[time-nuts] Noise and non-linear behaviour of ferrite transformers

[Stefan Heinzmann]

On 25.07.2014 20:07, Charles Steinmetz wrote:
> Stefan wrote:
>
>>> shield side of the coax goes to ground through a 0.1uF cap.
>>
>> For optimum results with respect to high RF frequencies, I'd expect
>> that you would need this cap to be annular, so that the cable can pass
>> through the middle, and the outside connects to the chassis hole all
>> around.
>
> The coax itself remains coaxial through the connector -- it is just the
> point on the shield that is bypassed to chassis that is not annular.
> The low impedance of the shield makes this non-critical up to
> frequencies where you should be using a waveguide anyway.  It is a
> well-known and proven technique for bypassing shields to chassis with no
> galvanic connection.  (In some cases, the capacitor is paralleled with a
> resistor of 10 ohms to 1k ohms -- this provides some DC/LF continuity
> while limiting the possible ground loop current to levels that
> [hopefully] don't cause noise problems.)
>
> Rarely is the cap as large as 0.1uF -- 0.01uF is most common, and 1nF is
> also quite common.  You want the cap to have low inductance (high first
> self-resonant frequency), and you need to keep the leads very short.

Well, yes, I've seen this done many times.

However, when making shield terminations, many would tell you that it is 
important to terminate the shield 360 degrees to the chassis in order to 
have the best effect. Surely, that also has to be true for RF 
frequencies when you choose to raise the near-DC impedance of the 
shield-to-chassis connection to combat a hum loop?

I'm just trying to take this to the logical conclusion, which would be 
an annular capacitor for the shield connection.

Cheers
Stefan