Ok, we spent ages at Knowles incorporating RF suppression in hearing aid mics.
Basically the process involves grounding the case to the negative terminal and fitting a low pass R/C filter. The original idea was to kill off the rfi signal from mobile phones in the 800/900 MHz, 1.6/1.8GHz and circa 2.4Ghz bands; you could hear the harmonics as a ‘motorboat’ sound.
There are limits though, if the RFi signal is too great, you’ll start to detect it with the various lengths of metalwork within the mic matching the wavelengths of the signal.
Which means only one thing in my book, persistent noticeable interference when the microwave is running, means the field density of microwave radiation is enough to overcome the RFi suppression.
The faraday cage is leaking, or the magnetron is pumping some RF out the wrong direction through its shielding.