On Sat 05 Oct at 10:24 John Clarke (3:713/730{at}fidonet) wrote to Russell Brooks:

 JC> The main problem with them is that they take a few 
 JC> micro-Seconds to fire during which time the over-voltage is 
 JC> presented to the computer/modem etc.  The best systems have 
 JC> arrestors backed up by high speed zener diodes (which react 
 JC> in nano-Seconds (1000th of a micro-Second).  The zener 
 JC> operates first and holds the voltage down until the arrestor 
 JC> has time to fire.

 JC> Another problem is that it is not easy to determine if the 
 JC> arrestor/zener combination is till operational or has failed 
 JC> for some reason.  If this has happened, and you don't know 
 JC> about it, then you are no better off than having nothing at all.

Wouldn't a high voltage, low current power supply be able to trigger the
zeners (exceding the zener breakdown voltage) without causing them to
destruct? Then have two sets of the zeners, and automatically test them one
at a time using a "make before break" switch so that even if
there is a power surge during a test, there will be one set of zeners to
carry the surge?

 JC> Perhaps the best arrangement is to also include a series 
 JC> fuse that blows when the arrestor operates.  This way you 
 JC> know when the arrestor has fired since the equipment won't 
 JC> work until the fuse is replaced.  At the same time the 
 JC> arrestor/zener combination could be checked to see if 
 JC> healthy and replaced.

True - I'd like to have something like this before using computer and
communications equipment in the tropics.

--- msgedsq 2.1