Transcript SuperGPS through optical networks

Long-distance fiber-optical
time transfer in the Netherlands
Jeroen Koelemeij
VU Amsterdam
Simultaneous fiber-optical delivery
of picosecond time and 10 Gb/s data
over 75 km distance
Jeroen Koelemeij
LaserLaB VU University, Amsterdam,
The Netherlands
Nikolaos Sotiropoulos
Chigo Okonkwo
Huug de Waardt
COBRA Institute, Eindhoven University of Technology,
The Netherlands
Roeland Nuijts
SURFnet (Dutch Research & Education Network),
Utrecht, The Netherlands
Timing is everything
Science
PNT
(GNSS)
Mobile
telecom
Astronomy
Aerospace
Time transfer &
Atomic clocks
Internet
Power grids
Oil & mineral
exploration
e-Financial
transactions
Time transfer – the state of the art
Optical fiber methods Satellite methods
Satellite and optical fiber methods
Method
Distance
Time uncertainty
Ref.
GNSS
>1000 km
3 – 50 ns
TWSTFT
>1000 km
1 ns
T2L2
>1000 km
0.2 ns expected
White Rabbit (fiber)
(1 Gpbs Ethernet, PTP)
10 km
0.1 - 1 ns
Optical fiber
(20 Mbps PRBS)
540 km
0.1 – 0.25 ns
Lopez et al., Appl.
Opt. (2012)
Dark optical fiber
(20 Mbps PRBS)
73 km
74 ps
Rost et al.,
Metrologia (2012)
Dark optical fiber
(10 MHz + 1pps)
69 km
(480 km)
8 ps
TDEV=0.6 ps Sliwczynski et al.,
(20 ps) @3×103 s
Metrologia (2013)
Fridelance et al.,
Exp. Astr. (1997)
www.ohwr.org
Time transfer through optical fiber
•
•
•
•
Need to correct for (unknown) propagation delay tAB
Can measure only round-trip delay: tRT = tAB + tBA
Generally, tBA  tAB (tBA = tAB + D)
Estimate: tAB = (tRT – D)/2
Optical fiber
A
B
Optical signals
Important requirement:
• Calibration of D
• Bidirectional light path through optical fiber to achieve <1 ms
THE disadvantage (so far) of OFTT
Hard to get access to fiber telecom infrastructure!
• Need ‘dark’ wavelength channel or unused ‘dark’ fiber, but…
• Telecom fiber infrastructure: mostly commercial business, fiber
owners want to earn back their investment (and more!)
• Governmental institutions possess fiber infrastructure, but often
interested only if it supports their mission/saves money
• Our approach to circumvent this: develop methods which are
compatible with high-capacity optical telecom
• Not entirely new (IEEE 1588, SyncE can make use of optical
connections)
• BUT key to sub-ns accuracy lies in the optical layer: bidirectional
optical light path is essential!
Our approach
• Test bed: 75 km, 10 Gb/s telecom link (spooled fiber) at TU/e
• Find delays via XCOR of 10 Gb/s bit streams (captured with oscilloscope)
Quasi-bidirectional amplifier
(Amemiya et al., IEEE IFCSE 2005)
25 km
50 km
Advantages:
• Time + 10 Gb/s data transfer functionality
- no telecom capacity sacrificed!
• Compatible with existing telecom methods & equipment
PRBS signals and cross correlation
50 GS/s
75 km
12.5 GS/s
150 km
PRBS signals and cross correlation
50 GS/s
75 km
12.5 GS/s
150 km
Sources of delay asymmetry
• Nonreciprocal paths in instruments, fiber patches, amps: DI
– Easily accumulates to >> 1 ns
– Calibrate: remove fiber spools and measure
– Calibrate dependence on ambient conditions and system parameters
• Choosing different wavelengths for downstream and
upstream communication adds to compatibility with existing
networks, BUT leads to asymmetry due to dispersion
– Chromatic dispersion (1.6 ns/nm per 100 km)
– Polarization mode dispersion (<10 ps for 100 km link)
• Calibrate chromatic dispersion by use of third wavelength
A l2
l3
l1
B
Chromatic dispersion calibration
1. Measure round-trip delays tAC12 and tAC13
2. For each combination (l1, lJ), calibrate all other
delay asymmetries (D = DI + DPMD)
3. Measure wavelengths lj (0.3 pm uncertainty)
4. Estimate one-way delay (qAB) using formula:
Few ns size, sub-ps uncertainty
q AB 


1
12
12
13
13
2
13
13
2
12
12
2
[
t

D

t

D
l

t

D
l

t

D
l








AC
AC
1
AC
2
AC
3
2  l2 2  l32 
 Ll12  l1  l2  l1  l3  l2  l3  n(l1 ) / c]
Few ps size, sub-ps uncertainty
(Estimate n‴ separately; see Sotiropoulos et al. Optics Express 21, 32643 (2013)
Polarization mode dispersion
25 km
50 km
• Measure two different round-trip delays with opposite
polarization states
• DPMD found by differencing the delays
• PMD leads to few-ps delay asymmetry in 75 km legacy fiber,
much less in newer optical fiber types
Every millimeter counts…
Effect of air-gap attenuator
D
Measure signal propagation delays
with 200 fs resolution!
Every dB counts…
• SOA input power
• SOA bias current
• Received power
D
Results
OWD  tAB(t) [ps]
Time difference= <OWDestimate>  <OWDdirect>
e.g.
OWDestimate =
OWDdirect
=
Estimated delay uncertainty: 4 ps
(agrees with observations)
75 km link
Bit-error rate (BER) below
10-9 :
Error free
communication
at 10 Gb/s
Measurement number
25 255(1)
50 405(1)
75 552(1)
q[ps]
q[ps]
q[ps]
DI
3.4
3.4
3.4
DPO time base stability
0.8
1.0
1.7
Fit uncertainty
1.5
1.5
1.0
VOAs
1.0
1.0
1.0
PMD correction
0.6
1.0
0.6
Wavelength measurement
0.2
0.5
0.7
XCOR interpolation
0.3
0.3
0.3
Estimate n′′′
0.05
0.1
0.1
SPM and XPM
<0.1
<0.1
<0.1
Fast fiber length fluctuations
<0.1
<0.1
<0.1
Sourcea
b
Link length uncertainty
Total
<10
4
4.0
<10
4
4.1
75 km
log BER
Link length [m]
369 287 563.9(4.2) ps
369 287 565.6(0.8) ps
25 km
0 km
50 km
<104
4.2
Received power [dBm]
Results
OWD  tAB(t) [ps]
Time difference= <OWDestimate>  <OWDdirect>
e.g.
OWDestimate =
OWDdirect
=
Estimated delay uncertainty: 4 ps
(agrees with observations)
75 km link
Bit-error rate (BER) below
10-9 :
Error free
communication
at 10 Gb/s
Measurement number
25 255(1)
50 405(1)
75 552(1)
q[ps]
q[ps]
q[ps]
DI
3.4
3.4
3.4
DPO time base stability
0.8
1.0
1.7
Fit uncertainty
1.5
1.5
1.0
VOAs
1.0
1.0
1.0
PMD correction
0.6
1.0
0.6
Wavelength measurement
0.2
0.5
0.7
XCOR interpolation
0.3
0.3
0.3
Estimate n′′′
0.05
0.1
0.1
SPM and XPM
<0.1
<0.1
<0.1
Fast fiber length fluctuations
<0.1
<0.1
<0.1
Sourcea
b
Link length uncertainty
Total
<10
4
4.0
<10
4
4.1
75 km
log BER
Link length [m]
369 287 563.9(4.2) ps
369 287 565.6(0.8) ps
25 km
0 km
50 km
<104
4.2
Received power [dBm]
OWD  tAB(t) [ps]
Results
Delivery of 10 Gb/s optical data with 4 ps
uncertainty over 75 km distance
75 km link
Measurement number
log BER
75 km
25 km
0 km
50 km
N. Sotiropoulos et al., Optics Express 21, 32643 (2013)
Received power [dBm]
Speed bonus
• Delay determination/synchronization requires a
single shot of 10 Gb/s data lasting less than 1 ms
– Comparison: state-of-the-art fiber methods require
10-100 s of averaging to achieve 4 ps stability
Dissemination of UTC(VSL) via WR
from Delft to Amsterdam
Tjeerd Pinkert
Jeroen Koelemeij
LaserLaB VU University, Amsterdam,
The Netherlands
Erik Dierikx
Martin Franssen
VSL Delft
The Netherlands
Nikolaos Sotiropoulos
Huug de Waardt
COBRA Institute, Eindhoven University of Technology,
The Netherlands
Rob Smets
SURFnet (Dutch Research & Education Network),
Utrecht, The Netherlands
Henk Peek
NIKHEF Amsterdam, The Netherlands
Work in progress…
• Demonstrate optical
time transfer from VSL (Delft)
to NIKHEF (Amsterdam)
• Fiber link provided by
SURFnet
• Amplified fiber link to
distribute UTC(VSL) via WR
WR link VSL – NIKHEF (2×80 km)
Bidirectional (dark) fiber link
l11471nm
l21491nm
VSL Delft
link VSL-TU Delft
Leiden
l2
Delft
l measurement
Rx
Atomic
clock
time & freq
comparison
l2
SOA
l1
WDM
SOA
WDM
WDM
Tx
WDM
WR node 2
WR node 3
time & freq
reference
UTC(VSL)
time & freq comparison
Rx
WDM
SOA
Tx
l1
time & freq reference
Atomic
clock
SOA
Amsterdam
WDM
Tx
WDM
WR node 1
WDM
Rx
NIKHEF
SURFnet link
Tx
WR node 4
Rx
Towards large-scale depoyment in live
networks
Under development:
v1 bi-di OLA
(developed i.c.w. TU Eindhoven / SURFnet /
VTEC Lasers and Sensors)
• Includes remote control & monitoring
through Ethernet
• > 20 dB gain (fiber-to-fiber)
• Delay asymmetry calibrated at ps level
Commercial v2 bi-di OLA
 Remote control & monitoring,
compatible with telecom infrastructure
& standards
 Including filters for bi-directional
bypass in live DWDM networks
(patent pending)
 Create bi-di optical path for WR TFT
w/o insertion loss
 Delay asymmetry calibrated at ps level
 Create bi-di optical path for ultrastable
frequency dissemination
 Commercially available
(expected launch Q4 2014)
Outlook: SuperGPS for science & society
Planned (VU Amsterdam, TU Delft & stakeholders)
4 ps  2.4 mm uncertainty (4D positioning)
Thank you!
Contact: [email protected]