.5 dB, buying back 3 dB), shorten the run, or move to a higher power-budget class (XGS-PON N1 at 29 dB, or a C+ optics class at 32 dB).
Can I mix a 1:8 splitter at the hub with a 1:4 at each cabinet to get 1:32 total?
Yes — this is the standard two-stage splitting architecture (a "1:128 (2-stage)" row is included in the table above for the equivalent case). Two-stage splitting is common because it lets you deploy the first-stage splitter early and add second-stage splitters closer to subscribers as they're actually turned up, rather than pre-splitting to the final ratio at the hub before demand exists.
Why is XGS-PON's power budget class (N1, 29 dB) so close to GPON's (B+, 28 dB) despite 4x the speed?
Power budget and bit rate are largely independent design choices — budget class is set by transceiver optics (launch power and receiver sensitivity), not by how fast the data is modulated. XGS-PON gets its extra speed from better modulation and receiver technology, not from a fundamentally different optical budget.
Does this calculator account for WDM/wavelength-specific loss differences?
Fiber attenuation is slightly wavelength-dependent (typically lower at 1550nm than 1310nm), but PON standards specify power budget as a single class figure already validated across their working wavelength plan, so this calculator uses the standard's published budget class directly rather than re-deriving it per wavelength.