RM>> a nearby lightning strike knocked out our power for a while
RM>> (according to our neighbour, we were all out). The cables to
RM>> the sprinkler solenoids must have acted as antennae

JB> why noy just use two zeners back-to-back (in series, but facing
JB> opposite directions)

 The same reason that you don't use an ant to tackle an elephant.

 By the time the lightning gets to the computer interface, it is
5,000 volts (the wires themselves arc over at around 5KV) with
basically zero impedance for the fast leading edge. The current can be
20,000 amps! So... two brave little zeners face the charging elephant
and are swept aside.

 The *first* thing you have to do, is get a few K in series with the
line, and then organise a spark gap to limit the fast leading endge of
the pulse. A 2K2 spiral-cut resistor works rather well in either
carbon of tin oxide (the voltage actually arcs across adjacent tracks
and dissipates energy) and a 1uF electro is a very good low-voltage
spark gap (30V or so) is you can arrange to make the system run
slowly. At 2400Bd, a 0.1uF barrierlayer capacitor works okay, but
occasionally shorts out. Zeners and varistors are basically useless.
The micro itself is stronger than the zener (it is designed to take
2KV, btw).

Regards,
Bob
 

--- BQWK Alpha 0.5