On Sun, 20 Dec 2020 12:58:29 +0000, Another John wrote:

... and I have sealed the bell-push with various gunges round the edges
every now and again to try and keep out damp ...

Mistake. It's very difficult to seal anything just with "sealant".
Getting a reliable water vapour proof bond between the selant and
substrate is next to imposible. Atmospheric pressure variations force
water vapour in. It then condenses and can't fully get out again.

... -- but after a period of solid functionality, it always becomes
unreliable again.

'cause it gets wet... Unless an enclosure is designed to seal, ie
have a complete, compressable, gasket of some sort between suitable
surfaces its far better to accept that water will get in and give it
some means to get out, ie at or near the lowest point. Doesn't need
to be very big 1 mm maybe 1.5 and sheltered from direct, wind blown,
rain if possible.

Few examples. Exterior "water proof" mains socket, with the cover
clipped down and no plug/cable in place, would end up very wet
inside. Drilled small hole hidden up a recess at the bottom of the
cover. No longer gets wet inside. Rear light cluster had condensation
in it, drilled small hole at the bottom on the back of the unit. Took
a couple of weeks to clear but now doesn't steam up at all.

One thing that bugs me about the kinetic one is: the bell-push travel
seems tiny (I've tried it through the packaging): how can such a tiny
movement generate enough [whatever].

Probably piezo electric but relatively gently flexing the crystal
rather than high stress/sudden release of a high voltage piezo
igniter.

--
Cheers
Dave.