Dr. Slack wrote ...

What's RF ground?

Let's start with an idealisation. Imagine an infinite sheet of metal, stretching to the horizon in all directions. You set up your TC in the middle of it, and drop a wire from the secondary base to the metal. That's your RF ground.

The RF ground forms the 'other plate' of the capacitor through which the secondary current flows, the 'top plate' being your toroid, and any (conductive) streamers coming from it. I like to think of RF ground being any piece of grounded metal on which 'lines of electric displacement current' can flow to from the top terminal, and return by conductors to the secondary base.

In a less ideal case, you're sitting in a 10th floor apartment. You connect the secondary base to the mains earth at the wall socket. That earth will be bonded to the earth wires all over your flat, capacitively coupled to live and neutral wires buried in the walls, bonded to the wiring in the flat above, and ultimately bonded to that relatively conductive dirt plane stretching out in all directions to the horizon. The main contribution to RF ground will be the wires in the floor, the walls, the ceiling, through which the capacitive displacement current returns to the secondary base. That current is more or less shielded by those conductors from the big earth outside. Which is why if you're operating in that sort of difficult situation, you're advised to put a few 10s of nF earth to live and earth to neutral at the socket where you plug in your TC, to make sure all three wires are officially working as your RF ground.

The current transformer just samples the return current to the base of the TC from the capacitive coupling of the top.

Dr. Slack wrote ...

No, I think you've got hold of the wrong end of the stick. However in your defence, I don't think I demonstrated the stick too clearly.

The RF ground that the TC 'sees' is the conductor around it. There's no choice about this, RF current is flowing into the conductors most local to the TC. If you're in the middle of a field, then the nearest conductor is the earth. If you're in an apartment, then the nearest conductors are the wires in the walls, the plumbing, the HVAC ducting etc.

A circuit (and we all know that current only flows in a complete circuit, don't we) might start at the secondary bottom, through the secondary, topload, capacitance to the nearest conductorsl, some net of impedances which connect them back to what you're calling your RF ground, and the wire from that to the coil bottom, that last wire possibly interrupted by your current transformer.

The key to managing your RF ground, and any interference it causes to your and neighbouring properties, is that phrase 'some net of impedances'.

If you bang a spike in the earth out there somewhere, and run a wire back to your TC secondary bottom, then that 'net of impedances' has got to get RF current from all the wires in your walls, floor and ceiling, into the protective ground maybe at your meter inlet, and back though this long wire. That's a long path, which includes several high inductance long wires, capacitive coupling between live and earth conductors in distribution wiring and any appliances plugged in, and who knows what else.

If you want to minimise interference, you shorten the path current has to take. At the mains inlet to your TC, you connect live, neutral and ground at RF with some X2 filter capacitors, to keep the RF out of the differential between those conductors. Then you connect the TC power supply earth conductor to the secondary bottom. The existing protective earth conductor, grounded at your meter inlet, is just as effective, more so, than a wire to another metal spike in the ground.

If you're in the middle of a field, or a yard, then the RF current will flow from the topload capacitivly to the ground. Rather than taking the long route through the ground, the ground bonding at the meter inlet, the earth conductor tot he TC, in this case you should spike the ground, and connect that to the TC secondary bottom.

In summary, connect all the conductors *nearest to the topload* by as short a connection as reasonable to the secondary bottom, which we then call RF ground.