Alan, am following your TCXO mod with interest, thankyou.
You wonder about the source of 50 Hz sidebands. May I posit a cause?....
Radiated RF has a habit of reflecting off adjacent semiconductor junctions that are strongly modulated by 50 Hz sources.
An example junction is a rectifying diode that creates DC from a power transformer's AC.
The diode effectively changes the length or reactance of the wires comprising that 50 Hz current loop...a 50Hz modulated
Reflected RF is thereby phase and/or amplitude modulated at 50 Hz rate. It can be the source of "tunable hum" in a direct
conversion receiver. It is often devilishly hard to track down the source of the reflecting RF junction - it may be buried in
another piece of gear completely unrelated to your experimental setup. Perhaps why your battery-powered attempt
So where's the direct conversion receiver in your setup?
It could be the TCXO itself. An unbuffered oscillator might act as a regenerative receiver driven hard.
I have experimented with microprocessor master xtal oscillators made with weak feedback to augment this
receiver-like property - it is able to lock to a weak signal introduced into the positive feedback loop.
A buffer stage after the TCXO oscillator (inside that nice shielding can) ought to reduce this effect....perhaps the
buffer is missing, or has too much internal output-to-input coupling (inadequate buffering).
Another possibility is that your monitoring receiver is combining a clean Si5351-generated signal mixed
with a re-radiated 50Hz-corrupted signal. Reducing 2M transmitted signal might help this.
There are many, many possible scenarios where this phenomenon is concerned.
This is really an aside, quite likely unrelated to the fast drift problem of primary interest in this thread.