> As usual, an attempt is under way to reduce the solution to a complex
problem
> to a one-liner. The closest I can get to that is : Stainless, or any
material
> with a conductivity significantly less than Copper, should not be used in
any
> current-carrying paths within the amplifier. Knowing how to apply this
rule
Apply it by considering the current density, or the total current flowing in
a given cross-section of material.
Contrary to all the notions about magnetic and non-magnetic, the only effect
of anything containing iron is a slightly shallower skin depth than non-iron
materials. There are no magnetic effects for RF, because eddy currents
create a counter magnetic field. You car, when it pulls over the VLF field
of traffic sensing coils at a light, causes the inductance (frequency) to
decrease. If it didn't work that way, people wouldn't waste so much time
slicing iron into thin insulated layers for transformers at 60Hz.
I used to believe all the folklore about brass screws and non-magnetic
materials until I started testing things both ways.
> But I came bearing gifts. One comment passed by here which I would like to
> offer some response to, in the form of hard data. It relates to the effect
of
> nearby conductive surfaces (i.e. metal sheets or structures) on the Q of
tank
> circuits. Like the writers of those comments, I have always tacitly
accepted
> this advice about keeping the coil away from the cabinet. Now the issue
seems
> to be further fragmenting into a consideration of whether that wall is
highly
> conductive, poorly conductive, ferromagagnetic, etc. I believe that some
> tabled data on this exists somewhere, but I can't put my finger on it
right
> now.
I've built HF amplifiers with everything from copper to steel, and never
noticed or measured the slightest change in performance one way or another.
VHF amplifiers, where the tank area becomes a major part of the tank system,
are another matter. Silver-plated copper or thick silver plated brass is
best there.
> Generally speaking, the degradation in Q was not even noticeable until the
> sheet got a distance from the axis of the coil about equal to the coil
> diameter (actually, slightly less, but let's not split hairs.) The Q then
> started changing, more rapidly as we got closer. In order to drop the Q
by
> 50%, the sheet had to get quite close, like under 1/8 inch. This pattern
was
> the same for ALL metals which I had available for the test: Copper, Brass,
> Steel, Manganin, Phosphor Bronze, Stainless, Aluminum, and Kryptonite.
Only
> with the sheet laying right up against the coil could I detect the effect
of
> the material conductivity. The greatest Q-reduction was, as expected, due
to
> the highest resistance material (Stainless). But in order to see it
produce a
> serious perturbation, I had to create a pathological situation.
The largest effect of sheetmetal is the reduction in inductance from eddy
currents and the resulting counter magnetic field, and an increase in stray
capacitance shunting the coil. The first tends to move EFFECTIVE inductance
downwards, the second increases EFFECTIVE inductance. Both combine to cause
circulating current paths to increase in length while increasing the amount
of circulating current. This is the root cause of Q is reduction. Losses in
the conductor making up the coil generally increase the largest amount, but
of course there are some small additional loss incurred in the sheetmetal
circulating currents.
What you measured is how it works. Steel or other metals with high RF
resistivity need to be VERY close to the coil axis to cause a noticeable
increase in loss due to the circulating currents induced in the external
materials. Unless you are jam-up against the metal and unless the metal
barrier is large (a steel 10-24 screw in a two inch coil field certainly
won't do it), it makes no difference at all if it is a conductor with the
slightly elevated RF resistance due to reduced skin depth from iron content.
I wouldn't hesitate for a second using stainless in any design with even
remotely reasonable component layout at HF.
I would, however, avoid it like the plague when it is not plated with a
conductive layer for VHF or especially UHF applications where striplines or
cavities use the walls as part of a system with standing waves.
73 Tom
|