Topband: Filters and grounding

[Tom Rauch]

> Good point,  I had to use one here when I had a 1KW on 
> 870KHZ in the
> next room from me,  thank goodness its gone.   I had to 
> mount the filter
> on the copper strap that runs across the desk to ground, 
> just putting it
> in line was not good enough,

 Perhaps it is worth beating this horse a little.

I know there is a popular myth, partially reinforced by 
improper instructions and articles, that filters need to be 
"grounded".

Filters, if the cabling is good and the shield have 
integrity, never have to be grounded.

Think about how the cable works and what the filter does in 
a coaxial system. The filter adds a high series impedance 
with the center conductor, and in the filter a shunt from 
the center to the shield. We all know (or should know) the 
only signal the receiver responds to is the differential 
signal between the center and the shield. We are in very 
deep trouble if the receiver responds to common mode, which 
would be the signal outside the shield. So once the filter 
does its job it does it exactly the same if the case is 
grounded or not.

As a matter of fact all my working life I've tested filters 
and traps without ever worrying about grounding the case. If 
the case is grounded or not, they all work the same. They 
have the same attenuation. Always.

The only exception to this is if there is a resistive or 
open shield someplace in the system between the measurement 
point or detector and the filter. Even an open on the source 
side of the filter doesn't change the notch, although it can 
cause ripple in the source because of the impedance 
mismatch.

If you have a filter that is "improved" by grounding 
someplace in that system there must be a poor connection 
between the filter case and the receiver. That's the way it 
has to work.

73 Tom