I have run a simple test of temperature drift of the two RockMites :
Using a fan blowing on the PCBs lying on the desk the RM80 drifts less than 10Hz when the fan comes on or off for about 5 minutes, that is quite acceptable, given that the RMs will be boxed, and likely I will use some insulating Styrofoam.
The RM30 is a bit more tricky. I tried to lower the frequency with a choke in series with the tuning trimmer capacitor I mounted, and the drift with fan was about 50Hz at on/off.
It turns out that grounding one side of the crystal resulted in a frequency of 10138.490. Not ideal for WSPR, but should be useable with WSPR2 and "BFO offset". Not tested yet, but will be in the near future. Temperature drift at the fan on/off is now 20Hz, and with a bit of insulation that should result in an acceptable temperature stability.
It looks like the drift is more due to the capacitors in the oscillator than the crystal in itself.
Oh, and I am a bit disappointed that the RMs have several zener diodes, but the oscillator itself runs from the raw supply voltage. Where did that idea come from ? I find it surprising that there is not more stability problems with the LO in the RMs
Next step with the RockMites : input crystal filter tuning - if necessary.
QRP-LABS 5Watt PA - I am still struggling to make a "simple" 5Watt PA for the 1.8 - 50MHz range The QRP-LABS 5W PA could not deliver more then 1W at 28MHz using the original...
5 minutes ago