Logically(?) if the frequency gets too high, then the signal will avoid entering the interior of the litz bundle at all and prefer to "crawl" across the external surface of the bundle - transferring power by "radiation" rather than conduction? in which case, solid wire or tube would be better, in that transfer is aided by some conduction as well, rather than having to radiate (leap) from strand to strand on the surface of the litz?
I knew that the reason HF and VHF transmitter coils are sometimes made of 1/4 or 3/8 or even 1/2 inch copper tube instead of soild wire was that there is much more "surface area" per lb. of copper using tubing instead of solid wire (and solid copper wire 3/8" diameter would cost a fortune and be really hard to work with).
But with transmitters you are dealing with alot of power compared to what flows in a X-tal set's coil and I always figured that litz wire large enough to handle those kinds of currents would be prohibitively expensive so they used tubing instead.
Litz wire mitigates skin effect up to about 1 MHz. I believe that as you go higher in frequency the coupling between the wires reduces the advantage of using Litz wire. Something called "proximity effect," where the field in one wire affects the fields in other nearby wires.
Sounds almost like "self capacitance" in a close wound coil. Or maybe it sets up eddy currents that interfer with resonance of the coil and kill the Q?
Every evaluation of BCB quality litz (44-46-48 gauge) that I have ever seen shows a substantial drop off after a couple of MHz. I've never actually seen or heard of 50+gauge litz so I can't guess if it could/would display the same characteristics.
Well... I quess I learned something today.
Besides...solid wire is much less expensive at that point!
I think solid wire is less expensive than litz at all points.