FIBEROPTIC POWER METERS: Calibration method addresses high-power nonlinearities

On Sept. 19 and 20 at a National Institute for Standards and Technology (NIST) symposium for optical-fiber measurements to be held on the same campus as the NIST laboratories in Boulder, CO, NIST researchers will describe newly developed calibration procedures for optical-fiber power-meter measurements at high powers.

Erbium-doped fiber amplifiers (EDFAs) for optical communications use high-power pump lasers at both 980 and 1480 nm, according to Igor Vayshenker, Calibration Leader, in the Sources, Detectors and Displays Group of the NIST Optoelectronics Division. And while modern optical-fiber power meters (OFPMs) are capable of measuring powers above several watts propagating through an optical fiber, existing calibration methodologies cannot calibrate OFPMs for the powers that can now be routinely achieved in optical fiber.

OFPMs are generally calibrated at relatively low powers of 10 to 100 µW (-20 to -10 dBm), but launching about 1 W (+30 dBm) from a high-power pump laser into a single-mode optical fiber can yield power densities on the order of a megawatt per centimeter-squared, which can destroy connectors and the fiber itself, Vayshenker said. Nonlinearities in power-meter performance can introduce errors as large as 4% in power readings, he added. And nonlinearities that lead to inaccurate meter readings at high power may not even show up at lower powers.

Two types of nonlinearity occur between the power-meter input and output signals, he said. One is saturation, in which the input signal might increase by a factor of two but the output signal increases by a factor of less than two. The other is superlinearity or superresponsivity. The meter acts almost like an amplifier in this case. If the input signal increases by a factor of two, the output signal may increase by a factor greater than two.

The nonlinear behavior is thought to be related to the solid-state physics of the material, which manifests in the wavelength-dependent electron and hole generation in the detector material. So the same detector could conceivably exhibit different nonlinear behavior at different wavelengths. In fact, nonlinear behavior for some detector materials, such as indium gallium arsenide, tends to be much worse (by a factor of two) at lower wavelengths, such as 850 or 980 nm, than at higher wavelengths, such as 1480 nm. The detector materials are the same for measuring high and low powers, but at high powers an integrating sphere is generally used to attenuate the signal prior to detection and measurement.

The NIST calibration method is based on a triplet superposition algorithm. When two signals are directed at the detector simultaneously the output should equal the sum of the two outputs obtained when those same two signals are directed at the detector separately. So far the NIST team has developed and reported on systems for 980 and 1480 nm measurements (see figure).^{1, 2, 3}The method could also be applied to other wavelengths, such as 1310 nm, Vayshenker said.

REFERENCES

I. Vayshenker, R. Swafford, S. Yang, Natl. Inst. Stand. Technol. Spec. Publ. 1024, 145 (2004).
2. Vayshenker, R. Swafford, S. Yang, Appl. Opt. 45(6) 1098 (2006).
3. Vayshenker, S. Yang, R. Swafford, Natl. Inst. Stand. Technol. Spec. Publ. 1055, 22 (2006).