Topband: Filters and grounding

[Jim Brown]

On Tue, 15 Jan 2008 04:32:04 -0500, Tom Rauch wrote:

>Raoul,

>Gary alluded to BC work. In well-designed BC station studios 
>the audio lines are all balanced shielded lines. The normal 
>procedure, and the proper procedure, is to ground the shield 
>at one end only. This ground normally is at the input end of 
>the equipment since that is the most sensitive spot for RF 
>ingress.

That is the "old" way. Bill Whitlock published an AES paper (JAES 
Jun 1995) that convinced everyone who read it that the proper 
shield connection for a "one end only" bond is at the SENDING end, 
not the receiving end. Further, virtually every audio authority is 
convinced that the shields of balanced audio cables ought to be 
one end only bonded at AUDIO frequencies, and BOTH ends bonded at 
RF. What Ott calls "hybrid" grounding.

However, BALANCED shielded cables are VERY different from coax 
with respect to the behavior of the shield. With balanced cables, 
the shield is effective only against the electric field, NOT a 
magnetic field. With coax, the shield is not really a shield, it 
is one conductor of a transmission line, and the rejection of 
external fields comes from the mutual coupling between the 
"shield" and the center conductor. With twisted pair balanced 
cables, the primary rejection mechanisms are mutual coupling and 
twisting, with twisting being the most powerful. 


>Large RF and AC differentials between equipment or rooms is 
>prevented by using a wide copper flashing that bonds 
>equipment cases together. In a BC environment with proper 
>balanced line installations the bonding of case does, as 
>Gary suggested, help the situation. But this is because the 
>cable shield is intentionally open at one end.

I disagree. With wide flashing, we lower the resistance a lot and 
the inductance a bit. 

The primary reason that balanced audio cable shields must be 
opened at one end is THE PIN 1 PROBLEM. I STRONGLY suspect a pin 1 
problem in the equipment being discussed here (that is, where an 
EARTH connection for the filter improves things). In this case, 
common mode current on the shield is likely flowing into the 
equipment via a pin 1 problem at the connector where that antenna 
is connected. I have found pin 1 problems in virtually every piece 
of ham gear I have ever opened. For more on this, see the 
publications section of my website. Pin 1 problems are a PRIMARY 
cause of baseband (hum, buzz) and RFI. MOST of the problems we 
call "RF in the shack" have pin 1 problems as their fundamental 
means of ingress (and poor antenna design often puts more RF in 
the shack). 

http://audiosystemsgroup.com/RFI-Ham.pdf

http://audiosystemsgroup.com/publish

>Actually any balanced line filter in a line with no, poor, 
>or open ended shield integrity, say for example a power line 
>filter with no shield or an audio cable shield that is open 
>at one end,  needs a bonding ground or a ground path.

More correctly, the "ground" connection needs to be to the 
shielding enclosure of the equipment being filtered, or in the 
absence of a shielding enclosure, the reference plane. 

Pin 1 problems can be "band-aided" by killing the common mode 
current with a good common mode choke at the point where the cable 
enters the equipment, or by bonding the shield to the shielding 
enclosure.

73,

Jim Brown K9YC