Removing ground plane under small cap trimmer

[jmw]

In my progress in building a active probe, I need some very small adjustable capacitance, and went with one of these: https://www.digikey.com/en/products/detail/knowles-johanson-manufacturing/27283-3R10/9094377

When I tried it out in a prototype, I found the circuit performed quite a bit out of spec, the adjustment afforded by the trimmer was measurable but barely so, and depopulating the trimmer had almost no effect on performance. The trimmer is connected as shunt to ground. With the recommended footprints[1], I computed the capacitance of the big circular pad to the ground plane: it was nearly 2 pF. The optimal trim value in simulation was 650 fF, so it looks like I completely swamped this trimmer with my SMD pads. When I put these parasitics back into the simulation, I was staring at nearly the same result on my instruments.

So it looks like the next step should be to cut out the ground plane under these pads, but how much further beyond the pad dimensions do I need to go before the coupling is negligible? Should I remove the ground plane under all the component pads from the probe tip up to the buffer amp input or perhaps just this one?

[1] https://www.knowlescapacitors.com/getattachment/Products/Capacitors/Trimmer/Multi-Turn/Giga-Trim/Giga-Trim-Solder-Pad-Layouts.pdf