Hi Geoff,

Did you and I run into the same issue with NMEA from old Jupiter
receivers? (many years ago at the now dead forum at gpskit.nl)

--

Björn

Björn wrote:

We may have Björn, I know we have communicated before, but I'm not
sure on what topic(s).

I have never owned or played with a Jupiter - so I have never made any
tests on one myself.

I have only run tests (NMEA w.r.t. PPS) on Trimble SV6, Motorola GT+,
Garmin 18 and 18x , Ashtech (Thales) AC12.

Geoff wrote:

David Taylor wrote:

I have only run extensive tests on Version 3.00 (for the 18x LVC) so I
know I can trust that version with respect the NMEA completing within
the second and stability of 1PPS.

I admit I just "assumed" subsequent versions were OK (without testing
myself) - so I got caught!

When / if Garmin fix this issue with the 18x, I will run thorough
tests next time before I commit to using  a new version of firmware.

Regards, Geoff (Christchurch, New Zealand)

Thanks, Geoff.  I have just look for that version on the Garmin Web site
but could not find it, at least by using the URL:

http://www8.garmin.com/software/GPS18xPC_LVC_300.exe
http://www8.garmin.com/software/GPS18xPC_LVC_30.exe
http://www8.garmin.com/software/GPS18xPC_LVC_3.exe

instead of:

http://www8.garmin.com/software/GPS18xPC_LVC_350.exe

I have yet to hear back from Garmin UK.

Cheers,
David

SatSignal software - quality software written to your requirements
Web:  http://www.satsignal.eu
Email:  david-taylor@blueyonder.co.uk

David Taylor wrote:

David, a quick Google Search found this:

http://www.gawisp.com/perry/oem_sensor/

and it is an exact file match for the file I downloaded from Garmin (Feb 2009).

Regards, Geoff (Christchurch, New Zealand)

Your Googling is better than mine!  What a very helpful Mr. Geoff and Mr.
Perry!  I've bookmarked the site for the next time - I did already have
something similar but the site had moved.

Cheers,
David

SatSignal software - quality software written to your requirements
Web:  http://www.satsignal.eu
Email:  david-taylor@blueyonder.co.uk

So, what is the final consensus as to which is the best firmware
version to use for NTP purposes?

Joe Gray
W5JG

On Thu, Dec 30, 2010 at 7:11 AM, David J Taylor
david-taylor@blueyonder.co.uk wrote:

I don't know, but I trust y'all know about:

http://support.ntp.org/bin/view/Support/ConfiguringGarminRefclocks

H

Any, providing you don't use the NMEA data to get the nearest second, but
use a secondary reference source.  Bit of a problem for a stand-alone
system, though.

I certainly would /not/ use 3.50.

Cheers,
David

SatSignal software - quality software written to your requirements
Web:  http://www.satsignal.eu
Email:  david-taylor@blueyonder.co.uk

If I am using the NMEA data (as in the Garmin is the only time
source), what firmware is recommended by all those using a GPS18x-LVC?

Joe Gray
W5JG

On Fri, Dec 31, 2010 at 1:52 AM, David J Taylor
david-taylor@blueyonder.co.uk wrote:

Kiwi Geoff recommended:

http://www.gawisp.com/perry/oem_sensor/

I would go back to this version, were it not for the act that I had a
GPS18x LVC which stopped working, and needed to be replaced by Garmin.
There was a suggestion at the time that it was a firmware problem.  Here's
the list of firmware changes:

http://www8.garmin.com/support/download_details.jsp?id=4055

I wonder about the 3.10 update "Improved robustness of internal error
handling".

73,
David

SatSignal software - quality software written to your requirements
Web:  http://www.satsignal.eu
Email:  david-taylor@blueyonder.co.uk

On 12/31/2010 10:20 AM, Joseph Gray wrote:

Here are measurements on the bad versions:

http://n1.taur.dk/permanent/gps18x-330-pps.jpg
http://n1.taur.dk/permanent/gps18x-350-pps.jpg
Note I am running 115200, not 4800

3.00 kills itself. I had 2 die simultaneously.
3.10 starts transmitting 350ms .. 480ms after PPS
3.30 starts transmitting 920ms .. 1220ms (!) after PPS
3.50 starts transmitting 700ms .. 1220ms (!) after PPS

/Kasper Pedersen

Joseph Gray wrote:

I thought I would run some tests for you Joe, using Version 3.0 and 3.2

Although I myself have run version 3.0 for a number of months (when it
first came out in 2009) without issues, I take note of what Kasper
Pedersen wrote about his experience  - so I have now loaded version
3.2 into my 18x, rather than any newer version (until Garmin fix the
NMEA latency issue).

I ran my tests for 2 hours each (7,200 one second samples) with GPS
sitting on an inside window sill with a very limited view of the sky,
so my readings should be taken as perhaps worse case rather than from
a perfect view of the sky. The GPS was using 9 - 10 satellites for the
fix solution during the test.

The RF sensitivity of the 18x is much better than the Garmin 18.

I measured the time of the "START" character ($) of the $GPRMC
sentence and  the "LAST" 0x10 (Linefeed) character, GPS 18x set to
4800 Baud.

Following times are in milliseconds from the leading edge of the 1PPS marker.

Version 3.00 firmware for Garmin 18x
START Min =  365
START Avg =  427 (N = 7200)
START Max =  499
LAST Min =  513
LAST Avg =  576 (N = 7200)
LAST Max =  648

Version 3.20 firmware for Garmin 18x
START Min =  370
START Avg =  433  (N = 7200)
START Max =  518
LAST Min =  519
LAST Avg =  582  (N = 7200)
LAST Max=  667

So if you use Version 3.20 firmware for the 18x, at 4800 baud the NMEA
RMC sentence will give the time of the current UTC second and will be
received by an average of 582 ms after the leading edge of the 1PPS
transition (with a maximum jitter of plus 85 ms / minus 63 ms).

Kasper Pedersen has done, I think we can say that currently version
3.2 is the "best" version if you want to use the 18x NMEA timestamps
to be given early in the second (rather than much later as per version
3.3 and 3.5).

Joe, you can download    GPS18xPC_LVC_320.exe    from here:

http://www.gawisp.com/perry/oem_sensor/

The above file is identical to the one I downloaded in 2009 from Garmin.

Hope that helps.

Regards, Geoff (Christchurch, New Zealand).

The orb in the sky made an appearance today, so I figured I'd photograph
the guts of the Starloc II. The HAG-240 is an active antenna a local
surplus shop was selling. Since it worked with the box, I went back and
wiped them out. $15 each, plus $2 for the threaded pipe to mount it.

Geoff,

Thanks a lot for comparing firmwares. That hard information helps tremendously.

Joe Gray
W5JG

On Tue, Jan 4, 2011 at 8:46 PM, Kiwi Geoff geoff36@gmail.com wrote:

I agree, very helpful, thanks Joseph!

I recently got two 18x that I upgraded to 3.50. (I'm starting to work on
integrating some M12+T modules to supplement/replace the 18x's).

Has anyone had trouble with reverting firmware upgrades on Garmin products?
Of course, feedback on going back from 3.50 to 3.20 on an 18x would be most
specific to my situation.

Thanks in advance!

Kevin Rosenberg
KR5F

On Jan 4, 2011, at 9:29 PM, Joseph Gray wrote:

On Jan 5, 2011, at 1:15 AM, David J Taylor wrote:

David, that's a fine idea. Do you have an email address (or other contact info)
where our reports of this issue have a greater chance of getting to the firmware
group compared to being forwarded to a write-once, read-never file?

Kevin

Kevin,

I only have a generic e-mail address:
product [period] support  {at-sign}  garmin  [dot] com

but perhaps by putting the case number in the subject and body of the text
it might get added to the case:  KMM19596806I15977L0KM  I have written to
the Garmin Europe person that I am happy to talk directly to the
developers and to try beta firmware, but I don't hold out much hope of
direct contact!

Cheers,
David

SatSignal software - quality software written to your requirements
Web:  http://www.satsignal.eu
Email:  david-taylor@blueyonder.co.uk

Hal Murray wrote:

I have run some more tests Hal, with V3.5, averaging ON and OFF.

I include V3.2 for comparison.

I measured the time of the "START" character ($) of the $GPRMC
sentence and  the "LAST" 0x10 (Linefeed) character, relative to the
leading transition of the 1PPS, GPS 18x set to 4800 Baud.

Version 3.20 firmware for Garmin 18x
START Min =  370
START Avg =  433  (N = 7200)
START Max =  518
LAST Min =  519
LAST Avg =  582  (N = 7200)
LAST Max =  667

Version 3.50 firmware for Garmin 18x  with Averaging ON
START Min =  768
START Avg =  813  (N = 4194)
START Max =  871
LAST Min =  917
LAST Avg =  962  (N = 4194)
LAST Max =  1020  ( 7 sentences finished over 1,000 ms)

Version 3.50 firmware for Garmin 18x  with Averaging OFF
START Min =  769
START Avg =  805  (N = 3563)
START Max =  858
LAST Min =  918
LAST Avg =  954  (N = 3563)
LAST Max =  1007 ( 4 sentences finished over 1,000 ms)

This agrees with what Kasper Pedersen wrote earlier:

The "averaging" appears to have about an 8 ms penalty in delaying the
NMEA data stream.

We can see in the above, that V3.5 RMC is infrequently not finishing
"within" the correct second !

For "my" application I need both RMC and GGA to finish within the
current second - which the Garmin 18x does very comfortably with
version 3.2 and under.

Kevin Rosenberg wrote:

I have "reverted" quite a number of times now Kevin - trying to
understand why the 18x can be 1 second slow. I have never had an issue
putting in an older version into the 18x, the updater software does
know what your doing, it says:

"Your current software version is newer than the update version.

Do you want to replace your current software version with an older version ? "

We click the YES box.

To do the above timing tests I had v3.5 loaded into my 18x, I then
reverted to v3.2 and everything worked fine Kevin (to answer your
specific question).

Regards, Kiwi Geoff (New Zealand).

On Jan 5, 2011, at 4:08 AM, Kiwi Geoff wrote:

Hi Geoff (of Kiwiland),

Thanks for the information and testing. Having had my 3.50 18x and
NTP report that NMEA time is -1.00 seconds compared to PPS time,
I'll give 3.20 a whirl.

I've certainly bricked my share of devices by attempting to
"reverse the timeline" of firmware images. Manufacturer's have
cavalierly deemed  such attempts to install an older firmware image
as being "unsupported". I tend to think of it more of as programmers,
for reasons as diverse as lack of time, talent, technology, or
imagination, who create insufficiently robust installers.

It's very nice of you to actually test going from 3.50 to 3.20 for
the rest of us.

Kevin (of New Mexicoland)