From kouba at geod.nrcan.gc.ca Fri May 5 13:43:05 2000
From: kouba at geod.nrcan.gc.ca (Jan Kouba)
Date: Fri, 05 May 2000 20:43:05 GMT
Subject: [IGSMAIL-2824]: IGS orbit/clock navigation & sat. clock interpolations with no SA
Message-ID: <200005052043.UAA25332@macs.geod.nrcan.gc.ca>
******************************************************************************
IGS Electronic Mail 05 May 13:43:09 PDT 2000 Message Number 2824
******************************************************************************
Author: J. Kouba
Subject: IGS Orbit/clock navigation & sat. clock interpolations with no SA
Dear Colleagues,
The removal of SA (as announced on May 2 (Wk 1060/day 2) at about 04:00 UTC
is expected to have a significant impact on precise point navigation with the
IGS orbit/clocks held fixed,as well as satellite clock and data interpolations.
With no SA it should now be possible to interpolate the IGS satellite clocks,
currently sampled at 5 min intervals, at or below the 20 cm precision level.
Consequently,it should now be possible to do precise navigation at any interval
(instant) at this (20 cm) precision level with IGS orbits/clocks fixed. The
navigation precision of better than 10 cm with IGS orbits/clocks fixed is being
demonstrated daily and weekly in the summaries of the IGS Rapid(IGR) and Final
combinations, respectively (at the 15 min sampling intervals of sp3
orbit files).
For your information, the first table below contains the expected clock errors
based on individual frequency standard specifications (Cs, Rb and HM). The
second table contains the actual Allan std's (for 100 sec interval), computed
for the second day (Wk 1060/d3) of the SA-free IGR combined orbit/clock
solutions.The first day (1060/d2) gave almost identical results. All but one Cs
(PRN 6) are well below 5.e-12, i.e. they are closer to the high performance Cs
tube specifications (1.e-12/100). Only one Rb (PRN8) has exceeded the usual Rb
specs (1.e-12/100sec).An interesting observation is that most Rb performed more
like Cs (a random walk,i.e. showing the sqrt(t) dependency at least for up to 4
hour intervals!) The only exception is the already mentioned PRN 8 Rb which has
shown a time dependency about half the way between the sqrt(t) and t-
dependency.
The Table 1 should give an idea what to expect for clock (station/sat)
interpolation and/or extrapolation (due to frequency standard errors only). For
example for a 5 min clock interpolation, a conservative clock error estimate
caused by a straight line interpolation is less or equal to the half of the
interval, I.e. the 2.5 min (150 sec) value. This should be quadratically
increased by up to a clock noise sigma (about .1 ns (3 cm) or less for IGS
combined clocks). In practice one'd do a polynomial (at least 3 point
interpolation) so the interpolation errors should be even less. As mentioned
above, the Rb satellites should be treated as Cs in Table 1, so that
interpolations within a 5 min interval the frequency standard errors should be
within 20 cm for all but one Cs' and below 3.6 cm for all but one Rb's
standards. For extrapolation use the full interval, eg. for a 60 min
extrapolation use 60 min (3600 sec) etc. HM's specs are included here for
comparisons and due to the fact that a number of IGS stations are equipped
with receivers that use unaltered (so far) external HM frequency standard
signals with the HM's of much better specs than the one shown in the Table 1.
Hope this helps
J. Kouba
PS. I've (re)used the following (old) references related to freq. standards:
1. Fell. P.J., 1980, "Geodetic Positioning using GPS Satellites, OSU- DGS
Report #299, June.
2. Kouba, J. and J.A. Orosz, 1982, Ecaluation of ELECTRAC receivers and
Oscillator effects on Doppler Data Quality .., Proc. 3rd Inter. Symp. on
Satell. Pos, Las Cruces, NM, pp. 937-952.
3. Vessot, R.F.C, 1976, Frequency and Standards, Method of Experimental
Physics, Vol. 12, Part C, Radio Observations (ed. M.L. Meeks), Academic Press
pp 198-227.
p
----------
Table 1
Inter/Fq Std. Primary Cs Super Tube Cs Standard Cs Rb HM
Specs/100s 1.E-13 1.E-12 1.E-11 1.E-12 1.E-14
1 sec .3 mm 3 mm 30 mm 3 mm .003 mm
30 1.6 16 160 16 .003
150 3.6 36 360 45 .2
300 5.2 52 520 90 .6
600 7.3 73 730 180 1.5
900 9.0 90 900 270 2.5
1800 12.7 127 1270 540 5.1
3600 18.0 180 1800 1080 10.5
7200 25.4 254 2546 2160 21.3
14400 36.0 360 3600 4320 42.9
Table 2: CURRENT BLOCK II/IIA/IIR SATELLITES
==================================
LAUNCH LAUNCH FREQ US SPACE
ORDER PRN SVN DATE STD PLANE COMMAND Allan/100s
----------------------------------------- ------------
*II-1 14 14 14-Feb-89 19802
II-2 2 13 10-Jun-89 Cs B3 20061 4.0E-12
II-3 16 16 18-Aug-89 Rb E5 20185 9.6E-13
II-4 19 19 21-Oct-89 Cs A5 20302 3.7E-12
II-5 17 17 11-Dec-89 Cs D3 20361
II-6 18 18 24-Jan-90 Cs F3 20452 3.5E-12
*II-7 20 20 26-Mar-90 20533
II-8 21 21 2-Aug-90 Cs E2 20724 2.7E-12
II-9 15 15 1-Oct-90 Cs D5 20830 3.6E-12
IIA-10 23 23 26-Nov-90 Cs E4 20959 4.1E-12
IIA-11 24 24 4-Jul-91 Rb D1 21552 6.6E-13
IIA-12 25 25 23-Feb-92 Cs A2 21890 2.8E-12
*IIA-13 28 28 10-Apr-92 21930 8.3E-13
IIA-14 26 26 7-Jul-92 Rb F2 22014 7.8E-13
IIA-15 27 27 9-Sep-92 Cs A4 22108 4.9E-12
IIA-16 1 32 22-Nov-92 Cs F4 22231 2.8E-12
IIA-17 29 29 18-Dec-92 Rb F1 22275 8.3E-13
IIA-18 22 22 3-Feb-93 Rb B1 22446 4.7E-13
IIA-19 31 31 30-Mar-93 Cs C3 22581 4.2E-12
IIA-20 7 37 13-May-93 Rb C4 22657 7.0E-13
IIA-21 9 39 26-Jun-93 Cs A1 22700 2.9E-12
IIA-22 5 35 30-Aug-93 Cs B4 22779 3.8E-12
IIA-23 4 34 26-Oct-93 Rb D4 22877 4.9E-13
IIA-24 6 36 10-Mar-94 Cs C1 23027 8.1E-12
IIA-25 3 33 28-Mar-96 Cs C2 23833 3.9E-12
IIA-26 10 40 16-Jul-96 Cs E3 23953 2.9E-12
IIA-27 30 30 12-Sep-96 Cs B2 24320 1.6E-12
IIA-28 8 38 6-Nov-97 Rb A3 25030 3.0E-12
***IIR-1 42 17-Jan-97
IIR-2 13 43 23-Jul-97 Rb F5 24876 9.3E-13
IIR-3 11 46 7-Oct-99 Rb D2 25933 3.7E-13
* Satellite is no longer in service.
** US SPACE COMMAND, previously known as the NORAD object number;
also referred to as the NASA Catalog number. Assigned at successful
launch.
*** Unsuccessful launch.