[time-nuts] My new GPSDO design

[Pluess, Tobias]

Good day dear colleagues

I am in the process of designing a new GPSDO. As my previous one worked
very nice, but had several small design issues, I want to make a new one
that is a bit improved. So I have drawn a new schematic that is in the
attachment.

I use again the Oscilloquartz / "UTC" branded oscillator and an integrated
dual buffer / distributor and a TDC7200 as interpolator. I did not find any
other time to digital converter that has both, higher resolution as well as
a package that can be hand soldered, so I stick to the 7200. I am happy,
however, if someone can give me a hint for a possibly even better TDC.

For the GPS module, I would like to use the newest ublox. The NEO-F10T. I
have also added a power supply for the active antenna, and a monitoring
capability such that the antenna supply can be measured by the
microcontroller.

Also, with this new GPSDO, I want to add a serial connector that connects
also the 1PPS signal, such that it could be used by a PC with GPS/NTP
software, such as chrony. I was also thinking about PTP, but I found no
elegant solution for this.

Later, I want to make an additional daughter board that should incorporate
a Raspberry Pi or similar small computer that could run a web server and
will connect to the GPSDO with the serial interface, such that the GPSDO
could be monitored via a web GUI, and, additionally, NTP could be used from
this. I would not want to implement Ethernet and such on my microcontroller.

Also, I would like to increase the resolution of the DAC that I use to
steer the OCXO. For this, I have 2 optional solutions. One is the PWM DAC,
that I copied from the Oscilloquartz STAR4+ GPSDO I have. The other one is
the classic 16-bit DAC, and here, I would like to use later some kind of
delta-sigma modulation to increase the resolution. For this reason, there
is this low pass filter after the DAC.

Also I would like to add something new to this, if the computing power of
the microcontroller is sufficient. I would like to collect some statistical
data about the OCXO aging, store this data in the flash memory that I added
for this purpose, and then try to extrapolate the aging. Then, in the case
of holdover, use the extrapolated aging data to steer the OCXO such that
its aging is compensated. Not sure if this is even possible, but I would
like to try. With my current design of the GPSDO, this is not possible, as
there is not enough storage for some statistical data.

Also, thanks to the Raspberry that I could add later, the data could be
displayed in some kind of web interface and easily accessed over the
network.

What do you guys think about my design? Would it be a nice GPSDO or
complete crap ideas.

Also, I was thinking about adding a 100MHz output. But I am not sure if
this really adds value to it, as I see that 100MHz are used by some
instruments, but only rarely, and I am not yet aware of good, obtainable
100MHz oscillators.

I would be happy about some criticism about my new GPSDO project.

thanks
best greetings
Tobias
HB9FSX
-------------- next part --------------
A non-text attachment was scrubbed...
Name: gpsdo_hb9fsx.pdf
Type: application/pdf
Size: 2462889 bytes
Desc: not available
URL: <http://febo.com/pipermail/time-nuts_lists.febo.com/attachments/20250717/7bd2367a/attachment.pdf>