The issue is the gap in a door normally forms a long line.
The fact they interfere with WiFi should make it obvious the average microwaves faraday cage is far from perfect.
Anyway, I don’t recall all the details but 1W from a microwave is probably an underestimate.
WiFi uses multiple separate frequencies and outside of what a microwave should be producing. So it’s significantly more energy to block it than you might think especially when the hub is closer to the device than the microwave.
PS: To be clear this is still a trivial amount of energy, just annoying when reheating food blocks WiFi.
So you were off by a factor of 200, and you can see that my referencing -40dB of attenuation was an underestimate. I do wonder what details it was that you thought you might recall.
Quite disappointed by HN standards of self-moderation yet again.
A 1 foot cube, has 6 faces of ~30 cm * 30 cm or 5,400 cm2. 5,400cm2 * 5mw = 27 watts. Of course 2 inches from the surface is a significantly larger box.
Of course that’s a legal maximum, most devices should be well below it.
Anyway, math is math but presumably this is why that’s the legal limit. It’s low enough to be safe while high enough not to be expensive for manufactures.