Topband: Mother of all ferrite common-mode coaxial chokes

[Jim Brown]

On 7/13/2012 2:04 PM, Paul Christensen wrote:
>> One of the dumbest things I've seen recently from a very good company is
>> RF chokes in series with the shield connections on analog and RS232 I/O
>> boards for the Elecraft K3.
> Right.  I think Joe, W4TV, was one of the first to identify that problem
> back around 2008.

I think I was the first. Joe and I discussed the K3 a lot.

>> Modern pro output stages have a very low source Z (typically
>> 100 ohms for line level) and high input Z (typically 10K for line
>> level), and consumer stuff is roughly 3-5X those values.
> Somewhere in my files from the late '80s is a white paper authored by
> Richard Cabot.  I believe Richard was the chief designer of the Audio
> Precision brand of audio test equipment.  The focus of the article was on
> the standarding of all audio output stages, balanced or unbalanced to a
> value of 40-50 ohms.  He created models showing the effect of changing drive
> Z from 1-ohm through 600-ohms into long audio cables (and independent of
> terminating Z) that start to take on transmission line qualities.  His
> conclusion was that a target of just under 50 ohms was optimum.

Dick is a sharp guy. The best work I've seen on this topic was by the 
late Deane Jensen, c.a. 1980, showing that the capacitive loading of a 
long cable could make an output stage unstable, and that 100 ohms was a 
good value to prevent that from happening. That's been pretty much 
accepted as definitive, which is why most pro output stages are in that 
range. This can, of course, be device-dependent, and the lower value 
Dick suggested would be fine if the output device was unconditionally 
stable for all loads.  It's VERY common in large installations for there 
to be 500-1,000 ft of cable bridging an output stage, and commonly used 
balanced audio cables range from as much as 40 pF/Ft (older stuff like 
Belden 8451) to as little as 13 pF/Ft for cable rated for AES3.

That said, a cable would have to be VERY VERY VERY long to require 
transmission line analysis at audio frequencies, and applying 
transmission line analysis is VERY complex, because Zo varies over two 
orders of magnitude through the audio spectrum, converging to 75-100 
ohms at HF for practical audio cables. See

http://audiosystemsgroup.com/TransLines-LowFreq.pdf

73, Jim K9YC