Improving IGS Timescale Stability and Tracking of UTC · NRC1 NRC * AOA SNR-12 ACT H-maser Ottawa,...
Transcript of Improving IGS Timescale Stability and Tracking of UTC · NRC1 NRC * AOA SNR-12 ACT H-maser Ottawa,...
Improving IGS Timescale Stability and Tracking of UTC
K. Senior, J. RayU.S. Naval Research Laboratory (NRL)
2004 IGS Workshop & SymposiumBern, Switzerland
Status of IGS Clock Products
• Sat clocks+ orbits (Final & Rapid) provide autonomous PPP at few cm level
• Station clock comparisons show accuracies ranging from 120ps to ~ 1ns (Metrologia 2003)
• Finals densification much improved (e.g., CODE & GFZ + new MIT)
• Rapids were previously most robust; fragile due to USNO/CODE dropouts
• Estimates referenced to IGS timescales have shown to be useful for diagnosing station problems.
Needed/Expected Improvements in IGS Clock Products
• Continued densification, now particularly with Rapids.– H-Masers + timing lab stations critical
• Timescale Improvements• Station operators could improve local clock performance
significantly in almost all cases – using clock estimates (referenced to IGS timescales) as
diagnostic tool, – thermally isolating receiver, phase-stabilized cabling, & other
known problems.
Clock “Fiducial” Stations
YAR2 (Rb)
YAR1 (Cs),
YAKT (Cs),
WUHN (Cs),
WROC (Rb),
WHIT (Rb),
VILL (Cs),
TSKB (Cs),
TROM (Rb),
TNML (Rb),
THU3 (Rb),
THU2 (Rb),
THTI (Rb),
SYOG (Cs),
SUTM (Rb),
STR1 (Cs),
SCH2 (Rb),
RIOG (Rb),
QAQ1 (Rb),
PRDS (Cs),
PERT (Cs),
OBET (Cs),
MCM4 (Rb),
MAS1 (Cs),
MAR6 (Rb),
LPGS (Rb),
KOUR (Cs),
KIRU (Cs),
KERG (Cs),
JPLM (Rb),
JOZE (Rb),
JOZ2 (Rb),
HLFX (Rb),
HERS (Rb),
HARB (Cs),
GRAZ (Rb),
GMAS (Cs),
DWH1 (Cs),
DARW (Rb),
DAEJ (Cs),
CHUR (Rb),
CAGZ (Cs),
BAN2 (Rb),
BAHR (Cs),
AREQ (Rb),
YELL (HM)YEBE (HM),
WTZR*(HM),
WTZA*(HM),
WSRT (HM),
WES2 (HM),
USUD (HM),
USNO*(HM),
USN1*(HM),
TWTF*(Cs),
TLSE*(Cs),
TIDB (HM),
TID2 (HM),
TID1 (HM),
STJO (HM),
SPT0 (HM),
SFER*(Cs),
PTBB*(Cs),
PIE1 (HM),
OPMT*(HM),
ONSA (HM),
OBE2*(Rb),
NYAL (HM),
NYA1 (HM),
NRC1*(HM),
NPLD*(HM),
NNOR (HM),
NLIB (HM),
MIZU*(Cs),
METS (HM),
MEDI (HM),
MDVO (HM),
MDVJ*(HM),
MATE (HM),
MAT1 (HM),
MAD2 (HM),
KOKB (HM),
KHAJ (HM),
KGN0*(Cs),
IRKT (HM),
IRKJ (HM),
IENG*(Cs),
HOB2 (HM),
GOL2 (HM),
GODZ (HM),
GODE (HM),
FORT (HM),
FAIR (HM),
DRAO (HM),
BRUS*(HM),
BREW (HM),
BOR1*(Cs),
AMC2*(HM),
ALGO (HM),
Primary (Hm or timing lab) Secondary (Cs or Rb not at timing lab)
List updated daily at https://timescales.nrl.navy.mil/
Wettzell, Germany H-maser AOA SNR-8000 ACT IFAGWTZR
Wettzell, Germany H-maser Ashtech Z-XII3T IFAGWTZA
Washington, DC, USA H-maser AOA SNR-12 ACT USNO * USNO
Taoyuan, Taiwan cesium Ashtech Z-XII3T TL * TWTF
Toulouse, France cesium AOA TurboRogueCNESTLSE
Boras, Sweden cesium JPS Legacy SP SPT0
San Fernando, Spain cesium Trimble 4000SSI ROA * SFER
Braunschweig, Germany H-maser AOA TurboRoguePTB * PTBB
Penc, Hungary rubidium Trimble 4000SSE SGOPENC
Paris, France H-maser Ashtech Z-XII3T OPOPMT
Oberpfaffenhofen, Germany rubidium AOA SNR-8000 ACT DLR OBE2
Ottawa, Canada H-maser AOA SNR-8100 ACT NRC * NRC2
Ottawa, Canada H-maser AOA SNR-12 ACT NRC * NRC1
Teddington, UK H-maser Ashtech Z-XII3T NPL * NPLD
Mizusawa, Japan cesium AOA Benchmark NAOMIZU
Mendeleevo, Russia H-maser Trimble 4000SSE VNIIMMDVJ
Koganei, Japan cesium Ashtech Z-XII3 CRL * KGN0
Torino, Italy Cesium Ashtech Z-XII3T IEN IENG
Brussels, Belgium H-maser Ashtech Z-XII3T ORB BRUS
Borowiec, Poland cesium AOA TurboRogueAOSBOR1
Colorado Springs, CO, USA H-maser AOA SNR-12 ACT AMC * AMC2
CityFreq. Std.GPS ReceiverTime LabIGS Site
IGS Stations Co-located at Timing Labs
IGS Time Scale• Internal IGS time scale implemented using a weighted
combination of included frequency standards
• 1 d stability improved from ~2 x 10-14 (when aligned to GPS time before) to ~1 x 10-15 now
• Steered to GPS time in the long term (~30 to 40 d)
• But longer term drifts can reach ~50 ns w.r.t. UTC
Station Calibration Bias
iiiB UTC-CLK=
( )( ) ( )
T TIGS
T IGS
CLK GPST (UTC GPST) (UTC UTC )
CLK UTC GPST GPSTGPST
i i i
i i
i
B
B
′ = − − − + −
= − + −
= + ∆
STATION CALIBRATION BIAS: includes internal GPS receiver/antenna calibration bias & intra-lab offset to UTCi
CLKi = GPS geodetic clock estimates at lab iUTCi = local realization of UTC for lab i
To relate IGS time scale to UTC, need to know prediction of (UTCi – UTC) and station timing bias:
From IGS clock products & BIPM Circular T, can compute:
small corrections due to different methods of observing GPS time
New Steering Reference
(UTC-IGSTu) = Σwi(CLKi – IGSTu) - Bi] + (UTC-UTCi)
Using LQG alg., steer to zero the quanity:
Must predictOutput of existing timescale algorithm
Can be determined provided bias is stable
3-state LQG currently used, but 2-state may suffice
AMC2 ( )B t′
BOR1( )B t′
BRUS( )B t′
KGN0 ( )B t′
NPLD ( )B t′
TWTF ( )B t′
USN1( )B t′
2.693.33639.59639.9218718552532-52729
1.561.6285287.1985287.1516415952360-52529
1.381.555.895.91807452260-52339
1.932.24-68.83-69.07737252180-52256
2.362.27-157.72-157.25282652131-52158
3.973.25619.53619.04616152069-52129
2.522.43-69.69-69.71656051922-51986USNO
1.892.241.650.76309552467-52729USN1
9.057.69289.84288.4516616052563-52729
3.664.64381243.81381242.55796852483-52562
6.134.91298.93298.87344152400-52453
7.34294.7016052316-52347TWTF
3.332.82-8075.78-8075.9011237652133-52569NPLD
2.683.44-29.26-28.9510316852421-52729KGN0
2.422.59586.15585.8323829252420-52729
0.303.46452.77454.83413652213-52374BRUS
2.182.63-25.60-25.63313252693-52724
3.513.98121.23120.79605652611-52678
2.623.3131.1531.72282652504-52531
2.022.75-352.91-353.05453452251-52296
12.6811.94-2955.26-2955.91546051854-52020BOR1
2.602.694.262.7123024752480-52729
1.472.87-452.38-451.404229952117-52437AMC2
FinalRapidFinalRapidFinalRapidRANGE
B'i WRMSB'i# ValuesMJDSTATI
ON
Determining ∆GPST
average difference in BIPM & IGS measurement of GPS time due to different satellite antenna offset values ~ -1 ns (-0.92 to -1.3 ns over last few years)
day-to-day variations in IGS clock alignment to GPS time & differences in satellites used by IGS & BIPM each day
• ∆GPST(t) causes common-mode variations in Bi' which can be used to estimate effect when multiple labs are available
lab-specific errors in local calibration & time transfer to BIPM
• Can separate ∆GPST into time-varying & quasi-static components:
GPSTGPST( ) GPST ...
i i
i
B BB t
′= − ∆
′= − ∆ − ∆ +
∆GPST Adjustments
1.222.01641.72641.592.693.33639.59639.9218718552532-52729
1.390.8385288.6885288.651.561.6285287.1985287.1516415952360-52529
1.010.456.757.241.381.555.895.91807452260-52339
1.620.91-67.80-68.301.932.24-68.83-69.07737252180-52256
2.240.30-156.46-155.082.362.27-157.72-157.25282652131-52158
3.973.25620.53620.043.973.25619.53619.04616152069-52129
2.392.15-68.85-68.782.522.43-69.69-69.71656051922-51986USNO
1.030.763.041.401.892.241.650.76309552467-52729USN1
8.607.42290.12288.279.057.69289.84288.4516616052563-52729
3.474.15381246.48381245.163.664.64381243.81381242.55796852483-52562
5.354.91301.33301.346.134.91298.93298.87344152400-52453
7.11294.607.34294.7016052316-52347TWTF
2.331.35-8074.25-8074.743.332.82-8075.78-8075.9011237652133-52569NPLD
2.442.64-28.31-27.922.683.44-29.26-28.9510316852421-52729KGN0
1.431.24586.89585.812.422.59586.15585.8323829252420-52729
0.061.31453.68454.420.303.46452.77454.83413652213-52374BRUS
1.771.55-25.73-25.692.182.63-25.60-25.63313252693-52724
1.721.92121.26120.323.513.98121.23120.79605652611-52678
2.332.8833.7734.402.623.3131.1531.72282652504-52531
0.981.49-353.19-353.192.022.75-352.91-353.05453452251-52296
12.5911.77-2954.49-2955.0012.6811.94-2955.26-2955.91546051854-52020BOR1
1.411.395.104.192.602.694.262.7123024752480-52729
0.751.93-452.02-450.721.472.87-452.38-451.404229952117-52437AMC2
FinalRapidFinalRapidFinalRapidRANGE
Bi WRMSBiB'i WRMSB'i# ValuesMJDSTATI
ON
Conclusions• Demonstrated an in situ method to transfer lab calibration to co-
located IGS geodetic receiver system– Requires no cable, connector, or configuration changes– Precision (RMS) good to ~1 ns– IGS Rapid & Final results consistent to <~1 ns
• Accuracy compared to absolute instrumental calibration:– AMC2 differences -- +4.2 ns (Rapids), +5.1 ns (Finals)– USN1 differences -- +1.4 ns (Rapids), +3.0 ns (Finals)
• Technique can be applied to continuously monitor intra-lab calibration stability
• Being used to improve steering of IGS time scales w.r.t. UTC (~EFTF April, 2004)