Transmitter Output 6 dB Low on a 10G Link
You are qualifying a 10GBASE-ER SFP+ before installation. The datasheet gives average launch power +4.0 dBm maximum and -1.0 dBm minimum at 1550 nm; receiver sensitivity is -16.0 dBm. You connect the SFP+ through a 2 m SC/UPC single-mode patch cord to a handheld power meter. The patch cord is known good, with 0.2 dB loss at 1550 nm. The meter reads -2.7 dBm. The meter's calibration certificate covers 850/1300/1310/1550 nm, but the display wavelength selector is set to 850 nm. The meter was calibrated 14 months ago. Room temperature is 23 °C. The tech wants to replace the SFP+ immediately because it is '6 dB low.' You have a known stable 1550 nm source at +1.0 dBm and a second power meter.
1. Calculate the expected meter reading if the SFP+ is at its maximum and minimum spec, and assess the -2.7 dBm reading. What is the first action?
Read the worked answer
At the meter, expected max = +4.0 - 0.2 = +3.8 dBm; expected min = -1.0 - 0.2 = -1.2 dBm. The -2.7 dBm reading is 6.5 dB below the max-case expectation and 1.5 dB below the min-case expectation. That looks like a fail, but the measurement is invalid because the meter is set to 850 nm while measuring 1550 nm. A wavelength-calibration mismatch changes the detector responsivity correction and can easily produce several dB of error. The first action is not to replace the SFP; it is to set the meter to 1550 nm, verify with the known +1.0 dBm source, then re-measure. Use the same patch cord and adapter, and inspect/clean both connectors.
2. List the checks in order, and state what you would accept as proof that the meter is not the problem.
Read the worked answer
Order: 1) set wavelength to 1550 nm; 2) inspect/clean meter port, patch cord, and SFP; 3) attach the known +1.0 dBm source through the same cord and meter; 4) compare reading to +1.0 dBm within the meter's specified uncertainty; 5) re-measure the SFP. If the known source reads +1.0 ±0.5 dB, the meter and cord are fit for this check. If it reads low by the same ~3 dB, the meter or cord is suspect. Also check that the meter is not in 'dB relative' mode, that the detector is not saturated, and that the adapter is not multimode. Only after a valid 1550 nm reading should you judge the SFP.
3. After correction, the SFP reads +3.6 dBm. For a 10 km link with 0.35 dB/km fiber, four 0.3 dB connectors, and two 0.1 dB splices, what is the received power and margin to sensitivity?
Read the worked answer
Fiber: 10 km × 0.35 = 3.5 dB. Connectors: 4 × 0.3 = 1.2 dB. Splices: 2 × 0.1 = 0.2 dB. Total link loss = 4.9 dB. Received power = +3.6 - 4.9 = -1.3 dBm. Margin = -1.3 - (-16.0) = 14.7 dB. That is comfortable for the stated component counts, but confirm the connector count includes both ends and any patch panels. If the actual link has more connectors, subtract 0.3 dB each. Also note the receiver sensitivity is at a defined BER and wavelength; do not add an arbitrary margin unless the design requires it.