Pete,
See inserted comments below. Hope this helps.
Keith
KB6B
Pete Smith wrote:
> After 48 years as a ham, there are still many big holes in my technical
> understanding of things. One of the biggest is shielding -- elementary
> questions like:
>
> Does a shield box for RF need to be grounded? What about a shield against
> magnetic fields?
The purpose of a shield is to "short circuit" the radiated fields either
emanating from or coupling to whatever you're trying to shield. In order to
work, the shield must be an electrically continuous surface, i.e., a faraday
cage. In practice, discontinuities which are very small in terms of
wavelength at the lowest frequency you're interested in shielding against are
generally not a problem. One rule of thumb is that any discontinuities in a
shield should be smaller that 0.01 wavelength.
If you have wires penetrating the shield, RF currents can flow in & out of the
shield & degrade the performance. One approach is to make the box and any
interconnecting cable shields a continuous faraday cage. In practice, this is
sometimes unworkable, and shielding just the areas most prone to field-to-wire
coupling may help. Sometimes what appears to be a "shield" actually works by
providing a return path for noise currents back to the source. An example of
this would be a shielded I/O cable connected to an unshielded (or poorly
shielded) computer. The cable shield actually provides a return path. There
are some pretty complex interactions involved here, and sometimes the best
approach is just to try it and see what works and what doesn't.
A connection to the physical earth would have no effect of the shielding
properties of a faraday cage. What you need is a low-impedance connection
between the shield and the circuit you're trying to protect, in order to
provide a "short circuit" path for fields coupling to or from the circuit.
What you really want to do with "grounding" is to maintain everything at a
common RF potential, which isn't going to happen at HF with a practical
connection to the physical earth. Sometimes connecting a "shield" to the
station ground reduces interference by providing a return path or moving a
resonance. The physical earth is not a sink into which we can dump noise
currents. Current only flows in loops, and in order for current to flow in a
ground lead there must be a complete loop somewhere. Remember that part of
this loop can be an electromagnetic field.
>
>
> Should high-impedance audio circuits always be installed in a metal box to
> prevent hum pickup?
Not necessarily. Most shielding materials have very little effect at 60 Hz.
What generally works better for high-impedance circuits is to control low
frequency coupling by controlling circuit geometry, i.e., twisted pair
interconnect cables and balanced interfaces with a high common-mode rejection
ratio. Preventing ground loops is also important for minimizing coupling at
low frequencies.
>
>
> Does the material of a shield box make any difference as long as it is a
> good conductor?
In general, no. If you really need low frequency shielding, there are special
high permeability materials (such as mu-metal) that make better shields
against the magnetic portion of the EM wave. Shielding material selection in
these cases is also a function of the circuit impedance. Low impedance
(current loop) circuits tend to radiate and respond to magnetic fields, i.e.,
are inductively coupled. High impedance circuits tend to radiate and respond
to the electric field, i.e., are capacitively coupled.
At RF, what's more important than the specific shielding material is maintaing
good electrical bonding across the shield joints and minimzing
discontinuities.
>
>
> Should signal grounds in a computer cable be grounded to the shield box at
> each end, or deliberately left to float?
Rule of thumb: If the cable is less than one twentyth of a wavelength long at
the highest frequency you're interested in shielding, terminate the shield at
one end only. If the cable is longer than this, terminate the shield at both
ends. Shield termination is also important...remember you're trying to
provide a low impedance connection to the circuit you're trying to protect.
>
>
> I have looked for a coherent discussion of shielding against various types
> of signal ingress, without success. There are passing references in the
> Handbook under various other headings, but typically they seem to assume
> knowledge I just don't have.
>
> Can anyone suggest a good, entry-level discussion of the subject anywhere?
If you don't mind slogging through a lot of government-ese, MIL-HDBK-419 and
MIL-STD-188-124 both contain extensive discussions on bonding, grounding and
shielding. They're large (multi hundred page) documents, but can be
downloaded in PDF format from several sources on the web if you have a
reasonably fast internet connection or lots of time. Try
.http://assist2.daps.dla.mil/quicksearch/ . Type the spec number above into
the "document ID" field. Once the search has returned the document
description, you'll go through several screens before you get to the one to
download the actaul .pdf image.
>
>
> Thanks,
>
> 73, Pete N4ZR
>
> 73, Pete N4ZR
> Sometimes a tower is just a tower
>
> _______________________________________________
> RFI mailing list
> RFI@contesting.com
> http://lists.contesting.com/mailman/listinfo/rfi
|