[TowerTalk] Lightning and wired networks was: Re: Lightning protection

[Jim Brown]

On 7/9/2013 6:40 AM, Pete Smith N4ZR wrote:
> Sure, but the network cables were connected directly to the computer 
> motherboards.  I have also lost a couple of RS-232 ports on my K3, 
> which now has a DB-9 surge protector on it - I'm inclined to think 
> it's a design problem. 

YES. Cable shields MUST be bonded to the SHIELDING ENCLOSURE (the 
chassis) at each end.  If there is no shield, the signal return must be 
bonded to the shielding enclosure at both ends. Connecting the shield or 
signal return to the circuit board is the fundamental cause of failure 
due to lightning, and also causes hum, buzz, and RFI.  The failure to 
connect the cable shield or signal return to the chassis is called "The 
Pin One Problem."

A contributing cause is the use of SHUNT-MODE surge suppression (MOVs)  
on branch circuits. In the event of a strike, each suppressor shoves the 
strike to the GREEN wire, and there is IR drop in every individual green 
wire as a result of its INDUCTANCE. If there is equipment plugged into 
multiple shunt-mode suppressors at different points, each suppressor 
sees  a DIFFERENT inductance on its green wire, so there can be a VERY 
LARGE difference in potential (voltage) between that equipment. That 
voltage is what causes destructive failiure.  There is a similar result 
when one end of the cable goes to a chassis plugged into a suppressor, 
and the other end goes to the station ground.

The guys working in pro audio have a LOT of very long cables plugged 
into equipment at widely separated points, and long ago figured out that 
shunt-mode surge suppression was a bad idea. Neil Muncy, working in pro 
audio, was the guy who figured out "The Pin One Problem."  I use nothing 
but series-mode suppressors on branch circuits. Mine are made by Surge-X.

There's a lot about these issues on my website. http://k9yc.com/publish.htm

73, Jim K9YC