[CUWiN-Dev] Lighthning arresters

[David Young]

On Thu, Nov 03, 2005 at 01:34:58PM -0600, Bill Comisky wrote:
> On Fri, 28 Oct 2005, David Young wrote:
> 
> >On Thu, Oct 27, 2005 at 11:43:48PM -0500, Kristjan Onu wrote:
> >>>Our new installation is much more permanent, but there is a little
> >>>work remaining---mainly sticking the new cable to the roof.
> >>
> >>Are you still thinking about using a lightning arrester?
> >
> >Yes, absolutely.
> >
> >>I'm a bit confused. This picture shows a schematic of a gas discharge
> >>tube arrester:
> >>
> >>http://www.hyperlinktech.com/web/copyrighted_images/coax_lp_diagram.gif
> >>
> >>It shows that a ground connection is necessary. On the roof of the
> >>IMC you pointed to a plastic cannister connected to Volo's equipment and
> >>said it was a lightning protector, but it had no obvious ground
> >>connection. I can't quite see how a lightning protector would work
> >>without a ground connection.
> >
> >I don't remember pointing out a lightning protector on Volo's equipment.
> >Anyway, there isn't one up there.
> >
> >Dave
> >
> 
> I'm curious, has anyone out there experienced a lightning strike?  If so 
> can you post-mortem the devastation for us?  I can understand the purpose 

OJC lost its firewall in a storm back in August (?), when one or two
ethernet cards were zapped.  I believe a nearby lightning strike induced
a current in the line between Volo's equipment and our network closet,
and that current caused the damage.

One of the ethernets on the firewall was connected directly to Volo's
node, however, I believe the damage was done by a surge on the AC,
not the ethernet.  The way that Volo's PoE injector was wired up (don't
ask), the surge could have gone straight into the building's AC supply.
Both Volo's PoE injector and our gateway were plugged into "surge-only"
ports on a hefty UPS, however, I suspect that the surge-only ports are
isolated from surges on the building AC, not from each other.

BTW, now Volo's PoE injector is a different type that should isolate
the ethernet from the AC with a transformer (at least).  Also, the
PoE injector is plugged into a different UPS than any other equipment.
Next step is to put a surge arrestor on the ethernet line.

> of a lightning arrestor when the antenna is separated from the radio by a 
> long length of coax, but we are mostly using a ~1 foot jumper.  In this 
> case it seems like anything hitting the antenna/pole would fry the node 
> regardless.  A surge supressor on the cat5 going into the building (maybe 
> built into the PoE injector) seems like a necessary precaution.. 
> and grounding too seems like a good idea.  Just not sure of the benefit of 
> the lightning arrestor in this scenario.

I mis-spoke when I said lightning arrestor.  I don't think we have a
prayer of protecting our equipment from a direct strike, and we may
have worse problems than a ruined node after a direct strike, besides.
Sounds like we agree, protecting hosts' gear (like OJC's firewall,
CUWiN's and OJC's servers, and switch) from surges induced by lightning
is necessary and feasible.

I'm not sure if it matters where I locate a surge arrestor, whether at
the entrance to the building or in the network closet.  What do you think?

Dave

-- 
David Young             OJC Technologies
dyoung at ojctech.com      Urbana, IL * (217) 278-3933