[Elecraft] Surge Protection

[Jim Brown]

On 1/14/2013 3:32 PM, Don Wilhelm wrote:
> We had a close electrical strike last summer.  Nothing died as a 
> result of being connected to the AC mains, but it somehow jumped onto 
> the wired Ethernet home network.  I lost 4 computers, a NAS file 
> server, two switches, the router, and an access point.  It tripped all 
> the GFCI receptacles in the outbuilding (which does not have a surge 
> protection but has separately metered power feed).  Only one GFCI 
> breaker/receptacle in the house was tripped, so something worked. 

The computer networking destructive failures are VERY common, and are 
usually CAUSED by the use of shunt mode (MOV) surge suppressors on 
branch circuits.  I've seen many documented occurrences of this in the 
home offices of very good engineers working in pro audio, and the 
problem also occurs with audio gear that is interconnected with other 
gear powered from different outlets, or that are connected to other grounds.

In the pro audio world, we learned that the solution was to NEVER use 
shunt protectors on branch circuits, but rather to use SERIES MODE surge 
protection.  The ONLY good place for MOVs on power systems is at the 
service entrance.  See pages 29-30 of

http://audiosystemsgroup.com/SurgeXPowerGround.pdf

Also see the discussion of Surge Suppression beginning on slide 81 of

http://audiosystemsgroup.com/InfoComm-PowerSystems2012.pdf

The heart of the failure mode is that MOVs are shoving the lightning 
surge current onto the Green wire, and the resulting IZ drop across the 
Green wire between the MOV and the breaker panel raises the 
instantaneous voltage on the chassis of the "protected" equipment to a 
very high value.  When that equipment has a low voltage connection to 
OTHER equipment that is grounded at a DIFFERENT point, surge-protected 
or not, the DIFFERENCE in the instantaneous potentials on their 
respective chassis appears between the electronics that connects them, 
and destroys that electronics.

73, Jim K9YC