Preparation of the Inter- Laboratories Comparison

Transcription

1 Preparation of the Inter- Laboratories Comparison Yi-Jiun Huang GPS time-transfer and calibration techniques Concluding Workshop, Taoyuan, Taiwan October 25-26, 2017

2 The Inter-Laboratories Comparison (ILC) Activities The goal is for attendees to be familiar with Understanding of the GPS calibration for UTC(k) Use and circulation of the travelling system Process of the receiver data Report of the calibration results Besides, TL applied to the GNSS working group of TCTF, for registering an BIPM calibration activity during the ILC

3 Understanding of the GPS calibration

4 Time interval by counter UTC(k) clock Time Interval Counter L. Marais, 2016, GPS time transfer and time scales: What does the BIPM do with my data?

5 How to compare two distant clocks? UTC(k) UTC(j) Time Interval Counter? a a L. Marais, 2016, GPS time transfer and time scales: What does the BIPM do with my data?

6 GPS time transfer GPS SV UTC(k) UTC(j) GPST GPST Time Interval Counter Time Interval Counter UTC(k) GPST Reading: 15 ns Common-view: [UTC(k) UTC(j)] = = 5 ns UTC(j) GPST Reading: 10 ns

7 Calibration need for time transfer UTC(k) UTC(j) GPS Receiver UTC(k) GPST Reading: 15 ns GPS Receiver UTC(j) GPST Reading: 10 ns [UTC(k) UTC(j)] = 5 ns + internal delay internal delay The calibration is to determine the value of the internal delays

8 The internal delay Antenna Internal Delay, X S GPS Antenna Internal delay (INTDLY) is defined by X S + X R Receiver Internal Delay, X R Here the receiver starts the measurement UTC(k) PPS GPS Receiver UTC(k) 5, 10MHz Using remote calibration, the INTDLY can be determined without power-off

9 GPST We want: UTC(k) UTC(k) - GPST PPS in the cables and the receiver GPS time transfer GPS Signal in the onboard circuit GPST GPS satellite Space vehicle (SV) GPS Signal on the air We have: Pseudorange GPS Receiver t GPS signal in the cables and the receiver PPS UTC(k)

10 The receiver looks like a counter Start of the counter when UTC(k) arrives Stop of the counter when GPS signal arrives We want: UTC(k) - GPST Correction!! Readings from the receiver: Pseudorange, ms GPST t UTC(k)

11 Corrections for time transfer GPST We want this: UTC(k) Reference delay, d GPST to k-th SV, d k GPST k-th GPS SV Geometry delay, ρk c Ionosphere, I k We get this: Pseudorange Troposphere, T k Internal (antenna), d GPS Receiver Antenna cable, d t Internal (receiver), d UTC(k)

12 Corrections for time transfer τ k = d k + ρk c + Ik + T k + d + t REFGPS τ k : pseudorange, readings of GPS receivers t REFGPS : time difference d k : delay from GPST to transmission antenna, or total delay of the k-th GPS SV In dual-frequency case, group delay differential is included d: delay from reception antenna to REF (e.g. UTC(k)), namely total delay of the receiver ρ k : geometry distance; c: speed of light Sagnac effect and the effect due to relativity are included I k /T k : Ionosphere / troposphere delay These are used in the BIPM calibration software

13 Corrections for time transfer For precise applications, several items should be taken into consideration: Precise orbit and clock for GPS SV Phase center variation in the antennas Phase wind-up effect due to relative rotation between SV and the antenna Solid earth tides, ocean loading, earth rotation variation, etc. These are not used for BIPM calibration below the noise level of pseudorange

14 Some models of the corrections Ionosphere dual-frequency combination Klobuchar model Global ionospheric map (single layer model, SLM) etc. Troposphere empirical model (height, elevation) Vienna mapping function etc.

15 Common view t REFGPS and some corrections in the CGGTTS data Data are generated by the BIPM software r2cggtts modified by NMIA to have modeled ionosphere delays For time transfer, receivers are connected to different clocks For calibration, two receivers are connected to the same clock common clock differential calibration

16 Common view for calibration GPS SV k τ L k = d k + τ T k = d k + x L k c x T k c + I k + T k + d L + t REFGPS + I k + T k + d T + t REFGPS x L k x T k GPST => τ T k τ L k x T k x L k obtained from broadcasted navigation data c = d T d L given one, pseudoranges: compute the other readings of the two receivers Local Traveling REF

17 More on common-view In geodetic receivers, carrier phases give more precise measurements Suppose that corrections for the two receivers are identical common clock antenna positions and hardware delays are stationary BIPM released the software DCLRINEX to calibrate without using CGGTTS The internal delays of G1 labs are determined by DCLRINEX in the report

18 More on common-view GPS SV k τ L k = d k + τ T k = d k + xk c + I k + T k + d L + t REFGPS xk + x + I k + T k + d c T + t REFGPS x k x k + x GPST => τ T k τ L k = xk + x x k where d T,L d T d L c + d T,L Local x Traveling UTC(k)

19 DCLRINEX software x = x 2 + y 2 + z 2 ρ x = x = y y ρ Taylor expansion: f(x+δx)=f(x)+f (x) Δx x + x = x + x x + y y + z z ρ ρ ρ => x + x x = x ρ x + y ρ y + z ρ z τ T k τ L k = x ρ x + y ρ y + z ρ z + d T,L

20 DCLRINEX software Four unknowns x, y, z, d T,L and many equations The unknowns are estimated in the least squares sense baseline internal delay

21 Target of the GPS calibration Check the connections and the components PPS/10MHz may be ill-functional Cables/antenna may be broken Renew the coordinates for the antenna Update the internal delay with respect to a BIPM reference (code: BP0R at present) follow the BIPM guideline

22 Preparation for the ILC activity

23 GPS receivers in TL Ashtech Z12T (Reference, code: TLT1) NMIA Topcon/Javad Euro-80 (Traveling, code: TRVL) Cal_ID: Use time interval counters to link the UTC(k) and the time base inside the receiver

24 Specs of TRVL NMIA Topcon/Javad Euro-80 Antenna: Hemisphere A45 Support C/A, P1 and P2 code RINEX and CGGTTS Stable in a long term Repeatable calibration values during power cycles User instruction provided by NMIA An LMR400 cable of 30 m for antenna cabling

25 Long-term stabilities According to the BIPM guidelines, the receiver must demonstrate sufficient stability over a time period comparable with the campaign Reference receiver AU01 Data provided by NMIA MJD C1 INTDLY / ns

26 For the equipment outbound Thread 5/8-11 for the antenna mount Do you have a pole to mount? Yes Length from rooftop to the lab Antenna cable of 30 m is sufficient? Yes Additional requirements Low-noise amplifier? No Tripod? No Appearance and weight of the box Take photos!! Measure the weight (~ 30kg)!!

27 Installation 2 4 PPS Cable Your Own Power Cord 3 AC V, V Antenna Cable PWR 1 ANTENNA 1PPS IN REF IN NMIA Topcon/Javad Euro-80 Your Own Cable UTC(k) 10MHz UTC(k) 1PPS Equipment List: 1. GPS Receiver 1 2. Antenna 1 3. Antenna Cable 1 4. PPS Cable 1 5. Case 1 Start measurement: 1. Check the items in the confirmation sheet 2. Carefully mount the antenna with its thread lubricated 3. Connect cables as shown above 4. Power on 5. Make sure if the data appear in the PC Complete measurement: 1. Power off 2. Disconnect 3. Complete, sign and send the confirmation sheet to TL by 4. Pack equipment as the original

28 Installation Mount the antenna at a location that has a clear view of the sky Connect a dedicated PPS cable to the UTC(k) if possible To minimize the effect of various cables on the reference delay A photo is required to show that, the baseline between the traveling and visited stations is short enough

29 Check the short baseline

30 Evaluation for the TRVL To see if the GPS system can reflect clock variations GPS time transfer v.s. time interval counter (TIC) GPS time transfer TLT1 TRVL UTC(TL) 1PPS 1PPS HP 5071A TIC (SR620) Trig. Lvl: 0.5V; Term.: DC; Impedance: HiZ

31 Evaluation for the TRVL 30 s averaging for GPS data 780 s averaging for GPS data

32 The ILC activity

33 BIPM Calibration BIPM separated UTC labs into two groups, the G1 and G2. UTC(G1) is calibrated by BIPM directly UTC(G2) is calibrated by G1 labs TL registered an ID Advised by NMIA(Australia), TL is in charge of the activity among NIMT(Thailand), NMIM(Malaysia) and VMI(Viet Nam).

34 Schedule Two weeks are expected for shipment and measurement Measurement at TL: 2 w Taiwan to Thailand: 2 w Measurement at NIMT: 3 w Thailand to Malaysia: 4 w Measurement at NMIM: 2 w Malaysia to Vietnam: 2 w Measurement at VMI: 2 w Vietnam to Taiwan: 2 w Measurement at TL: 2 w Expected 18 weeks for a round Actual 21 weeks

35 Equipment arrival / departure Unpack and check if anything failed / broken Sign and submit the confirmation sheet Mount / unload the GPS antenna Take a photo to confirm the baseline between GPS antennas is short enough Start recording GPS data Pack everything into box For avoiding any trouble during shipment, colleagues are asked to take a photo and measure the weight

36 Data Processing

37 Raw (common clock) difference The raw values T (traveling receiver) and L (local receiver) are computed from CGGTTS data T (or L) = REFSV + MDIO + MDTR + INT DLY + CAB DLY REF DLY T L : raw difference (RAWDIF) or common clock difference (CCD) Compute RAWDIF for each time, each code, and each satellite Find the median of RAWDIF over satellites and times for each code The physical meaning of T, L and T L : τ T k x T k c τ L k x L k c = d T d L

38 INT DLY ns INT DLY INT DLY Calibration Trip MTTO TLT1 LS2P UTC(NIMT) VM12 UTC(TL) UTC(NMLS) *Results for C1 code only UTC(VMI)

39 INT DLY ns INT DLY INT DLY INT DLY Calibration Trip MTTO TLT1 TRVL LS2P UTC(NIMT) VM12 UTC(TL) UTC(NMLS) *Results for C1 code only UTC(VMI)

40 Check the common clock TLT1 TRVL SDI HPDA-15RM SDI HPDA-15RM 20 MHz 10 MHz SDI FS040 SDI FS020 UTC(TL) SDI HPDA-15RM SDI PD10-RM 5 MHz AOG-110 1PPS H. Maser HM3011

41 CGGTTS GENERIC DATA FORMAT VERSION = 2E REV DATE = RCVR = Ashtech Z12T CH = 12 IMS = LAB = TL X = m Y = m Z = m FRAME = ITRF COMMENTS = NO COMMENTS INT DLY = ns (GPS C1), ns (GPS P1), ns (GPS P2) CAL_ID = CAB DLY = 0.0 ns REF DLY = 0.0 ns REF = UTC(TL) CKSUM = FF SAT CL MJD STTIME TRKL ELV AZTH REFSV SRSV REFSYS SRSYS DSG IOE MDTR SMDT MDIO SMDI MSIO SMSI ISG FR HC FRC CK hhmmss s.1dg.1dg.1ns.1ps/s.1ns.1ps/s.1ns.1ns.1ps/s.1ns.1ps/s.1ns.1ps/s.1ns G02 FF L1C FF G02 FF L1P FF G02 FF L2P FF G02 FF L3P FF CGGTTS GENERIC DATA FORMAT VERSION = 2E REV DATE = RCVR = NMIA Topcon/Javad Euro-80 CH = 12 IMS = LAB = TL X = m Y = m Z = m FRAME = ITRF COMMENTS = NO COMMENTS INT DLY = ns (GPS C1), ns (GPS P1), ns (GPS P2) CAB DLY = 0.0 ns REF DLY = 0.0 ns REF = UTC(TL) CKSUM = FF CAL_ID = xxxx-2017 SAT CL MJD STTIME TRKL ELV AZTH REFSV SRSV REFSYS SRSYS DSG IOE MDTR SMDT MDIO SMDI MSIO SMSI ISG FR HC FRC CK hhmmss s.1dg.1dg.1ns.1ps/s.1ns.1ps/s.1ns.1ns.1ps/s.1ns.1ps/s.1ns.1ps/s.1ns G02 FF L1C FF G02 FF L1P FF G02 FF L2P FF G02 FF L3P FF

42 Calibration Using CGGTTS Compute RAW for each station, each PRN, and each time stamp, RAW = REFSV + MDIO + MDTR + INTDLY + CABDLY REFDLY Use INT, CAB, REF values in CGGTTS data For C1, RAW TLT1 = = For C1, RAW TRVL = = Compute RAWDIF TRVL-TLT1 = median(raw TRVL RAW TLT1 )= Apply new reference delay to TRVL. New REFDLY: 0.0 Compute ΔSYSDLY TRVL-TLT1 = RAWDIF TRVL-TLT1 + REFDLY TRVL REFDLY TLT1 = Apply new cable delay to TRVL. New CABDLY: 0.0 Compute ΔINTDLY TRVL-TLT1 = ΔSYSDLY TRVL-TLT1 CABDLY TRVL + CABDLY TLT1 = Since ΔINTDLY TRVL-TLT1 = INTDLY TRVL INTDLY TLT1 We have INTDLY TRVL = ΔINTDLY TRVL-TLT1 + INTDLY TLT1 = = 176.7

43 RAWDIF / ns Time Deviation / s 1.0E-08 RAWDIF in TL, closure C1, median = ns P1, median = ns P2, median = ns 1.0E E C1-241 P1-242 P MJD 1.0E E E E E E+06 Averaging Time / s

44 INT DLY ns ns INT DLY ns INT DLY Calibration Trip TRVL MTTO TLT1 TRVL UTC(NIMT) LS2P UTC(TL) VM12 UTC(NMLS) UTC(VMI) *Results for C1 code only

45 Check the short baseline

46 Check the common clock Topcon Euro 80 1 PPS 1.77 ns Universal Counter TRVL 10 MHz 1 PPS 7.07 ns 1 PPS 10 MHz UTC(NIMT) RF distribution Amplifier 1 PPS 10 MHz RF distribution Amplifier 10 MHz 1 PPS 88.1 ns HROG 10 MHz 5 MHz 1 PPS 8.4 ns Cs 5071A-001 BIPM code

47 GGTTS GPS DATA FORMAT VERSION = 01 REV DATE = RCVR = NML Australia Topcon Euro-80 L1/L2 CH = 12 IMS = NML Euro-80 L1/L2 Pseudorange differences LAB = NIMT (system #2) X = m Y = m Z = m FRAME = 2005/08/26 COMMENTS = None INT DLY = 42.9 ns CAB DLY = ns REF DLY = 7.07 ns REF = UTC(NIMT) CKSUM = 67 To be updated PRN CL MJD STTIME TRKL ELV AZTH REFSV SRSV REFGPS SRGPS DSG IOE MDTR SMDT MDIO SMDI MSIO SMSI ISG CK hhmmss s.1dg.1dg.1ns.1ps/s.1ns.1ps/s.1ns.1ns.1ps/s.1ns.1ps/s.1ns.1ps/s.1ns 24 FF EB CGGTTS GENERIC DATA FORMAT VERSION = 2E REV DATE = RCVR = NMIA Topcon/Javad Euro-80 CH = 12 IMS = LAB = NIMT X = m Y = m Z = m FRAME = ITRF COMMENTS = NO COMMENTS INT DLY = ns (GPS C1), ns (GPS P1), ns (GPS P2) CAB DLY = 0.0 ns REF DLY = 0.0 ns REF = UTC(NIMT) CKSUM = FF CAL_ID = xxxx-2017 SAT CL MJD STTIME TRKL ELV AZTH REFSV SRSV REFSYS SRSYS DSG IOE MDTR SMDT MDIO SMDI MSIO SMSI ISG FR HC FRC CK hhmmss s.1dg.1dg.1ns.1ps/s.1ns.1ps/s.1ns.1ns.1ps/s.1ns.1ps/s.1ns.1ps/s.1ns G24 FF L1C FF G24 FF L1P FF G24 FF L2P FF G24 FF L3P FF

48 Calibration Using CGGTTS for MTTO Compute RAW for each station, each PRN, and each time stamp, RAW = REFSV + MDIO + MDTR + INTDLY + CABDLY REFDLY Use INT, CAB, REF values in CGGTTS data For C1, RAW MTTO = = For C1, RAW TRVL = = Compute RAWDIF TRVL-MTTO = median(raw TRVL RAW MTTO )= Apply new reference delay to MTTO. New REFDLY: 7.1 Compute ΔSYSDLY TRVL-MTTO = RAWDIF TRVL-MTTO + REFDLY TRVL REFDLY MTTO = Compute ΔSYSDLY MTTO-TLT1 = ΔSYSDLY TRVL-TLT1 ΔSYSDLY TRVL-MTTO = (-238.3) (-19.9) = Apply new cable delay to MTTO. New CABDLY: Compute Consider the same as before ΔINTDLY MTTO-TLT1 = ΔSYSDLY MTTO-TLT1 CABDLY MTTO + CABDLY TLT1 = We have INTDLY MTTO = INTDLY TLT1 + ΔINTDLY MTTO-TLT1 = (-383.8) = 31.2

49 RAWDIF / ns Time Deviation / s 1.0E-07 RAWDIF in NIMT E C E C1, median = ns MJD 0.1ns 1.0E E E E E E+06 Averaging Time / s Visited receiver: NMIA Topcon/Javad Euro-80 (code: MTTO) The stability cannot reach 0.1 ns due to the salt-and-pepper noise

50 INT DLY ns ns INT DLY ns 31.2 ns Calibration Trip TRVL MTTO TLT1 UTC(NIMT) TRVL LS2P UTC(TL) VM12 UTC(NMLS) *Results for C1 code only UTC(VMI)

51 Check the short baseline

52 Check the common clock LSM1 TRVL LS2P TRVL SDI PD-5 10 MHz 10 MHz 1PPS 1PPS 10 MHz 10 MHz 1PPS 1PPS UTC (NMLS) UTC (NMLS) SDI HPDA 15RM-B 10 MHz Caesium 5071A (High Performance) SDI HPDA 15RM-B 10 MHz Caesium 5071A (High Performance)

53 CGGTTS GENERIC DATA FORMAT VERSION = 2E REV DATE = RCVR = Septentrio PolaRx2e TR CH = 20 IMS = LAB = NMIM X = m Y = m Z = m FRAME = ITRF COMMENTS = NO COMMENTS INT DLY = 0.0 ns (GPS C1), 0.0 ns (GPS P1), 0.0 ns (GPS P2) CAL_ID = xxxx-2017 CAB DLY = ns REF DLY = 26.9 ns REF = UTC(NMLS) To be updated CKSUM = FF SAT CL MJD STTIME TRKL ELV AZTH REFSV SRSV REFSYS SRSYS DSG IOE MDTR SMDT MDIO SMDI MSIO SMSI ISG FR HC FRC CK hhmmss s.1dg.1dg.1ns.1ps/s.1ns.1ps/s.1ns.1ns.1ps/s.1ns.1ps/s.1ns.1ps/s.1ns G10 FF L1C FF G10 FF L1P FF G10 FF L2P FF G10 FF L3P FF CGGTTS GENERIC DATA FORMAT VERSION = 2E REV DATE = RCVR = NMIA Topcon/Javad Euro-80 CH = 12 IMS = LAB = NMIM X = m Y = m Z = m FRAME = ITRF COMMENTS = NO COMMENTS INT DLY = ns (GPS C1), ns (GPS P1), ns (GPS P2) CAB DLY = 0.0 ns REF DLY = 0.0 ns REF = UTC(NMLS) CKSUM = FF CAL_ID = xxxx-2017 SAT CL MJD STTIME TRKL ELV AZTH REFSV SRSV REFSYS SRSYS DSG IOE MDTR SMDT MDIO SMDI MSIO SMSI ISG FR HC FRC CK hhmmss s.1dg.1dg.1ns.1ps/s.1ns.1ps/s.1ns.1ns.1ps/s.1ns.1ps/s.1ns.1ps/s.1ns G10 FF L1C FF G10 FF L1P FF G10 FF L2P FF G10 FF L3P FF

54 Calibration Using CGGTTS for LS2P Compute RAW for each station, each PRN, and each time stamp, RAW = REFSV + MDIO + MDTR + INTDLY + CABDLY REFDLY Use INT, CAB, REF values in CGGTTS data For C1, RAW LS2P = = For C1, RAW TRVL = = Compute RAWDIF TRVL-LSP2 = median(raw TRVL RAW LSP2 )= 55.9 Apply new reference delay to LS2P. New REFDLY: Compute ΔSYSDLY TRVL-LS2P = RAWDIF TRVL-LS2P + REFDLY TRVL REFDLY LS2P = Compute ΔSYSDLY LS2P-TLT1 = ΔSYSDLY TRVL-TLT1 ΔSYSDLY TRVL-LS2P = (-238.3) (-203.2) = Apply new cable delay to LS2P. New CABDLY: Compute Consider the same as before ΔINTDLY LS2P-TLT1 = ΔSYSDLY LS2P-TLT1 CABDLY LS2P + CABDLY TLT1 = We have INTDLY LS2P = INTDLY TLT1 + ΔINTDLY LS2P-TLT1 = (-192.6) = 222.4

55 RAWDIF / ns Time Deviation / s RAWDIF in NMIM 0 1.0E C1-6 C1, median = ns MJD 1.0E E E E E E+06 Averaging Time / s Visited receiver: NMIA Topcon/Javad Euro-80 (code: LSM1)

56 RAWDIF / ns Time Deviation / s 1.0E-08 RAWDIF in NMIM E C1, median = ns P1, median = ns P2, median = ns 1.0E C1 P1 P MJD 1.0E E E E E E+06 Averaging Time / s Visited receiver: Septentrio PolaRx2eTR (code: LS2P)

57 176.7 ns INT DLY ns ns ns 31.2 ns Calibration Trip MTTO TLT1 UTC(NIMT) TRVL LS2P TRVL VM12 UTC(NMLS) UTC(TL) UTC(VMI) *Results for C1 code only

58 Check the short baseline VM12

59 Check the common clock VM TRVL VM12 VN3P TRVL NV 142K HF-Multi Coupler NV 142K HF-Multi Coupler 10 MHz 10 MHz 6602 Pulse Distribution 6502 RF Distribution XSRM-Z Frequency Converter 6602 Pulse Distribution 6502 RF Distribution XSRM-Z Frequency Converter UTC(VMI) AOG 110 UTC(VMI) AOG PPS 5 MHz 5071A (option 001) Master Clock 1 PPS 5 MHz 5071A (option 001) Master Clock

60 CGGTTS GENERIC DATA FORMAT VERSION = 2E REV DATE = RCVR = Septentrio PolaRx3E TR CH = 99 IMS = LAB = VMI X = m Y = m Z = m FRAME = ITRF COMMENTS = NO COMMENTS INT DLY = 0.0 ns (GPS C1), 0.0 ns (GPS P1), 0.0 ns (GPS P2) CAL_ID = xxxx-2017 CAB DLY = ns REF DLY = ns REF = UTC(VMI) To be updated CKSUM = FF SAT CL MJD STTIME TRKL ELV AZTH REFSV SRSV REFSYS SRSYS DSG IOE MDTR SMDT MDIO SMDI MSIO SMSI ISG FR HC FRC CK hhmmss s.1dg.1dg.1ns.1ps/s.1ns.1ps/s.1ns.1ns.1ps/s.1ns.1ps/s.1ns.1ps/s.1ns G04 FF L1C FF G04 FF L1P FF G04 FF L2P FF G04 FF L3P FF CGGTTS GENERIC DATA FORMAT VERSION = 2E REV DATE = RCVR = NMIA Topcon/Javad Euro-80 CH = 12 IMS = LAB = VMI X = m Y = m Z = m FRAME = ITRF COMMENTS = NO COMMENTS INT DLY = ns (GPS C1), ns (GPS P1), ns (GPS P2) CAB DLY = 0.0 ns REF DLY = 19.0 ns REF = UTC(VMI) CKSUM = FF CAL_ID = xxxx-2017 SAT CL MJD STTIME TRKL ELV AZTH REFSV SRSV REFSYS SRSYS DSG IOE MDTR SMDT MDIO SMDI MSIO SMSI ISG FR HC FRC CK hhmmss s.1dg.1dg.1ns.1ps/s.1ns.1ps/s.1ns.1ns.1ps/s.1ns.1ps/s.1ns.1ps/s.1ns G04 FF L1C FF G04 FF L1P FF G04 FF L2P FF G04 FF L3P FF

61 Calibration Using CGGTTS for VM12 Compute RAW for each station, each PRN, and each time stamp, RAW = REFSV + MDIO + MDTR + INTDLY + CABDLY REFDLY Use INT, CAB, REF values in CGGTTS data For C1, RAW VM12 = = For C1, RAW TRVL = = Compute RAWDIF TRVL-VM12 = median(raw TRVL RAW VM12 )= Apply new reference delays. New REFDLY: for VM12 and 19.0 for TRVL Compute ΔSYSDLY TRVL-VM12 = RAWDIF TRVL-VM12 + REFDLY TRVL REFDLY VM12 = 3.5 Compute ΔSYSDLY VM12-TLT1 = ΔSYSDLY TRVL-TLT1 ΔSYSDLY TRVL-VM12 = (-238.3) (3.5) = Apply new cable delay to VM12. New CABDLY: Compute Consider the same as before ΔINTDLY VM12-TLT1 = ΔSYSDLY VM12-TLT1 CABDLY VM12 + CABDLY TLT1 = We have INTDLY VM12 = INTDLY TLT1 + ΔINTDLY VM12-TLT1 = (-366.1) = 48.9

62 RAWDIF / ns Time Deviation / s 1.0E-07 RAWDIF in VMI E C E C1, median = ns MJD 1.0E E E E E E+06 Averaging Time / s Visited receiver: NMIA Topcon/Javad Euro-80 (code: VM )

63 RAWDIF / ns Time Deviation / s RAWDIF in VMI E C1, median = ns P1, median = ns P2, median = ns 1.0E C1 P1 P MJD 1.0E E E E E E+06 Averaging Time / s Visited receiver: Septentrio PolaRx3eTR (code: VM12)

64 176.7 ns 48.9 ns ns ns ns ns 31.2 ns Calibration Trip MTTO TLT1 TRVL UTC(NIMT) LS2P UTC(TL) TRVL VM12 UTC(NMLS) UTC(VMI) Step3: Additional uncertainty due to the change of INTDLY of 0.1 ns, 0.7 ns and 1.1 ns for C1, P1 and P2

65 RAWDIF / ns RAWDIF / ns RAWDIF in TL, closure C1, median = ns P1, median = ns P2, median = ns Misclosure of P2: 1.1 ns C1, median = ns P1, median = ns P2, median = ns MJD MJD The variation of RAWDIF (misclosure) is the largest components in the ub

66 Time Deviation / s RAWDIF, Topcon receivers 1.0E E E ns MTTO LSM1 VM 1.0E E E E E E+06 Averaging Time / s TDEV of MTTO cannot down to 0.1 ns even we measured for 30 days Therefore, MTTO dominates the ua

67 Uncertainty Budget Unc. Value C1/P1 (ns) Table 6. Uncertainty contributions. Value P2 (ns) Value P1-P2 (ns) Value P3 (ns) Description u a (T V) RAWDIF (traveling visited) u a (T R) RAWDIF (traveling reference) u a Misclosure u b, observed misclosure Systematic components related to RAWDIF u b, Position error at reference u b, Position error at visited u b, Multipaths at reference u b, Multipaths at visited Link of the Traveling system to the local UTC(k) u b, REFDLY T (at ref lab) u b, REFDLY T (at visited lab) u b,tot Link of the Reference system to its local UTC(k) u b, REFDLY R (at ref lab) Link of the Visited system to its local UTC(k) u b, REFDLY V (at visited lab) u b,sys Components of equation (2) u CAL Composed of u a and u b,sys

68 Uncertainty Budget u b,1 : RAWDIF variation during the trip, 0.7 ns for C/A or P1, 1.1 ns for P2 u b,11 (u b,12 ): position error of the baselines, 0.1 ns u b,13 (u b,14 ): multipath, 0.3 ns u b,21 (u b,22 ): time error of UTC(k) PPS for TRVL 0.0 ns If PPS = UTC(k); 0.5 ns otherwise u b,31 (u b,32 ): time error of UTC(k) PPS, 0.5 ns However, BIPM used total 2.5 ns for G2 labs

69 Uncertainty Budget u b,1 : RAWDIF variation can be reduced if we shorten the calibration period u b,11 (u b,12 ): position error is very small u b,13 : multipath can be avoided with a big effort u b,21 (u b,22 ): time error of UTC(k) PPS for TRVL This can be eliminated if we connect UTC(k) to the TRVL directly u b,31 (u b,32 ): time error of UTC(k) PPS is difficult to avoid

70 Before and after Receiver Reference receiver Date C1 INTDLY (ns) New C1 INTDLY (ns) AU01 BP0N (BIPM ± calibration) MTTO AU ± LSM1 TRVL ± VM AU ± ± 4.5 * 16.2 Data provided from NMIA

71 RAWDIF / ns Some observations Some GPS receivers cannot be used for time transfer MSVP (a Topcon Euro-80 receiver) of NIMT show a sawtooth in RAWDIF (or CCD) plot

72 Some observations If the reference station can only obtain C/A data, a larger uncertainty would be given For Septentrio PolaRx2e TR, the start time of the receiver measurement depends on the firmware version Schedule always changes Unfair to the last lab

73 Some observations Misclosure is the largest component of uncertainties A shorter traveling time may avoid this Door-to-door shipment is time-consuming Import procedure / license The biggest variation dominate the uncertainty for all labs Time interval counter of TRVL did not work well in NMIM, but it becomes normal

74 Welcome to use our calibration service One lab per trip: TL lab TL Duties and rights are clear for the lab To minimize the aging of the traveling equipment To shorten the calibration period The lab needs to pay the shipment and insurance The lab must be responsive If any problems, let us know as soon as possible Thank you for your attention