[Starlink] GPS Dongle?

[Starlink] GPS Dongle?

[Nathan Owens]

[-- Attachment #1: Type: text/plain, Size: 103 bytes --]

Does anyone know if any MAC1-X GPS dongles or anything similar are
available anywhere?

Thanks,
Nathan

[-- Attachment #2: Type: text/html, Size: 174 bytes --]

Re: [Starlink] GPS Dongle?

[Nathan Owens]

[-- Attachment #1: Type: text/plain, Size: 490 bytes --]

Sorry, it was MACx-1, but either way, I decided to go for an Adafruit
Raspberry Pi GPS "Hat", which does 1PPS. I'll setup a raspberry pi with
RT_PREEMPT and run as a Stratum-1 Clock locally. This should help with
better irtt timings. This can be done for <$100 if anyone else wants super
accurate time.

On Mon, Jun 14, 2021 at 7:55 AM Nathan Owens <nathan@nathan.io> wrote:

> Does anyone know if any MAC1-X GPS dongles or anything similar are
> available anywhere?
>
> Thanks,
> Nathan
>

[-- Attachment #2: Type: text/html, Size: 849 bytes --]

Re: [Starlink] GPS Dongle?

[Gary E. Miller]

[-- Attachment #1: Type: text/plain, Size: 1433 bytes --]

Yo Nathan!

On Mon, 14 Jun 2021 10:40:36 -0700
Nathan Owens <nathan@nathan.io> wrote:

> Sorry, it was MACx-1, but either way, I decided to go for an Adafruit
> Raspberry Pi GPS "Hat", which does 1PPS.

I do not recommend that HAT.  There are much better ones available at
similar prices.  

> I'll setup a raspberry pi
> with RT_PREEMPT

RT_PREEMPT is overkill.  The best you can get on a RasPi is about 54ns
accuracy.  Due to the clock granularity of 54ns.

> and run as a Stratum-1 Clock locally.

But always have other chimers in your configuration file.

> This should
> help with better irtt timings. This can be done for <$100 if anyone
> else wants super accurate time.

For some definition of "super".

> > Does anyone know if any MAC1-X GPS dongles or anything similar are
> > available anywhere?

They pop up now and then.  Right now GPS parts are hard to find, much
less at good prices.  Search for Navisys GR-601W, GR-701W, or GR-801W.
The "W" suffix says it outputs PPS.

The MAC1-X, or any PPS over USB, will only get you to > 1 milli second
because of the USB.

RGDS
GARY
---------------------------------------------------------------------------
Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
	gem@rellim.com  Tel:+1 541 382 8588

	    Veritas liberabit vos. -- Quid est veritas?
    "If you can't measure it, you can't improve it." - Lord Kelvin

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 851 bytes --]

Re: [Starlink] GPS Dongle?

[Nathan Owens]

[-- Attachment #1: Type: text/plain, Size: 1911 bytes --]

Thanks for the info Gary!

Why wouldn't you recommend the hat? Anything you'd recommend that is
actually purchasable? I can't seem to find any of the Navisys units, and
ideally non-USB would be better?

On Mon, Jun 14, 2021 at 12:07 PM Gary E. Miller <gem@rellim.com> wrote:

> Yo Nathan!
>
> On Mon, 14 Jun 2021 10:40:36 -0700
> Nathan Owens <nathan@nathan.io> wrote:
>
> > Sorry, it was MACx-1, but either way, I decided to go for an Adafruit
> > Raspberry Pi GPS "Hat", which does 1PPS.
>
> I do not recommend that HAT.  There are much better ones available at
> similar prices.
>
> > I'll setup a raspberry pi
> > with RT_PREEMPT
>
> RT_PREEMPT is overkill.  The best you can get on a RasPi is about 54ns
> accuracy.  Due to the clock granularity of 54ns.
>
> > and run as a Stratum-1 Clock locally.
>
> But always have other chimers in your configuration file.
>
> > This should
> > help with better irtt timings. This can be done for <$100 if anyone
> > else wants super accurate time.
>
> For some definition of "super".
>
> > > Does anyone know if any MAC1-X GPS dongles or anything similar are
> > > available anywhere?
>
> They pop up now and then.  Right now GPS parts are hard to find, much
> less at good prices.  Search for Navisys GR-601W, GR-701W, or GR-801W.
> The "W" suffix says it outputs PPS.
>
> The MAC1-X, or any PPS over USB, will only get you to > 1 milli second
> because of the USB.
>
> RGDS
> GARY
> ---------------------------------------------------------------------------
> Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
>         gem@rellim.com  Tel:+1 541 382 8588
>
>             Veritas liberabit vos. -- Quid est veritas?
>     "If you can't measure it, you can't improve it." - Lord Kelvin
> _______________________________________________
> Starlink mailing list
> Starlink@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/starlink
>

[-- Attachment #2: Type: text/html, Size: 2779 bytes --]

Re: [Starlink] GPS Dongle?

[Nathan Owens]

[-- Attachment #1: Type: text/plain, Size: 2200 bytes --]

I am able to buy the same chip, the uBlox M8Q as hat:
https://www.amazon.com/MAX-M8Q-Raspberry-Multi-Constellation-Receiver-Support/dp/B08HJFYPGX/

On Mon, Jun 14, 2021 at 12:09 PM Nathan Owens <nathan@nathan.io> wrote:

> Thanks for the info Gary!
>
> Why wouldn't you recommend the hat? Anything you'd recommend that is
> actually purchasable? I can't seem to find any of the Navisys units, and
> ideally non-USB would be better?
>
> On Mon, Jun 14, 2021 at 12:07 PM Gary E. Miller <gem@rellim.com> wrote:
>
>> Yo Nathan!
>>
>> On Mon, 14 Jun 2021 10:40:36 -0700
>> Nathan Owens <nathan@nathan.io> wrote:
>>
>> > Sorry, it was MACx-1, but either way, I decided to go for an Adafruit
>> > Raspberry Pi GPS "Hat", which does 1PPS.
>>
>> I do not recommend that HAT.  There are much better ones available at
>> similar prices.
>>
>> > I'll setup a raspberry pi
>> > with RT_PREEMPT
>>
>> RT_PREEMPT is overkill.  The best you can get on a RasPi is about 54ns
>> accuracy.  Due to the clock granularity of 54ns.
>>
>> > and run as a Stratum-1 Clock locally.
>>
>> But always have other chimers in your configuration file.
>>
>> > This should
>> > help with better irtt timings. This can be done for <$100 if anyone
>> > else wants super accurate time.
>>
>> For some definition of "super".
>>
>> > > Does anyone know if any MAC1-X GPS dongles or anything similar are
>> > > available anywhere?
>>
>> They pop up now and then.  Right now GPS parts are hard to find, much
>> less at good prices.  Search for Navisys GR-601W, GR-701W, or GR-801W.
>> The "W" suffix says it outputs PPS.
>>
>> The MAC1-X, or any PPS over USB, will only get you to > 1 milli second
>> because of the USB.
>>
>> RGDS
>> GARY
>>
>> ---------------------------------------------------------------------------
>> Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
>>         gem@rellim.com  Tel:+1 541 382 8588
>>
>>             Veritas liberabit vos. -- Quid est veritas?
>>     "If you can't measure it, you can't improve it." - Lord Kelvin
>> _______________________________________________
>> Starlink mailing list
>> Starlink@lists.bufferbloat.net
>> https://lists.bufferbloat.net/listinfo/starlink
>>
>

[-- Attachment #2: Type: text/html, Size: 3399 bytes --]

Re: [Starlink] GPS Dongle?

[Gary E. Miller]

[-- Attachment #1: Type: text/plain, Size: 1889 bytes --]

Yo Nathan!

On Mon, 14 Jun 2021 12:09:19 -0700
Nathan Owens <nathan@nathan.io> wrote:

> Why wouldn't you recommend the hat?

A good recommendation needs to know a bit more about your use case
and expectations.  

The Adafruit chip is several generations old (MT3333?), and the built-in
antenna is no good inside a building.  You want a newer chip that does
multi constellations, and has an antenna connector.  The Adafruit has a
u.fl conenctor, I much prefer SMA.

> Anything you'd recommend that is
> actually purchasable?

Uputronics makes good HATs:

https://store.uputronics.com/index.php?route=product/category&path=60_64

NEO-M8Q for 44 Pounds, ZED-F9P for 240 Pounds.  They have good support
and fast ship to USA.

Also common to get a GPS on a breakout board and jumper it to the Pi.

Sparkfun has a lot of models:

https://www.sparkfun.com/search/results?term=gps

Abay is often the besst place to look.  Here is an interesting new part:

https://www.ebay.com/itm/264810984552?_trkparms=aid%3D1110006%26algo%3DHOMESPLICE.SIM%26ao%3D1%26asc%3D232899%26meid%3Ded97b963bf6d4fe88b2164c0af008895%26pid%3D101113%26rk%3D4%26rkt%3D12%26sd%3D384222062283%26itm%3D264810984552%26pmt%3D0%26noa%3D1%26pg%3D2563228%26algv%3DDefaultOrganic&_trksid=p2563228.c101113.m2108

The MAX-M8Q is good, not great.

> I can't seem to find any of the Navisys units,
> and ideally non-USB would be better?

You have to call them direct.  Minimum order quantity is non-trivial.
YOu see them on etsy and ebay when somoeone buys a case of them.


RGDS
GARY
---------------------------------------------------------------------------
Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
	gem@rellim.com  Tel:+1 541 382 8588

	    Veritas liberabit vos. -- Quid est veritas?
    "If you can't measure it, you can't improve it." - Lord Kelvin

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 851 bytes --]

Re: [Starlink] GPS Dongle?

[Gary E. Miller]

[-- Attachment #1: Type: text/plain, Size: 1003 bytes --]

Yo Nathan!

On Mon, 14 Jun 2021 12:13:55 -0700
Nathan Owens <nathan@nathan.io> wrote:

> I am able to buy the same chip, the uBlox M8Q as hat:
> https://www.amazon.com/MAX-M8Q-Raspberry-Multi-Constellation-Receiver-Support/dp/B08HJFYPGX/

Yeah, I saw that.  Looks good, but no reports on it yet. u.fl is OK,
but a PITA.  The antenna does not specify it works with Galileo.  You
want at least 15MHz to 20 Mhz for Galileo E1.  A "GPS" antenna is often
not good enough.  A good antenna would be marked as good for Galileo.
GLONASS needs a much wider bandwidth, but timing people turn off the
GLONASS as unstable.

For the price you can buy a better antenna if required.

RGDS
GARY
---------------------------------------------------------------------------
Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
	gem@rellim.com  Tel:+1 541 382 8588

	    Veritas liberabit vos. -- Quid est veritas?
    "If you can't measure it, you can't improve it." - Lord Kelvin

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 851 bytes --]

Re: [Starlink] GPS Dongle?

[Ricky Mok]

[-- Attachment #1.1: Type: text/plain, Size: 1285 bytes --]

Attached a paper on this topic from my old colleagues.

Ricky

On 6/14/2021 1:18 PM, Gary E. Miller wrote:
> Yo Nathan!
>
> On Mon, 14 Jun 2021 12:13:55 -0700
> Nathan Owens <nathan@nathan.io> wrote:
>
>> I am able to buy the same chip, the uBlox M8Q as hat:
>> https://www.amazon.com/MAX-M8Q-Raspberry-Multi-Constellation-Receiver-Support/dp/B08HJFYPGX/
> Yeah, I saw that.  Looks good, but no reports on it yet. u.fl is OK,
> but a PITA.  The antenna does not specify it works with Galileo.  You
> want at least 15MHz to 20 Mhz for Galileo E1.  A "GPS" antenna is often
> not good enough.  A good antenna would be marked as good for Galileo.
> GLONASS needs a much wider bandwidth, but timing people turn off the
> GLONASS as unstable.
>
> For the price you can buy a better antenna if required.
>
> RGDS
> GARY
> ---------------------------------------------------------------------------
> Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
> 	gem@rellim.com  Tel:+1 541 382 8588
>
> 	    Veritas liberabit vos. -- Quid est veritas?
>      "If you can't measure it, you can't improve it." - Lord Kelvin
>
> _______________________________________________
> Starlink mailing list
> Starlink@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/starlink

[-- Attachment #1.2: Type: text/html, Size: 2388 bytes --]

[-- Attachment #2: 2987443.2987476.pdf --]
[-- Type: application/pdf, Size: 1312104 bytes --]

Re: [Starlink] GPS Dongle?

[Gary E. Miller]

[-- Attachment #1: Type: text/plain, Size: 735 bytes --]

Yo Ricky!

On Mon, 14 Jun 2021 13:25:26 -0700
Ricky Mok <cskpmok@caida.org> wrote:

> Attached a paper on this topic from my old colleagues.

Not bad.  It skipped the important details of how they configured their
RasPi's.  It also made the very common mistake of over smoothing their
ADEV measurements.  Lacking a few now standard metrics.  Otherwise
sound.

It is possible to do a little better now.

RGDS
GARY
---------------------------------------------------------------------------
Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
	gem@rellim.com  Tel:+1 541 382 8588

	    Veritas liberabit vos. -- Quid est veritas?
    "If you can't measure it, you can't improve it." - Lord Kelvin

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 851 bytes --]

Re: [Starlink] GPS Dongle?

[Nathan Owens]

[-- Attachment #1: Type: text/plain, Size: 1908 bytes --]

Hey Gary,

Got this running:

pi@realtimepi:~ $ ntpq -crv -pn
associd=0 status=0115 leap_none, sync_pps, 1 event, clock_sync,
version="ntpd 4.2.8p12@1.3728-o (1)", processor="armv7l",
system="Linux/4.19.71-rt24-v7l", leap=00, stratum=1, precision=-20,
rootdelay=0.000, rootdisp=1.195, refid=PPS,
reftime=e4753908.36ea5b4a  Wed, Jun 16 2021 19:59:20.214,
clock=e4753915.668e763d  Wed, Jun 16 2021 19:59:33.400, peer=23018, tc=4,
mintc=3, offset=0.298251, frequency=-22.915, sys_jitter=0.149930,
clk_jitter=1.980, clk_wander=0.051

     remote           refid      st t when poll reach   delay   offset
 jitter
==============================================================================
o127.127.22.0    .PPS.            0 l   13   16  377    0.000    0.298
0.150
*216.239.35.12   .GOOG.           1 u   36   64  377   61.574   -4.184
 24.605

Let me know how it looks, or if there's anything I could improve!

On Mon, Jun 14, 2021 at 1:38 PM Gary E. Miller <gem@rellim.com> wrote:

> Yo Ricky!
>
> On Mon, 14 Jun 2021 13:25:26 -0700
> Ricky Mok <cskpmok@caida.org> wrote:
>
> > Attached a paper on this topic from my old colleagues.
>
> Not bad.  It skipped the important details of how they configured their
> RasPi's.  It also made the very common mistake of over smoothing their
> ADEV measurements.  Lacking a few now standard metrics.  Otherwise
> sound.
>
> It is possible to do a little better now.
>
> RGDS
> GARY
> ---------------------------------------------------------------------------
> Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
>         gem@rellim.com  Tel:+1 541 382 8588
>
>             Veritas liberabit vos. -- Quid est veritas?
>     "If you can't measure it, you can't improve it." - Lord Kelvin
> _______________________________________________
> Starlink mailing list
> Starlink@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/starlink
>

[-- Attachment #2: Type: text/html, Size: 2787 bytes --]

Re: [Starlink] GPS Dongle?

[Nathan Owens]

[-- Attachment #1: Type: text/plain, Size: 2358 bytes --]

I should note this is very fresh and the jitter is dropping.

     remote           refid      st t when poll reach   delay   offset
 jitter
==============================================================================
o127.127.22.0    .PPS.            0 l    5   16  377    0.000    0.285
0.014

On Wed, Jun 16, 2021 at 8:00 PM Nathan Owens <nathan@nathan.io> wrote:

> Hey Gary,
>
> Got this running:
>
> pi@realtimepi:~ $ ntpq -crv -pn
> associd=0 status=0115 leap_none, sync_pps, 1 event, clock_sync,
> version="ntpd 4.2.8p12@1.3728-o (1)", processor="armv7l",
> system="Linux/4.19.71-rt24-v7l", leap=00, stratum=1, precision=-20,
> rootdelay=0.000, rootdisp=1.195, refid=PPS,
> reftime=e4753908.36ea5b4a  Wed, Jun 16 2021 19:59:20.214,
> clock=e4753915.668e763d  Wed, Jun 16 2021 19:59:33.400, peer=23018, tc=4,
> mintc=3, offset=0.298251, frequency=-22.915, sys_jitter=0.149930,
> clk_jitter=1.980, clk_wander=0.051
>
>      remote           refid      st t when poll reach   delay   offset
>  jitter
>
> ==============================================================================
> o127.127.22.0    .PPS.            0 l   13   16  377    0.000    0.298
> 0.150
> *216.239.35.12   .GOOG.           1 u   36   64  377   61.574   -4.184
>  24.605
>
> Let me know how it looks, or if there's anything I could improve!
>
> On Mon, Jun 14, 2021 at 1:38 PM Gary E. Miller <gem@rellim.com> wrote:
>
>> Yo Ricky!
>>
>> On Mon, 14 Jun 2021 13:25:26 -0700
>> Ricky Mok <cskpmok@caida.org> wrote:
>>
>> > Attached a paper on this topic from my old colleagues.
>>
>> Not bad.  It skipped the important details of how they configured their
>> RasPi's.  It also made the very common mistake of over smoothing their
>> ADEV measurements.  Lacking a few now standard metrics.  Otherwise
>> sound.
>>
>> It is possible to do a little better now.
>>
>> RGDS
>> GARY
>>
>> ---------------------------------------------------------------------------
>> Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
>>         gem@rellim.com  Tel:+1 541 382 8588
>>
>>             Veritas liberabit vos. -- Quid est veritas?
>>     "If you can't measure it, you can't improve it." - Lord Kelvin
>> _______________________________________________
>> Starlink mailing list
>> Starlink@lists.bufferbloat.net
>> https://lists.bufferbloat.net/listinfo/starlink
>>
>

[-- Attachment #2: Type: text/html, Size: 3495 bytes --]

Re: [Starlink] GPS Dongle?

[Gary E. Miller]

[-- Attachment #1: Type: text/plain, Size: 721 bytes --]

Yo Nathan!

On Wed, 16 Jun 2021 20:00:05 -0700
Nathan Owens <nathan@nathan.io> wrote:

> Got this running:

Cool!

> o127.127.22.0    .PPS.     0 l   13   16  377    0.000    0.298 0.150

Where are you getting PPS?

> Let me know how it looks, or if there's anything I could improve!

Let ntpviz chew on it for a day, then look at that data.  Prolly good enough
for testing starlink.

RGDS
GARY
---------------------------------------------------------------------------
Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
	gem@rellim.com  Tel:+1 541 382 8588

	    Veritas liberabit vos. -- Quid est veritas?
    "If you can't measure it, you can't improve it." - Lord Kelvin

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 851 bytes --]

Re: [Starlink] GPS Dongle?

[Nathan Owens]

[-- Attachment #1: Type: text/plain, Size: 1081 bytes --]

It's coming from the GPS Hat, MAX-M8Q with an external antenna.

Let me set that up... thanks!

On Wed, Jun 16, 2021 at 8:10 PM Gary E. Miller <gem@rellim.com> wrote:

> Yo Nathan!
>
> On Wed, 16 Jun 2021 20:00:05 -0700
> Nathan Owens <nathan@nathan.io> wrote:
>
> > Got this running:
>
> Cool!
>
> > o127.127.22.0    .PPS.     0 l   13   16  377    0.000    0.298 0.150
>
> Where are you getting PPS?
>
> > Let me know how it looks, or if there's anything I could improve!
>
> Let ntpviz chew on it for a day, then look at that data.  Prolly good
> enough
> for testing starlink.
>
> RGDS
> GARY
> ---------------------------------------------------------------------------
> Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
>         gem@rellim.com  Tel:+1 541 382 8588
>
>             Veritas liberabit vos. -- Quid est veritas?
>     "If you can't measure it, you can't improve it." - Lord Kelvin
> _______________________________________________
> Starlink mailing list
> Starlink@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/starlink
>

[-- Attachment #2: Type: text/html, Size: 1810 bytes --]

Re: [Starlink] GPS Dongle?

[Gary E. Miller]

[-- Attachment #1: Type: text/plain, Size: 864 bytes --]

Yo Nathan!

On Wed, 16 Jun 2021 20:11:10 -0700
Nathan Owens <nathan@nathan.io> wrote:

> It's coming from the GPS Hat, MAX-M8Q with an external antenna.

Then you can expect 100x better after it settles down.  Assuming
optimal configuration: fixed CPU clock, few cron jobs, etc.

PPS is usualy on SHM(1) and gnss time on SHM(0).  Looks like you used
the ntpd native driver.  gpsd will perform slightly better, and give you
a ton more data.

https://gpsd.io/time-service-intro.html
https://gpsd.io/gpsd-time-service-howto.html


RGDS
GARY
---------------------------------------------------------------------------
Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703
	gem@rellim.com  Tel:+1 541 382 8588

	    Veritas liberabit vos. -- Quid est veritas?
    "If you can't measure it, you can't improve it." - Lord Kelvin

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 851 bytes --]