For example on 7MHz, I see 3vpp to 600mvpp as I tune the phase potentiometer, in the circuit I have prototyped. Whereas on simulation the amplitude is almost constant.
What you did is ok. You removed the potentiometers (used as potential dividers) and directly varied the control and input AC voltages.
The frequency of operation is anywhere between 1.8-30MHz.
You will find that the shift is minimal on lower frequencies with just the varicap. To add more range, I will add parallel capacitors switched in parallel to the varicap on lower frequencies. I have simulated that and it seems it works.
For for the shake of simplicity let's consider one frequency to evaluate it and whereas you see in your simulation very small amplitude variations, in practice my prototype shows huge (example on 7MHz, I see 3vpp to 600mvpp as I tune the phase potentiometer).
It is to be used for phasing the noise out of a noise antenna and then canceling it from the main antenna. I opt for a 360 degrees phaser, so more than one stages need to be cascaded. But the performance of the single stage shown here, as loaded to the 1M probe scope, is not what is expected from the simulation... What could cause this?
I use a 10:1 probe at the output of the phaser and a 1:1 probe (a coaxial cable) to ptobe the 50R output of the generator. The output voltage of the generator is 300mv to 380mv depended on how I set the phase potentiometer in the phaser. I do not know why there is this interaction.
Another weird interaction is that, when I set the gain potentiometer near maximum, the phase does change, where it should not!