Upload
lytuong
View
219
Download
1
Embed Size (px)
Citation preview
Degradation of Optical Performanceof Fiber Optic Connectors in aManufacturing Environment
T. Berdinskikh, N. Albeanu and S. StaffordCelestica International,
[email protected], [email protected], [email protected]
D. Silmser and H. TkalecAlcatel Canada,
[email protected], [email protected]
J. NguyenSolectron Corporation,
Degradation of return loss in connectors, due to frequent reconnection, in a manufacturing environment has been investigated.
Degradation by contamination and damage to the connector endface causes an air gap between matching connectors.
Degradation of Optical Performance of FiberOptic Connectors in a Manufacturing Environment
ÿ Degradation of return loss in connectors, due tofrequent reconnection, in a manufacturingenvironment has been investigated.Degradation by contamination and damage tothe connector endface causes an air gapbetween matching connectors.
ÿ A group of 27 fiber optic cables with singlemodeSC connectors was evaluated. The cables wereoriginally used for traffic testing and opticsverification during the functional testing processof a line interface input/output card. Eachconnector had undergone approximately 1200mating and demating operations during theeighteen months of operation in a standardmanufacturing environment. An average particlecount of particles 0.5 mm or greater measured480,550 particles/m3 in this manufacturingenvironment. This is classified as an ISO Class 8environment. The cleaning process was notstrictly enforced during the time that thesecables were in use.
ÿ As seen in figures 1 and 2, connectorendfaces appeared to be severely damaged.Pits, scratches, chips and contaminationwere all clearly visible with fiberscopes.Based on the SEM analysis the typical sizeof the pits at the cladding area was 1-6 mm.Several samples showed the presence ofnon-removable contamination with a smallpresence, (approximately 5%), of tinparticles. The size of non-removablecontamination at the cladding layer ofseveral samples was 20-22 mm with athickness of 0.18-0.2 mm.
ÿ Figure 3 shows a histogram of RL for thegroup of degraded connectors. Theaverage RL is 41.9 +/- 8.3 dB at 1310 nmand 42.9+/-8.3dB at 1550 nm. Thedegraded cables have poor RL whencompared with good cables, which typicallyhave a RL of less than -55dB.
ÿ Our previous studies have shown that pitsand scratches have a greater effect on RL,while contamination on the core region ofthe fiber affects the IL of the connector.The high RL can therefore be explained bythe presence of a large number of pits andscratches on the connector endface. Thesepits and scratches are likely the result ofcontamination on the endface . Ifcontaminants were present on theconnector during the connection, theywould be pressed against the endfaceleaving marks, scratches, or pits andpossibly even embedded particles.
Typical Return Loss Measurements
010203040506070
Connector ID RL-1310 RL-1550
Figure 1 – Severely damaged andcontaminated connector enface
Figure 2 - Severely damagedconnector enface
Figure 3 – Typical Return LossMeasurements
Degradation of Optical Performance of FiberOptic Connectors in a Manufacturing Environment
ÿ Contamination on the connector endface due tomicroparticles which, if not removed, arepressed against the endface causing an air gap.The air gap created by trapped particles betweenthe two ferrule endfaces is called thecontamination layer and is characterized by thethickness, dc, and the refractive index, nc.
ÿ In the case of contamination with microparticlesthat do not block the core of the fiber, nc=1 andthe contamination layer can be characterized bythickness only. A diagram identifyingcontamination layer is shown in figure 4. To ourknowledge these types and sources ofcontamination have not been identified and theeffects of the connector contamination onoptical performance have not been investigatedin detail.
ÿ Figure 5 shows the calculated RL values in whichthe air gap was caused only by geometric factors(undercut and apex offset with forcecompensation). No contamination layer has beenconsidered in this case. Three out of 54connectors, (about 5.5%), show a good matchbetween calculated and measured values. Theremaining connectors had higher measured RLthan calculated RL. The air gap caused bycontamination and/or the rough damagedendface surface may be responsible for thedegradation in RL.
ÿ Figure 6 shows the calculated RL where thecontamination layer, in addition to geometricparameters, is considered and comparedwith the measured RL. The thickness of thecontamination air gap is calculated, and itfalls in the range of 1-4 nm. As seen on theplot, the calculated RL values follow closelythe measured RL values. The model canexplain 48 out of 54 samples (about 89%).
ÿ In this paper the degradation of RL of fiberoptic connectors in a standard environmenthas been investigated. It was shown that aninsufficient cleaning process combined withfrequent reconnection resulted in thecontamination and damage of the connectorendface. As a result, the physical contactbetween matching connectors can not beachieved. A RL model has been developed forthe presence of small (1-4 nm) air gapsbetween the fiber ends. Proper cleaning andhandling processes can decrease long termdegradation of fiber optic connectors in astandard manufacturing environment.
n nnc
dc
High index layer (h, nh)due to polishing
Figure 6 - Measured RL values compared withcalculated RL values where air gap wasassumed to be caused by geometric factorsand the contamination layer
Figure 4 – Contamination layer causing anair gap
Figure 5 – Measured RL values compared withcalculated RL values where air gap wasassumed to only be caused by geometric factors
Figure 7 – Contamination layer thickness
0 10 20 30 40 50 60 0
5
10
15
20
25
Contamination Layer Thickness [nm]
Num
ber
of s
ampl
es
Contaminated connectors: histogram of the contamination layer thickness at 1310nm
˙˚
˘ÍÎ
Ș̃¯
ˆÁÁË
ʘ¯
ˆÁË
Ê--= ccdnRdBRL
lp4
cos12log10)(
)cos21()cos2( 212
22
1212
22
1 dd rrrrrrrrR ++++=
hnnn
nnr
nn
nnr h
ch
ch
h
h
lp
d4
,, 21 =+
-=
+
-=
0 10 20 30 40 50 6060
50
40
30
20
10
0
Sample Number
Ret
urn
Los
s dB
RL values - air gap caused by geometricparameters and contamination layer
0 10 20 30 40 50 6060
50
40
30
20
10
0
Sample Number
Ret
urn
Los
s dB
RL values - air gap caused only bygeometric factors
Calculated RLMeasured RL