[RFI] SolarEdge Finds New Source of Solar System RFI

[Jim Brown]

I led the writing group for On 6/3/2019 3:54 PM, David Eckhardt wrote:
> One additional item I should mention:  The low-frequency designers 
> tend to stick with "grounding" (returning) the shield at only one end 
> (to prevent "ground" loops).  The high-frequency designers know this 
> will enable the shield to become a wonderful antenna and they better 
> "ground"  (return!!!   !!!   !!!) both ends of the shield at the 
> penetration of the 'cage' - the shield should be viewed as an 
> extension of the internal gutts of the Faraday Cage design 
> philosophy.  If both ends of the shield are "grounded" and there is 
> still a radiation / susceptibility problem, the shield is too thin 
> and/or not of properly covering the shielded conductors. 

I led the writing group for AES48, AES54-1, AES54-2, and AES54-3, which 
specifies good engineering practice for balanced audio 
interconnections.  Everyone in the writing group was fully aware of the 
issues you raise. There ARE good reasons for lifting the shield 
connection at one end, and by correctly analyzing the balanced audio 
interface as a classic Wheatstone bridge, Bill Whitlock proved that it 
should never be lifted at the sending end. His work subsequently caused 
IEC Standards with respect to the balance of such circuits to be revised.

The Standard (AES4-1) specifies that, and recommends a capacitor with 
good properties at HF through UHF.  The most commonly used connector for 
shielded twisted pair audio cables is the XL3, of which Neutrik and 
Switchcraft are the principal manufacturers. I came up with the idea of 
a cylindrical capacitor terminating the shield to the connector shell, 
with a ferrite bead around the designated shield contact, Pin 1. The 
bead serves two purposes. First, it lowers the Q of the parallel 
resonance between the capacitor and the inductance of the connection via 
Pin One to the shielding enclosure.  Second, it blocks the flow of 
VHF/UHF shield current into equipment that is non-compliant with AES48 
("the Pin One Problem" Standard, which specifies terminating Pin One to 
the shielding enclosure).   We recommended capacitance in the range of 
10-50 nF.

Neutrik, an active participant in AES Standards activity, developed and 
manufactured a practical implementation based on our concept.  At the 
time of publication of these Standards, existing products experienced 
serious RFI because they were not AES48 compliant, and I consulted with 
two major mic mfrs on the solutions to their problems.

In part as a result of this work (we all did lots of other things), 
Muncy, Whitlock, and I were subsequently elected Fellows of the AES, 
along with another important contributor, Ray Rayburn.

73, Jim K9YC