A recent experiment exposed significant IMD of a 7MHz locally radiated signal.
The source used for these tests is a battery powered low power transmitter driving a 0.6m square loop. Radiated power is very low (of the order of -60dBm EIRP), and received signal on the station receiver is less than -73dBm.
Loop near steel shed #1
The loop was leaned against the colorbond sheet steel wall of shed #1. Shed #1 is about 40m from the receiving antenna, and is connected to the power mains. The building has colorbond sheet steel screwed to a steel frame. Colorbond is painted Zinc/Aluminium coated steel.
For this test, the submain was turned off at the main switchboard, so there is not equipment in the shed powered up, but the supply neutral is still connected and bonded to the shed steel work and shed PES ground electrode.
Above, the spectral response of receiver output, there are a number of side products at 100Hz intervals, presumably some form of IMD. At higher radiated power, products at odd 50Hz intervals become visible, though at much lower level than the 100Hz products.
Turning the power on to the shed made no noticeable difference.
Loop near steel shed #2
The loop was leaned against the colorbond sheet steel wall of shed #2. The building has colorbond sheet steel screwed to a steel frame. Shed #2 is about 40m from the receiving antenna, and is not connected to the power mains, but it is on a concrete slab that connects to that of shed #1 so they are not truly galvanically isolated.
The audio output of the receiver was recorded for both tests.
Above, to the left of the cursor is the first test, a small amount of amplitude modulate at about 100Hz is quite evident. To the right of the cursor is the second test which shows very little amplitude modulation.
A very similar pattern to shed #1 is observed when the source is located inside the main residence which has a 300m^2 sheet steel roof with sarking on timber trusses.
Locating the transmitter further away from the buildings reduces the modulation products, and they are not visible in the band noise at 30m distance from buildings.
There is some significant intermodulation process that seems due to steel sheet construction where the sheet is bonded to the PES, even where the mains is turned off.
The presence of dominant 100Hz modulation suggests that the modulation process might be caused by the zero crossing of 50Hz currents in contacts between sheets.