[time-nuts] Garmin 18X LVC UTC issue

Kiwi Geoff geoff36 at gmail.com
Thu Mar 26 19:27:50 EDT 2009


The Garmin 18 LVC is a useful low cost way to get 1PPS and the UTC time.

The new replacement of the 18 is the 18X, which has an amazing 20 dB
in RF sensitivity over the 18, which makes indoor experiments a lot
easier.

However one issue that I recently uncovered is worth mentioning -
especially to Time Nuts.

If one gives the 18X the command "Set Autolocate" (which intitiates a
cold start) the 18X  (after it gets a FIX) starts transmitting the UTC
time one second FAST. It is not until the 18X has downloaded a current
almanac from the satellites, does it create a LEAP SECOND in its UTC
time stream, and so come into line with current UTC time - this can
take at least 12 minutes after it has obtained a FIX to achieve. So
every time the "Set Autolocate" command is issued with an 18X, the GPS
is one second FAST until it creates its own LEAP SECOND. This is with
the latest firmware (3.00).

Here is an experiment I did a few days ago, to show what happens to
the NMEA data stream. I have edited the logged data to make it easy to
see what happens.

I firstly issued a "Set Autolocate" command to the Garmin 18X, we can
see the GPS received my command because it repeated the command back.

$GPRMC,225137,A     (at this stage the UTC time is correct to my other clocks)
$GPGGA,225137,
$PGRMF,500,255112,240309,225137,15,  (15 second delta GPS - UTC)
$GPRMC,225138,A
$GPGGA,225138,
$PGRMF,500,255113,240309,225138,15,
$PGRMI,,,,,,,A                  (this is my echoed Set AutoLocate command)
$GPRMC,225138,A         (notice the UTC time is the same as previous second)
$GPGGA,225138
$PGRMF,500,255112,240309,225138,14   (14 second delta GPS - UTC)

Notice also the "GPS seconds of week" (255112) has jumped back 2
seconds, I assume this is the GPS using its own clock to get the
initial "approximate" time to start searching for satellites. This
timing error gets corrected as soon as the GPS gets a FIX.

44 seconds later the 18X GPS gets a position FIX

$PGRMF,500,255136,240309,225202,14,
$GPRMC,225203,A,
$GPGGA,225203,,1,00,1.4,6.3,M,9.6
$PGRMF,500,255137,240309,225203,14,
$GPRMC,225204,A,
$GPGGA,225204,,1,04,1.4,6.3,M,9.6   ( FIX with 4 satellites)
$PGRMF,500,255138,240309,225204,14,
$GPRMC,225206,A,                 (GPS time now jumps 1 second with actual FIX )
$GPGGA,225206,,1,05,2.9,0.2,M,9.6,M,,*6B
$PGRMF,500,255140,240309,225206,14,     (UTC still 1 second FAST )

Approximately seven minutes later, the 18X had downloaded a new
almanac and so we see the UTC time changing in the various sentences.

$GPGGA,225839,
$PGRMF,500,255533,240309,225839,14,
$GPRMC,225840,A,
$GPGGA,225840,
$PGRMF,500,255534,240309,225840,14,
$GPRMC,225840,A    (LEAP SECOND - time stamp same as previous second)
$GPGGA,225840,
$PGRMF,500,255535,240309,225840,15,   (GPS - UTC now 15 second delta)
$GPRMC,225841,A,
$GPGGA,225841,
$PGRMF,500,255536,240309,225841,15,
$GPRMC,225842,A,
$GPGGA,225842,

Notice in the above sequence, the GPS time remains uniform (255533 to 255536),
it is only the "UTC time" that is shifted back one second.

At this stage, the 18X GPS UTC time, is correct to my other precision clocks.

It is interesting to note, the older Garmin 18, does not respond at
all to the "Set Autolocate" command.

Another issue I have found with the 18X:

I tried to download an almanac from the 18x sensor using the command

$PGRMO,GPALM,1

and it doesn't work with the latest firmware version (3.00) of the
18X, yet it works OK with an "18" sensor.

I was hoping to find the Toa (Time of Applicability) of the original
almanac held in firmware, but the 18X sensor wont provide the almanac
(although the manual / data sheet for the 18X says the command can be
used).

Regards, Kiwi Geoff.
http://www.geocities.com/kiwi_36_nz




More information about the time-nuts mailing list