Suggest starting over with troubleshooting, changing one thing at at time and recording conditions for each step.
1st, where is the noise? Radio receive? intercom with radio off? Only through intercom with radio on? Noise in nav radio only? Noise in comm radio only? Is the noise transmitted on top of your voice to another radio in another plane?
Verify that there's no noise when the light is powered from a separate battery. If there's no noise with separate supply (in any mode listed above), then shielding isn't likely to help, because the noise isn't being radiated into the other devices. (If it were, then the substitute battery/wiring would radiate into the a/c systems, just like the installed wiring.)
If all's good with independent power/wiring, grab a 15'-20' piece of 'zip' cord (lamp cord) or other pair of wires & go directly from the battery terminals to the light, making sure the light isn't grounded to the airframe. Noise now? If yes, then the noise isn't radiating but is being conducted on the power wires. If no noise, then you likely have ground loop issues in the airframe. Try leaving the ground side temporarily in place, and hook up your supply side wiring that's installed in the plane. (This leaves the ground for the light connected directly to the battery without a 'local' ground.) If still good, then you *definitely* have a ground loop issue.
You mentioned a rear mounted battery (which I forgot was a possibility). I think you also mentioned grounding the battery locally (near the battery) to the airframe. If that's the case, you have the potential for many and big ground loops because some of your locally grounded devices may well follow a ground path past other locally grounded devices before making it back to the battery negative terminal (the 'ultimate' ground point). Not saying it should never be done; many certified planes do it that way, and many homebuilts, too, and many get away with it. But some don't. With the forest of tabs bonded to the a/c at the firewall, and the battery bonded to the a/c in the rear fuselage, the forest of tabs becomes a 'local ground' for all the stuff that's grounded there.
If the dedicated ground wire keeps the noise away, the next test you can try is to temporarily lift the battery's ground to the airframe, & run a light gauge ground wire from the battery negative directly to the firewall forest of tabs. (To state the obvious, don't try to run high current loads while in this state, especially engine start. :-O ) Then reconnect the installed supply & ground wires from your switch and the forest of tabs to the light (keeping the light isolated from a local ground). Still quiet? Buy some fat welding cable for your new ground wire. Noisy again? I give up....
Shielding, and connections for shielding, can be confusing. Grounding only one end only works if there's a separate ground path *in addition to the shield*. The whole 'ground one end' thing probably arose from magneto P-leads. Even magneto P-leads run current on the shields, at least until the motor stops, because the shield is the ground path for the magneto primary coil when you turn the mag 'off'. Be aware that when you install many intercom/audio systems, the shield is the return (ground) path for the audio signal. When you isolate the jacks from the airframe, you aren't 'grounding the shield at one end', because the shield *is* the ground. Isolating the jacks from the airframe prevents having *2* return paths for the audio, where the airframe path might have other, noisy returns from other devices.
Charlie
Edit: Just reviewed some of the previous posts, & realized that you mentioned 'wigwag'. That's another variable. If you have a wigwag controller in the circuit, when you ran your independent battery test was the wigwag controller in the circuit? (hint, hint...)
