OCXO's are awesome - this is what I learned while trying to get WSPR working reliably on the 2m band. Here is all you need for reliable WSPR'ing on the 2m band: A RAW OCXO "can" works fine too! Note: This modded Si5351 board image comes from https://www.simonsdialogs.com site. Essentially we are modding the CLK2 output port to become an input port which accepts the OCXO's output signal. ...
The https://github.com/rxt1077/wspr_spread work is pretty awesome. It helped us debug why our TCXO-powered-WSPR beacon was pretty decent on 4m band but flaky on 2m. Doppler spread value on 70 MHz: ... 70.0924994 VU3CER MK68 23 0 0.29 1 1 0 1 44 1 810 0.517 Now see the problem on 144 MHz: ... 144.4905417 VU3CER MK68 23 -3 0.12 1 1 0 1 39 1 810 0.865 Pretty all over the place but we are operating on 2m - so probably expected for a 25 MHz TCXO based system! ...
I am posting an early design sketch of a 2m WSPR beacon. Here are the key points: 25 MHz HCI 0.5ppm TCXO powering the Si5351 module instead of the 26 MHz TCXO. If this fails, we can use a 10 MHz OCXO instead (a bit out of spec but works fine). Tokmas CID10N65F GaN FET might work @ 144 MHz as the 'final'. If not, use Mitsubishi RD15HVF1-501 MOSFET as backup. ...