The first aspects concern both designs.
The laser wavelength is used as phase reference and thus an acurate knowledge
of the wavelength and its stability is
essential. The tunable laser used in the time multiplexing set-up guaranties a
wavelength stability better than 1 pm and the DFB laser used in the
wavelength multiplexing set-up is temperature stabilized to ensure the same
stability. A wavemeter has been used in parallel to track the exact wavelength
position and eventual drifts (laser wavelength drifts smaller than 0.1 pm
were observed). The thermal stability of the test FBG is fundamental as the
measurement time goes from several minutes to several hours. Temperature
changes in the test FBG modify its spectral properties (in first approximation
a frequency shift) that are seen in the OLCR measurement by a phase difference
slope modifications due to a different local Bragg condition.
The time multiplexing design shows
several advantages over the wavelength multiplexing scheme :
-
The laser reflective reference is the
FBG itself and thus cancels completely the laser phase noise problem
encountered with the cleaved fiber end as the reference
-
The phase difference is ranged in tens
of radians compared to the millions of radians range of the phase itself for
the same scan distance (e.g. a few centimeters)
-
The sampling is not limited by the
Nyquist condition and this reduces considerably the measurement time (several
minutes instead of several hours) and the amount of data (factor 10 to 100)
-
The single wavelength operation
divides by two the number of optical and electronic components (adding on the
other hand an optical switch) and guaranties the perfect 2p modulation for
both OLCR and laser phase signals
The wavelength multiplexing exhibits
nevertheless interesting features :
-
The wavelength symmetry enables
measurements of FBG in the 1310 and 1550 nm range by simple source
exchange
-
For a dynamic measurement
configuration (§4.3.2), the laser signal can be used directly as reference
signal for the lock-in amplifier and then the phase difference is directly
extracted without any other operation (accurate constant mirror velocity is
required)
-
The time delay between the laser and
OLCR phase measurement is very small and can either be totally suppressed if
the voltage difference of both lock-in is measured, reducing the marginal phase
drifts well bellow the p/100 encountered for the time multiplexing design
We want now to discuss the minimal
sampling distance required for FBGs OLCR measurements when the OLCR phase is
obtained from the difference with the laser reference phase. The phase
difference change Df for two different wavelength l1 and l2 over the same distance d is given by
|
(4-45) |
where Dl = l2-l1. The following table gives the
minimal distance d2p for which a complete 2p phase change is
obtained for l1 = 1310 nm.
l2 [nm]
|
Dl [nm]
|
d2p [mm]
|
1310.1
|
0.1
|
17'161
|
1311
|
1
|
1'716
|
1320
|
10
|
172
|
1410
|
100
|
17
|
1550
|
240
|
8.5
|
When measuring the phase difference
between the laser and OLCR signals, the sampling has to be half the minimal
distance d2p to fulfill the Nyquist criteria. We observe that the wavelength multiplexing
design could also be used in the difference phase mode but with smaller
sampling intervals and an independent distance measurement. In this discussion,
we have neglected the phase changes introduced by the test FBG itself. For
homogeneous FBGs, additional p shifts in the impulse response phase are observed due to global
reflections at the grating interfaces and they are spread in the OLCR
measurement over the broadband light source coherence length by convolution.
For chirped grating, the Bragg condition change is usually smaller than 10 to
20 nm and then the corresponding d2p remains
under 100 mm. The typical OPLD sampling distance (that is twice the incremental
change of the mirror position) used in time multiplexing design was 20 mm, corresponding
to 10 mm mirror step. Several experiments have also been conducted with
1 mm mirror steps for precise OLCR measurements and 100 mm for fast
measurements.
|