Yes, but the problem is that regular thin sheets of metal such as steel are
not a particularly good approximation of a 'perfect conductor' at 60Hz.
This is related to the skin depth (which I seem to recall is ~8mm in
copper -- a much better conductor than steel -- at 60Hz?) -- if the skin
depth is large compared to the condutors thickness, you can't build a very
good Faraday cage out of it.
Conduction skin depth is not the consideration. Magnetic properties
are. 8mm of steel is NOT ever needed. That is utterly ridiculous.
I made three shields last week. One was from a simple cookie sheet,
which I then laminated with transformer tape for isolation. It
reduced the field in question by a factor of 5. I then made one from
18 Ga steel, and it went to 6. We then had one fabricated from .032"
steel, and it ha hard tie points to mae the sides mate more intimately
with the sides of the chassis. That got us to seven.
That would be 15mV of ripple on a 1000 volt supply at .1 volt
regulation over 250 Watts. It will power 450 PMTs at one time, very
accurately. It started t over 100mV when we placed the circuitry in
the case. My shields made the difference, and the engineering
director (my boss) couldn't believe that I achieved that much with a
simple shield after we chased after this thing for two days.
I realized that the noise was much less out of the case, and the
only thing that changed was the proximity of the amp, and control
boards. I reached in, and picked up the Ctrl board, and raised it up
a quarter inch, and the noise cut in half. That is when I knew a
simple shield was all we needed. My first shield was a mere backplane
for the ctrl board. The two subsequent shields were full partitions
for the case. Big difference.
Sure, a Mu metal chassis and partition would be better, but one has
to shoot the engineers, and get on with production t some point. That
point would be the point at which cost of manufacture and meeting
customer spec are optimized. Mu metal cases and shields are not in
that equation. If they wanted a million pieces.. maybe. But for 500
a year... no thanks.
Faraday cages are typically for keeping rf noise inside a device
chassis, and for keeping external rf influences and lf influences from
"getting in".
We're also shielding magnetic field, not electrostatic field, so
conduction is not an issue. A complete magnetic circuit, however is.
That is why the shield I made with tie points worked better, despite
being thinner than the 18Ga sheet was. I had a complete magnetic
circuit.
Steel is fine for many purposes in this regard.
Nearly all rack mount chassis made for mil use are NOT Mu metal, and
meet all mil specs for EM shielding. Why would that be were it not
sufficient enough at attenuating EM fields?
This is why people are always suggesting mu
metal -- its relative permittivity is on the order of thousands so the
requisite thickness of metal required is reduced by an order of magnitude or
so.
It is also about magnetic fields, not conduction properties.
It's unfortunate that the price is quite high and performance can
degrade remarkably due to mechanical stresses (bending, hammering, drilling,
etc.).
The order of magnitude more shielding achieved far outweighs any
"losses" in original spec incurred by deformation(s) of the original
sheet.
However... cost IS an issue, unless one is a huge conglom like Sony
or such, where incorporating the best is not a big impact on the
company's operational costs.
For us... we need NRE and development funding for such ventures.
Or a customer that actually pays for the quality we make, instead of
trying to think in chinese mass production pricing scheme numbers.
We do make the best, lowest noise HV & EHV supplies in the world,
though.
In radiology, for instance A lower noise supply means a higher
contrast ratio in the imagery. I have a 4kV supply that is at 2mV
ripple through its entire range of operation. That is like 0.00005%
A shielded multiplier (a mere partition), and a HV coaxial output
got us there, down from 11mV previously.
I AM the noise abatement crew! Heheheh...