Nice that he recovered his loop and shared the details with us, wonder what the mfgr thinks of it. However, the 1n4148 is a real imd generator in strong sig environments and if used as an rf switch, say for bandpass input and IF filter switching, think I'd use 4001s in their place. The 4001s have a quasi-PIN structure rather than the 4148s PN structure, wich makes the 4001s a bit less prone to become a mixer and are hardier by far. The loop should help with keeping signal inputs low compared to a "long wire" or other antenna, wich could provide enough rf to 4148s to create serious imd headache.