[RFI] RF & Ferrites,

[Jim Brown]

On 2/25/2020 11:03 AM, David Eckhardt wrote:
> Considering multiple turns through a ferrite 'bead':  Yes, more turns will
> yield better isolation.  However, much beyond 3 or 4 turns, and one
> approaches a limit of effectiveness.  Cram as many turns as is possible on
> a single bead increases capacitive coupling between the turns and
> ultimately limits the effectiveness.  The windings should be 'ordered' and
> not wound in a hap-hazard manner which also reduces inter-winding
> capacitance.

This is VERY oversimplified. One look at the impedance plot for any 
Fair-Rite core on their data sheet clearly shows that a wire passing 
through a core, and multiple turns through a core, is a parallel 
resonant circuit.

https://www.fair-rite.com/product/round-cable-emi-suppression-cores-2631480002/ 


k9yc.com/RFI-Ham.pdf  addresses this, and about ten years ago, these 
concepts were added to the ARRL Handbook.
> 
> Best practice:  Employ shielded cables with the shield terminated on both
> ends. 

YES, and HOW it is terminated matters a LOT. To function, the shield 
must be terminated to the shielding enclosure at the point of entry. A 
common construction error is to terminate the shield to a circuit board 
rather than to the chassis, with the "return" bus or layer eventually 
finding its way to the chassis. This error, first observed by a ham 
working in pro audio, called it "The Pin One Problem" because pin 1 of 
XLR connectors is the cable shield, and in a landmark paper, observed 
that it was a major cause of hum, buzz, and RFI. The mechanism is that 
any current flowing on the cable shield wanders around the "ground" bus 
until it eventually finds the chassis, with the resulting IR drop 
injected into signal paths at the whim of the PC board layout artist.

Sadly, nearly all equipment is built with Pin One Problems. The last 
time I was at Dayton, about four years ago, I looked at the rigs in 
every booth, and EVERY rig had Pin One Problems at their audio and 
control connectors. The Elecraft 2M and 220 MHz transverters I bought 20 
years have their BNC connectors insulated from the shielding enclosure, 
and both have stability problems.

Indeed, shielding is useless if the shield is not continuous, and is NOT 
terminated to the shielding enclosure at both ends.

CAUTION:  Many of the modern "shielded cables" may be rather
> compromised due to cost cutting regarding integrity of the shield
> installation and type.  When choosing a shielded cable, its best to test a
> reasonable length before buying a significant amount of the cable.

Sadly, it has long been common practice by these cost-cutters to use a 
single wire in place of what the cable definition calls for a shield. We 
first saw this in audio cables at least 25 years ago.

73, Jim K9YC