[RFI] ethernet switch with shielded case or low emission spectrum

[David Eckhardt]

The usual and almost ubiquitous answer to taming an RFI noisy Ethernet
switch (and a lot more) is, "Put it in a Faraday Cage".  Well, guys and
gals, as Ed wrote, there's a WHOLE lot MORE to it that that!!.  The devil
is in the details.

One item I've constantly run into in this respect since the RoS initiative
(elimination of Pb and Cr from electronics) is metal passivation.  When RoS
came along (mandated by the EU), Europe and the Western world went with
Cr-free CONDUCTIVE coatings.  The PacRim went with NON-CONDUCTIVE organic
compounds.  I had one instance where we were using an Ethernet Switch on
manufactured from the PacRim (China) but otherwise from a reputable
European company where the organic NON-CONDUCTIVE coating was used.  The
box was dirty as anyone could imaging from HF through UHF.  We had to ship
our product in which this switch was a part of (make money).  Therefore, I
set up a small al-hoc "production line" to scrape, sand, grind, and,
file..... the PacRim coating off the switches such that all the metallic
parts were in intimate and RF contact with each other.  That rendered our
product passing to EMC/RFI requirements and we could ship (and make
money).  We passed some 200 of the switches through the small ad-hoc
'production line' to clean them up due to their previous bad RF emissions.
I became so frustrated at delivering the same message over and over and
over again that I finally termed the PacRim coatings, "PacRim Spit".  Those
in the EMC/RFI discipline, BEWARE!!!!!  The basic design philosophy of the
switch was true to the Faraday Cage concept, but totally laid to ruin with
the non-conductive metallic passivation.  Heavy and nearly indestructible
powder coating (paint) can precipitate the same problem as PacRim Spit.
Beware of just throwing a "Faraday Cage" at a product to tame a nasty RF
emitter.

Dave - WØLEV

<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon>
Virus-free.
www.avast.com
<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link>
<#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>

On Thu, Mar 29, 2018 at 7:50 PM, Edward Mccann via RFI <rfi at contesting.com>
wrote:

> This overview came over the transom today. Interesting background,
> supporting what our RFI gurus have been saying.
>
> Passed for information.
>
> http://images.link.pentondes.com/Web/PentonDES/%7B31de5541-
> facc-4790-8147-385e74703ca2%7D_wp-design-by-desense-
> workflow-interference-free-design-electronic-devices.pdf
>
> 73,
> AG6CX
>
> Sent from my iPhone
>
> > On Mar 28, 2018, at 11:34 PM, Jim Brown <jim at audiosystemsgroup.com>
> wrote:
> >
> >> On 3/28/2018 11:22 PM, Jeff Blaine wrote:
> >> I was hoping someone may have found a "quite" switch that could be
> recommended.  Some of the semi-pro (e.g. Netgear home-office rack mount)
> stuff has metal cases -- but I don't know if that actually means they are
> quieter.
> >
> > Hi Jeff,
> >
> > I have no equipment recommendations, but can at least offer these
> thoughts.
> >
> > Unfortunately, shielding requires FAR more than putting circuitry in a
> conductive enclosure. Many such boxes are "clamshells" with paint
> insulating the places where they SHOULD meet, so they're not really
> shielded. Every cable SHIELD that enters a shielding enclosure must be
> bonded to the enclosure AT THE POINT OF ENTRY. Any unshielded conductors
> must go through feed-through caps that shunt RF to the chassis, but doing
> that with data circuits would likely trash the data. :) This problem can be
> solved by one or more ferrite common mode chokes on the cable.
> >
> > 73, Jim K9YC
> >
> >
> > _______________________________________________
> > RFI mailing list
> > RFI at contesting.com
> > http://lists.contesting.com/mailman/listinfo/rfi
> _______________________________________________
> RFI mailing list
> RFI at contesting.com
> http://lists.contesting.com/mailman/listinfo/rfi
>



-- 

*Dave - WØLEV*
*Just Let Darwin Work*