[RFI] re: It's the old story of peeling the onion --

David Garnier dgarnier at wi.rr.com
Sun Dec 5 17:11:37 EST 2004


Jim,

Bravo for bring up this concept, it's a great visual way of
illustrating the problem, (we repeating this same refrain
when project engineers or management declares victory
to soon on the OATS.)  ;-)

I have been reluctant to jump into this fray of personal
experience, opinion or standard emi compliance debug
practice, but here goes.

I am an engineering technician who has 10 + years of emc
debug and compliance experience for an major medical
electronics company in the midwest.  I have a couple
comments and observations that maybe of some benefit...

1) If you have a couple of these wall-warts or external
switching power supplies plugged into the same outlet.
Rule #1: "Broad-band noise (BB) is usually additive."

2) The wall-warts & switching power supplies (that I
have taken apart) are NOT internally encased in a metal
box! Given that fact, do not lay them over nearly cables.
Rule #2: "You will (take your pick) magnetically or capacitively
coupling the EMI (BB) energy unto other cables or your
metal desk top!

3) I like the idea of getting rid of these switching supplies
and using an external linear supply, but you need to filter
the digital noise before you connect it to the linear supply.

4) We have discovered at work that the plastic clamp on
EMI ferrite beads (FB), the plastic soon fatigues and the
ferrite halves don't make connection anymore!  Solution
is to place a Tyrap on each ferrite bead.

5) If all else fails, don't operate on multiples of crystal
harmonics!!!  This is a well known problem within the EME
moon bounce community (due to antenna & preamp gain.)

73's & GL de wb9own

dave garnier





Message: 3
Date: Sat, 04 Dec 2004 11:55:00 -0600
From: "Jim Brown" <jim at audiosystemsgroup.com>
Subject: Re: [RFI] routers & wall warts
To: "rfi at contesting.com" <rfi at contesting.com>
Message-ID: <20041204175501.7793E7D44 at gw1.nlenet.net>
Content-Type: text/plain; charset="us-ascii"

On Sat, 4 Dec 2004 12:20:17 -0500, Jim Jarvis wrote:

>So, I had concluded that it was 10baseT ethernet signals radiating
which
>were the source of the problem.

Your conclusions are in agreement with mine that this is the dominant
mechanism
for the best behaved products.  Some products may also radiate directly
due to a
combination of poor circuit layout and poor shielding, and, as noted,
the power
supply can be a bad guy.

It's the old story of peeling the onion -- you get rid of one bunch of
noise, and you
now hear what is lurking below it, so the search begins anew.





More information about the RFI mailing list