Kingfisher KI 2300 Series, Operation And Maintenance Manual

Page: 37 / 74

ACCURACY CONSIDERATIONS
KI2000 UM-6
-
35 -
Keep optical connectors clean and in good condition.
To reduce the effect of polarisation changes, the fiber system should
be neat, coiled and physically stable.
In multimode systems, modal noise and general uncertainty are much
worse than in single mode systems. Optimum measurement
repeatability will be obtained by use of LED sources and a mandrel
wrap.
Power Meter
For best absolute power testing accuracy, the meter calibration
wavelength should be set close to the measured wavelength. Note
that the wavelength uncertainty of the source may be 30 nm or so.
Problem areas tend to be InGaAs below 1000 nm and Ge above 1550
nm.
Use of an APC connector at the power meter interface will reduce line
reflections and improve measurement stability in laser based single
mode systems. This is important when measuring small connector
losses.
For high precision use, power meters provide better accuracy when
measuring a non-coherent source such as an LED. So for best
accuracy when testing small losses on single mode patch leads etc,
an LED source may be preferred.
High power (attenuated) meters tend to suffer overall reduced
accuracy. Kingfisher high power meters have better general accuracy
than most others, since our meters use a more sophisticated
attenuating element which reduces the adverse effects of reflections
and wavelength sensitivity.
At levels below about -45 dBm, spurious readings can be caused by
stray light leaking into fiber and patch cables. This can be a particular
problem with extended range linearity measurements.
Ge power meters have limited linearity, sensitivity and accuracy
compared to all other types.
Light Source
Light source power may drift. When you have finished a test, go back
to the start position to check if the light source power is still within
acceptable limits. Note that light source specifications are usually for
typical drift after a warm up period, but the actual drift will vary between
instruments and test situations.