[Elecraft] slightly OT -- acceptable network switch (aka Linksysproduces birdies)

Sun Feb 7 07:38:53 EST 2016

Gooooooood Morning!  Isn't it a wonderful day to be alive!  :-)

Ed, what follows is not directed specifically AT you... or anyone else
for that matter.  I am just trying to put this whole thread into
perspective.

I can only speak to the subject from the perspective of my experience,
which at the time was to design and construct networks for customers
that exceeded the requirements of "normal customers".  For these
customers, for whom money was less of an object, and things like signal
integrity, operating "overhead", and signal security were essential
priorities.

1) As best as technology provided, allow no external signal to invade
the data network's signals paths to either alter or obstruct that which
was intended to travel within it.
2) As best as technology provided, allow no internal signal to escape
via the data network's signals path to either be intercepted and/or
become an interference source to coincident systems.
3) As best as technology provided, design should provide an operating
margin (or overhead) of NLT 50% excess to that contained in the minimum
requirements list.

The equipment/systems I was to provide was to be housed in a series of
(for lack of a better word) vaults with something like a "super Faraday
cage" build into the panels.  The rooms were connected by a series of
separate cable-ways having the same construction (separated by signal
and power and by signal type).  If you zoom out on the whole structure,
it would resemble a series of 100% shielded device chassis,
interconnected by shielded, multi-conductor cables, with the 100%
shielding terminated at each chassis structure to which it connected. 
Sound familiar?  It should certainly LOOK familiar...  Add a good
connection to ground, and you have a series RF to ground system to keep
RF off the device chassis in the shack...

MY task was to construct the data network within the space described
above to include the provision of power systems to feed the network. 
Yet, everything I provided was required to perform with the ASSUMPTION
that ALL of what I JUST described above had been totally compromised as
to be assumed NON-existent.  So, the testing phase of our build-out was
done under as close as possible to "real world" operations utilizing a
bogus data set IN FREE AIR, meaning in an otherwise unprotected
environment (like an office building or your home or out in the center
of your front yard really).

Point is...  We build that network and tested it. Then the customer
tested it with equipment I was not even allowed to look at (really
well-made device chassis with little to no indication as to function,
separate security/shipping boxes that contained what I can only presume
were remoted control heads, etc, etc...  racks of the stuff.  I and my
folks were put out of the facility, and they tested items 1-3 (not
inclusive) above "IN FREE AIR".

We passed... The mystery test equipment and people went away.  We went
to work.

I tell you ALL of the above to tell you this.

I hand-made every single Ethernet jumper cable used in the system above
(RJ-45 on either end).  EVERY SINGLE ONE.  I made the terminations for
almost every single modular plug.  I made every single termination on
every patch panel. I used off-the-shelf, quality CAT-5/5e UTP cable
stock in stranded/solid varieties rated for where it was to be
installed.  I used quality RJ-45 (4P8C) modular connectors and connector
boots and/or modular plugs.  I used quality tools. I terminated each end
of every cable using the specification for the standard. I tested every
single cable for signal routing and quality of signal.

But I did not use SHIELDED TWISTED PAIR cable.  There was not ONE SINGLE
INSTANCE of my CAT-5/5e standard cables being the SOURCE of a "birdie"
or any other RF signal that violated the stringent signal requirements
placed upon me. NOT ONE.
Were there some RF issues to deal with?  A few...  But as I remember it,
none in HF or low VHF.  Using my and other equipment we used logic and
proper troubleshooting to drill down to the SOURCE OF THE OFFENDING
SIGNAL... every single problem SOURCE was DEVICE related.  We either
corrected the issue, designed a mitigating solution (rarely), replaced
the device causing the issue, etc.

Signal cabling REVEALS problems.  If properly made, they do not CAUSE them.

Bottom Line:

1) Do not pinch pennies on cabling.  But do not THROW money at cabling
to create a band-aid over the REAL source of a problem
2) Use logic and proper troubleshooting to ID the source of the problem
instead of trying to HIDE the symptoms.  Once the source of the problem
is located, ONLY then do you start searching for a solution to ELIMINATE
THE PROBLEM, not cover it up.

I and an unrepentant over-engineer.  But even I choose the RIGHT places
to over-engineer.... and in this case, cabling is not one of the places
I'd choose.  Maybe I'll change my mind some day.  I doubt it.  ;-)

Maybe I should come out of retirement...  <big grin>

Y'all have a wonderful morning.  LIFE is GOOD because you read this...
which means you are alive!

73,

______________________
Clay Autery

On 2/7/2016 4:04 AM, Edward R Cole wrote:
> Clay,
>
> Perhaps.  Several years ago I decided to try to minimize network
> caused birdies by obtaining some clamp-on ferrites, but also bought
> some shielded Cat 6 cable to replace generic Cat 5e in the shack.  The
> shielded Cat 6 was very effective in minimizing many 2m birdies.  The
> clamp-on's were largely ineffective.
>
> I did this mainly to improve 2m for eme operation.  I did not check HF
> frequencies.  Generally atmospheric noise floor is high enough to bury
> birdies; 10m and 6m seem most effected.  We bought a new Netgear
> router (replaced Linksys) and I had some thoughts to enclose it in a
> shielded enclosure with ferrites on outputs and cooling fan, but then
> my wife acquired a ipad and kindle which all run on wifi...so
> shielding would interfere.
>
> I am considering building onto the garage a new 18x26 foot ham shack
> so that might improve isolation from the router which would be 72 feet
> at the other end of the house.  I still run wired internet cabling to
> shack computers. I suspect the wifi in neighborhood is bigger issue as
> is picked up thru the eme antennas and preamps.
>
> I am about to disconnect the XP desk top computer from internet and
> run e-mail and internet only on my laptop.  Be interesting to see if
> that impacts VHF birdies any appreciable amount (K3 and KX3 are
> connected to the XP computer).  I can eliminate the 5-port ethernet
> switch when I do that and only run a single 25-foot +cable from the
> router to laptop.  Other thing will be elimination of all wall wart PS
> in new shack (only have the one laptop PS unit).
> ----------
> My laptop just acquired a 21-inch monitor (from my wife's old XP desk
> top which is being recycled; she bought a new laptop last year with
> win7 now upgraded to win10...and the desk top was never used after
> that.  Going thru a general "spring cleaning" to reduce household
> clutter and computer got added.
>
> So the shack now has two full-time computers with monitors which will
> be handy when monitoring prop loggers and running special ham sw,
> simultaneously.
>
> 73, Ed - KL7UW
> -----------------------
> Date: Fri, 5 Feb 2016 07:43:37 -0600
> From: Clay Autery <cautery at montac.com>
> To: elecraft at mailman.qth.net
> Subject: Re: [Elecraft] slightly OT -- acceptable network switch (aka
>         Linksysproduces birdies)
> Message-ID: <56B4A709.5020102 at montac.com>
> Content-Type: text/plain; charset=windows-1252
>
> If changing the cabling was the solution for your "birdies" then it
> would appear the issue was not with the switch hardware, but with poorly
> constructed cables.
>
> There are a LOT of CAT-5 and CAT-5e labels on cables out there that
> simply aren't.  There's a list of specs involved, fairly lengthy one,
> all of which must be met to qualify a cable at a specific grade.
> Most cables that fail a check because they missed something in the
> termination procedure (not signal routing).
>
> ______________________
> Clay Autery
> MONTAC Enterprises
> (318) 518-1389
>
>
>
> 73, Ed - KL7UW
> http://www.kl7uw.com
>     "Kits made by KL7UW"
> Dubus Mag business:
>     dubususa at gmail.com
>
> ______________________________________________________________
> Elecraft mailing list
> Home: http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:Elecraft at mailman.qth.net
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
> Message delivered to cautery at montac.com