I am curious, did you discipline the PRS10 with 1PPS as well or it is just
free-running?  Is your PRS10 new or used?

I bought mine brand-new (fully optioned out), directly from Stanford
Research Systems, this past summer and have been slow rolling a
master-clock build for my home lab (have not even ordered boards yet).
Since my PRS10 is destined to be hold-over mostly, the board will be using
the Ublox ZED-F9T directly on the PCB.

I do recall from the manual (
https://www.thinksrs.com/downloads/pdfs/manuals/PRS10m.pdf) that there are
some internal calibration registers you access from the serial port to tune
the control loop - but you want a stable reference input to align with
first, a GPS receiver with a known location can better solve for time
errors.  I've found the same procedure used for RTK base stations is
suitable for good timing, Sparkfun did a blog on setting up am
RTK reference station a few years ago, which includes a correction of the
location by  post-processing 24-hours of RAW survey data (
https://learn.sparkfun.com/tutorials/how-to-build-a-diy-gnss-reference-station/all#introduction
).

-TS

On Tue, Feb 11, 2025 at 12:30 AM time-nuts-request@lists.febo.com wrote:

Hi

If you want to set up the ZED-F9T (or any of the similar “timing” parts):

1. You very much want an antenna location that gives you > 6 sat’s in view from each
   system you are using 24 hours a day. Ideally you want to be locked on to 6 sats from
   each at all times.

2. You want an antenna that is suitable for base station use. That’s not to say it’s
   crazy expensive. You just would like to avoid magnetic mount “hockey puck” style units.

3. Yes, this is overkill, but this also is Time Nuts :). Run a week or so of data out of the device.
   Package that up and submit it to one of the many free services that will run corrections on it.
   Do this about 3 weeks after the data was collected. Net result should be a very accurate
   idea of your location.

4. Put your module into position hold mode at the location the correction service comes up
   with. If there are speed settings make sure they are set to “static”.

At this point you’ve done pretty much everything you can to improve the accuracy of the
device. The weeks worth of corrected data also should give you some neat plots of what’s
going on. That will help you spot anything weird that just might be a problem as you move
forward.

Just what correction service(s) you decide to use depends a bit on the version of the F9T
you happen to have. A L1/L2 version gives you a few more choices than an L1/L5 version.

Fun !!

Bob

It is used.  I decided to connect it to the PPS from the Z3805A, and thankfully it is working.  In the last 60 minutes, the frequency has decreased 14 mHz.  I think it is heading in the right direction.

Walt

-----Original Message-----
From: Tim S via time-nuts time-nuts@lists.febo.com
Sent: Tuesday, February 11, 2025 9:35 AM
To: time-nuts@lists.febo.com
Cc: Tim S tim.strommen@gmail.com
Subject: [time-nuts] Re: PRS10 - how do the values look?

I am curious, did you discipline the PRS10 with 1PPS as well or it is just free-running?  Is your PRS10 new or used?

I bought mine brand-new (fully optioned out), directly from Stanford Research Systems, this past summer and have been slow rolling a master-clock build for my home lab (have not even ordered boards yet).
Since my PRS10 is destined to be hold-over mostly, the board will be using the Ublox ZED-F9T directly on the PCB.

I do recall from the manual (
https://www.thinksrs.com/downloads/pdfs/manuals/PRS10m.pdf) that there are some internal calibration registers you access from the serial port to tune the control loop - but you want a stable reference input to align with first, a GPS receiver with a known location can better solve for time errors.  I've found the same procedure used for RTK base stations is suitable for good timing, Sparkfun did a blog on setting up am RTK reference station a few years ago, which includes a correction of the location by  post-processing 24-hours of RAW survey data ( https://learn.sparkfun.com/tutorials/how-to-build-a-diy-gnss-reference-station/all#introduction
).

-TS

On Tue, Feb 11, 2025 at 12:30 AM time-nuts-request@lists.febo.com wrote:

time-nuts mailing list -- time-nuts@lists.febo.com To unsubscribe send an email to time-nuts-leave@lists.febo.com

Awesome!

IIRC, the default integration time is something like 18 hours (it's in the
manual, and can be changed via serial commands).  Give it a day of good
PPS, and it should be okay to unplug the PPS input for up to around 48
hours and it should still very well match the GPSDO reference (but will
likely drift again over time).

Good advice from Bob (kb8tq) on the Physical setup of a reference GPS
station.  I have my GPS antenna (a used Topcon PN-A5) clamped to the
structure of a masonry building with clear line of sight horizon to
horizon, LMR600 to the rack.

-TS

On Tue, Feb 11, 2025 at 12:06 PM Walt walt@drizzle.com wrote: