I have tried many too..both on active loops and mag loops ( feed loop.)
My conclusions where..
Shielded vs unsheilded no difference at all..on mag and active.so not worth the effort of making shielded loop..
Below are comparisons of 1m dia wire loop on active antenna.( wire 3mm dia.)
Mobius on active loop..no difference or worse due to higher inductance
19mm copper tubing..big difference..19mm wins.lower inductance. Lf-30mhz.
Parallel loops of 12-19mm ..big difference..lf-30mhz.
Crossed parallel loops of 19mm or 12mm copper tubing ( each loop 1m square.) best by far..lf-30mhz..but now one has an array of 2x2 m so not small!
Cardioid loop..poor..needs adjustment of input matching into loop amp i suspect..too high inductance for loop amp..TOO be continued as not investigated loop enough yet..if can get loop amp working well with, will investigate phased array of..
Another Issue with cardioid loop is poor nulling of noise compared to standard loop..but does have lower “ take off” angle..hence why further investigation required..However phasing two parallel non cardioid loops via delay line brings their “ take of” angle down to about 30 degrees..and they perform much better in regards to signal strength.( which matters as delay line plus combiner robs signal..) currently building adjustable delay line ( see lz1aq site.) to play with this..AND also phasing cardioid loops when i get the amps working well with..( as should get even lower angle “ take off”.
All above tested with ccw and lz1aq amps..all in urban high noise qth.