[IGSMAIL-2190] TurboRogue L2 tracking update

[J]

******************************************************************************
IGS Electronic Mail      Thu Mar  4 17:12:16 PST 1999      Message Number 2190
******************************************************************************

Author: J F Zumberge, H J Kunze
Subject: TurboRogue L2 tracking update

Over the last three months some progress has been made on what's causing
the loss of L2 tracking in a TurboRogue (TR) receiver running at 30 sec.
(For reference, see IGSmail numbers 2071 and 2075.)

The symptom is that, when running at 30 sec in cross correlation mode
(normal for TRs when AS is on), tracking on L2 stops when the value of
P2 - P1 (actually, P2 - C/A) exceeds 12 m.  (A related symptom is that,
for 8 m < P2 - C/A < 12 m, the value of P2 - C/A is quantized.)

Because P2 - C/A is essentially a measure of the ionosphere, these
symptoms are most evident in low-latitude sites.  ASC1 (Ascension
Island, about -8 deg latitude) exhibits these symptoms in the extreme,
and MKEA (Mauna Kea, about +20 deg latitude) to a much lesser, but still
noticeable, extent.

When running at a 1-sec rate, the TR does _not_ exhibit the symptoms.

Limited resources have been available to examine the TurboRogue receiver
code.  Some of the symptoms are understood, and some very recent
findings look promising.  In fact, this week we have begun tests at MKEA
with modified firmware.  At this point we are cautiously optimistic
because (i) the quantization symptom has gone away and (ii) the number
of low L2 SNR events has decreased significantly (compared, for example,
to nearby KOKB, where these events continue to occur in large numbers
near local noon).

We will be looking at precise analysis of the 1-day data over the next
several days.  An initial test with the new firmware in a receiver at
JPL showed good results from such data.  Nevertheless, we need several
additional days of testing at MKEA before we are confident that we've
found a firmware mod to solve this problem.  At this point we expect to
have a firmware upgrade that solves the L2 tracking problem along with
the GPS week-1024 and Y2K rollover problems.

Independently, we have run the receiver at Mauna Kea at 1 sec during
much of January, and developed a perl script to mimic the TR's formation
of a 30-sec data set.  Analyses of these show performance similar to
what is typically achieved with normal TR IGS data.  We have also run
the ASC1 receiver at 1-sec data interval for two days, and are looking
at those results now.  In the event that the firmware mod does not solve
the problem, we expect to be in a position to operate some of our
low-latitude TR receivers at the 1-sec data rate.  An external PC at the
site will periodically offload the data and form the standard 30-sec IGS
product.

Of course, we should recognize that the TRs represent mature technology.
Operators of IGS sites with TRs should consider replacing them with more
modern technology, as budgets permit.

Sincerely,

J F Zumberge, Jet Propulsion Laboratory
H J Kunze,    Allen Osborne Associates Inc.


[Mailed From: jfz at cobra.jpl.nasa.gov]