Thanks to all... some good ideas. Yesterday I experimented: 1) antenna in long copper pipe (fadaday cage?), grounded and ungrounded. Result= no attenuation of signal, 2) then put a 20' length of coaxial video cable in series with the antenna, grounded the shield. Result - no attenuation. 3) Put the coaxial cable inside the copper pipe, both grounded. Result = no attenuation of signal. 4) Tried Dr. Pepper's clever idea...
grounded coaxial shield at one end, bifurcated the signal at the other end (and vice-versa), putting the signal down the core and also simultaneously down the coaxial shielding. This should have been equivalent to running two lengths of antenna in counter parallel orientation. Result = no attenuation.
I really don't see why solution #4 does not seem to work, since doubling back on the antenna so they are positioned parallel and close, kills the emitted signal when using the normal single conductor unshielded antenna wire. Perhaps the ground on one end of the shielding was inadequate, I did not confirm the effectiveness of the ground.
Anyway... giving up. Will use a rather contorted loop-back layout that is conventional and similar to some proposed in the manufacturers examples of layouts that work.
Still would like to learn more about how this system works, why any parallel or twisted arrangement of the antenna seems to lead to effective destructive signal cancellation regardless of the lengths of the loop-back (which presumably affects the phasing of the signal, allowing destructive interference). And why it is so hard to attenuate the signal with a fairly generous Faraday Cage ...
j