Transcript SOPMD (ps 2 ) - New Ridge Technologies

Test Plan for PMD Testing of
a WDM Receiver
Henry Yaffe, Principal
January 2004
Recommended Test Program
•
•
Process

Establish test procedure (automation recommended)

Schedule time in front of systems
Experiments:
1.
Measure bare receiver
2.
Use NRT PMD Source to map to Joint Probability Distribution Function
•
Analyze data - Calculate Total Outage Probability for different mean
span PMDs.
•
Addendum

If desired, add more impairments (CD, non-linearities) and re-measure
www.NewRidgeTech.com
2
Test Basis: System Margin (OSNR) as
Performance Benchmark
BER vs. OSNR
Back-to-back
10-6
40 ps PMDC
60 ps PMDC
80 ps PMDC
60 ps uncompensated
BER
10-12
Relative OSNR
Penalties
Back to
Back
OSNR
Recommended: 10-12 BER as limit (else, each test is too time-consuming)
www.NewRidgeTech.com
3
Recommended Tests to Acquire OSNR
Performance for 25 Ps PMD
Ex.: 17 states can adequately cover the 25psec (25%) contour
DGD
(ps)
SOPMD
(ps2)
0
0
20
0
566
1183
40
0
894
1386
2263
55
1097
1598
2297
60
0
70
2051
75
1183
1488
80
0
1587
55/2297
40/2263
40/1386
70/2051
55/1598
80/1587
75/1488
75/1183
20/1183
40/894
55/1097
20/566
0/0
20/0
40/0
60/0
www.NewRidgeTech.com
80/0
4
Acquiring the data
Test basis:

Set & maintain all system parameters

Vary OSNR for a given PMD state & measure BER
Characterization process:
1.
Set PMD state
2.
Set OSNR
3.
Set RX power
4.
Record: payload, B3 BERs, error counts, elapsed time, OSNR, and RX
power every 30 seconds
5.
Continue recording for >5 min and <30 min (or 10 hits)
6.
Increment OSNR, repeat
www.NewRidgeTech.com
5
Test Stand Required for PMD Testing
2
1
PMDE
Tx
12
3*
25 km
5
SMF-28
Polarization
scramblers
BERT
6
7
EDFA
Attenuators
OSA
10*
Rx
90/10
10 %
11
PMDC*
9
Tunable
Filter
10 %
8
90/10
Power
meter
Automated BER v OSNR at fixed RX power
•Allows efficient measurement of many PMD states
• Decreases data acquisition time by 3-4x
•* Polarization scrambler 3 is not needed if no optical PMDC is used
www.NewRidgeTech.com
6
Test Setup – Equipment Key
1.
Launch Polarization Scrambler – ensures no launch at SOP, which could create
artificially high-quality eye
2.
PMDE – controllable PMD source
3.
PMDE Output Polarization Scrambler (optional, only use with 10) – randomly
varies SOP to provide “speed” input
4.
Fiber spool – source of “residual” chromatic dispersion (appropriate for chirped
systems)
5.
EDFA input attenuator – ASE noise control (OSNR)
6.
EDFA – ASE source
7.
EDFA output attenuator – power control
8.
OSA – OSNR measurement
9.
DWDM Filter – models DWDM effects
10.
PMDC (optional) – compensates for PMD
11.
Power meter – power measurement to maintain constant Rx dBm
12.
BERT – measure BER as performance output
www.NewRidgeTech.com
7
Test Conditions
• To establish consistent and verifiable test results

Testing should be conducted PMD as the only impairment

A benchmark system operating level should be set and maintained
(Rx power, etc.)
• For a given WDM system:

Maintain a constant received power (dBm) level

Turn FEC off (or measure pre-FEC)

Turn SBS Suppression off (if used)

Maintain a fixed (preferably zero, to avoid confusing effects) level
of residual chromatic dispersion at the RCVR
www.NewRidgeTech.com
8
Benefits of Automation Can Rapidly Acquire
Data to Model Real-world Performance
•
•
•
For network-realistic PMDC characterization:

Reasonable sampling of PMD states is required

Try to define points where DGD & SOPMD performance degrades

Probability of reaching any those states is relevant to system performance
Testing each data point is time-consuming:

Need ~1hr for good statistics at 10-12 BER

Measure for minimum of 5-10 minutes for full SOP coverage

BER v OSNR curves (3-4 points) takes 1 to 2 hours
Automation improves confidence of characterization

Can characterize 10 to 20 PMD states in 24 hours with automation
– Compare to ~ 5 to 8 manually per 8 hour shift

OSA, power meter, 2 attenuators, BERT, PMDE are GPIB interfaced

PMD, test time, polling time, OSNRs, and RX powers set by user
www.NewRidgeTech.com
9