Bring up and learn about Calnex GPSDO

Through a happy chance, I'm able to use a [Calnex CN-HW-PX-OPT132](https://info.calnexsol.com/acton/attachment/28343/f-bf485192-7219-4d19-a172-1cc951764a91/1/-/-/-/-/CX2007v2.0_Rb-Ref-Source_DS_w.pdf) as a very stable external GPSDO (GPS Disciplined Oscillator).

With a GPS lock, it has:

Output Frequency Accuracy/Stability:

- 12 mins warm-up time @ 25°C: 5E-10 for frequency stability
- 24 hrs warm-up time @ 25°C and GPS locked: 1E-12, typical (requires a stable temperature
of ±1°C)

Output time accuracy/stability:

- 24 hrs warm-up time @ 25°C and GPS locked: <50ns (requires a stable temperature of ±1°C)

But using its internal Rubidium oscillator, even if GPS goes away, you get:

Output Frequency Accuracy/Stability:

- 12 mins warm-up time @ 25°C: 5E-10 for frequency stability
- 24 hrs warm-up time @ 25°C and Holdover (lost GPS): <3E-11/month (requires a stable
temperature of ±2oC)

Output Time Accuracy/Stability:

- 24 hrs warm-up time @ 25°C and Holdover (lost GPS): <2µs/48hr or <1µs/24hr (requires a stable
temperature of ±2°C)

Unfortunately my existing HVAC is probably not going to guarantee ±1°C, but I could probably hit 2°C...

The issue names no repository files, tests, or entry point. Start by clarifying how the Calnex CN-HW-PX-OPT132 should integrate with time-pi and what observable behavior is required; the issue currently does not define acceptance criteria or what done looks like.

Written by the indexing model from the issue text.