AVR-Chat

>
>Oh yes, a failed diode could cause a mess in the system, but a mess would
>also caused by a bad contact at the backup battery leads by corrosion, then
>no backup will be available.

That's a survivable failure. What I'm thinking of here, is a failed diode charging a lithium backup battery, which definitely can lead to an explosion, even at relatively low current. UL requires double protection mechanisms for this.


>Backup systems use to be a real pain in the neck, since the only moment you
>can be sure it is working, is when you most need it.  If not working, you
>wasted a lot of effort for nothing.

Yes, and the added complexity makes it MORE likely to fail. 
Dammit.


>This is almost the same as hiring someone to be your bodyguard, to feel safe
>you need to see him from time to time watching over you.
>
>Of course, there are many ways to produce an efficient backup switch, but as
>much components you insert into the thing, more complex it turns to be,
>thus, more subjected to fail.  In all my prof life, I saw much more, much
>more bad and oxidized contacts than open or shorted diodes where low current
>is involved.

I won't use slide-in lithium coin cells for that reason. Not enough wetting current, even with high contact force.  Solder-in only for me.

>Just as information, several metal traffic bridges use a small AC current
>through the estructure, just to reduce oxidation and rusting at the screws,
>bolts, joints, etc.  This small AC current also serves as to measure the
>oxidation level of the bridge.  There are some kind of gadgets that transmit
>such information via RF straigh up, so it can be read by a county airplane
>or chopper.

Cute!  I see dashboard application. Unsafe bridge ahead! :)
Fish floating in the water, due to open circuit on the bridge might be a bad sign..