This site is no longer maintained. You will be redirected in

seconds
 

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[IGSMAIL-5902]: Final 30s clock products modified



******************************************************************************
IGS Electronic Mail      19 Feb 05:40:53 PST 2009      Message Number 5902
******************************************************************************

Author: Jim Ray & Jake Griffiths

Starting with today's Final clock products for wk 1517, the
combination procedure for the 30s satellite clocks provided
in the files igsWWWWd.clk_30s has been modified slightly.

Recall that 6 Analysis Centers contribute clock products but
only 3 of them provide 30s satellite clocks.  As a consequence
the AC clock biases average differently for the 5m and 30s
epochs.  (For information, the average AC biases for each
clock are tabulated in the 2nd table of each daily .cls file.)
This can cause relative time-varying bias differences between
the 5m and 30s combined clocks at the level of roughly 30 ps
(1 cm).  This is well below the true uncertainty in the IGS
clock products and has no effect whatsoever on positioning
applications using these products.  However, the effect is
systematic and can affect one-way time transfer applications
over intervals <5m at the <100 ps level.

In the new procedure starting today, the bias differences
between the 5m and 30s combinations are determined at each
5m epoch and removed from the 30s values by interpolation.
In this way the clock values reported at the 5m epochs (which
are also used to form the IGS ensemble timescale) are
unchanged.  The combined 5m and 30s clocks should now attain
the same level of internal consistency provided by the 30s
ACs themselves.

We thank Sylvain Loyer, Jan Kouba, and their colleagues for
reporting this matter to us and for helping test our new
procedure.