[thedanbob]

> While it is theoretically possible that the relays could fail on through some sort of physical failure, this is so unlikely that we did not design for it.

Anecdotally, I've had a relay fail on when I inadvertently pulled more amps through it than it was rated for, so it's definitely possible.

[Animats]

The relay they're using is quite likely to fail. It's a no-brand imitation of a relay where the real one is only rated for 10^5 cycles driving an inductive load.[1] Also, they needed a DPDT relay, which they are emulating using two SPDT relays operated together. If the software ever operates only one of them, the door will remain locked regardless of what the entryphone box does. Also, no fuse or snubbing. There's a whole industry of really crappy control relays from China, especially on the solid state relay side.

A useful device to know about is the Relay In A Box line.[2] This is exactly what it says - a relay in a box, for when you need to switch power with a low-voltage control signal. UL and CE approvals, fits standard electrical conduit fittings, and will pass code inspection. Rated for 10 million cycles. Boring, but useful.

[1] https://www.alldatasheet.com/html-pdf/1132639/SONGLERELAY/SR...

[2] https://www.functionaldevices.com/category/building-automati...