[RFI] just what everyone needs!

[Dale Svetanoff]

Andy, Cortland, and All,

I have pasted Cortland's comments to Andy into this message so that I can
address comments and issues to both of you without the time and bandwidth
of separate replies.  After all, your messages are certainly related.  

Since this system operates in plain text mode, I can't use color to denote
my comments, so please scroll down carefully and slowly.  I've isolated my
comments with the use of dashed separators.  

73, 
Dale Svetanoff, WA9ENA
Sr EMC Engineer
E-N-A Systems, LLC
Specializing in shielding applications, system grounding, and lightning
protection
- - - - - - - - - - - - - - - - - - - - -
First, Cortland's reply to Andy:
 
Andy,
 
The way I unravel this is to ask how a ferrite BEAD works. It 
concentrates magnetic fields along a wire -- adding inductance --and in 
a lossy medium too, where the changing field is partly transformed into 
heat. This works even on a single wire.
 
Now imagine the ferrite has very low cross section, but extends along 
the wire for some considerable distance. It will still work, but the 
volume of ferrite per unit length will be much less, and it will 
concentrate the field in each unit length much less (= lower inductance) 
as well. Given inductive reactance rises with frequency, the higher 
frequencies will be where a thin covering of ferrite material is more 
effective.
 
Will it filter in a coax? Maybe -- if permeability, loss and thickness 
are right, skin effect isolates the center and outer parts of the 
ferrite from each other, so they act like separate DM chokes. Otherwise 
the fields from shield and center conductor will pass through ferrite 
(with some loss) and cancel (or partly cancel) like a CM choke on a 
differential pair. Nothing is 100 percent and the change occurs at what, 
6 dB/octave?
- - - - - - - - - - - - - - - - - -
The issue here is NOT one of adding inductance to form a discrete filter. 
Rather, the "lossy material" is a bulk absorber of UHF and microwave
energy.  (Think of it as having a lot in common with that patch of lossy
stuff embedded in every bag of microwave popcorn.  That material has a very
high loss tangent at 2.45 GHz and so converts most of the impinging RF into
heat.)  That is what this lossy material in the wire spec was meant to do. 
- Dale
- - - - - - - - - - - - - - - - - - 
 
On a ferrite covered single wire in an RF field... If the skin depth is 
small enough, fields won't much penetrate the ferrite, some being 
dissipated, some flowing lengthwise (being dissipated) and the rest 
will be reflected -- shielding the wire without being a return circuit.
- - - - - - - - - - - - - - - - - -
Due to the lack of loss below 500 MHz or so, let's call the net effect
"partial shielding of the wire".  The ferrite material does little below
that point.  - Dale
- - - - - - - - - - - - - - - - - - 
 
Cortland Richmond
KA5S
 
_______________________________________________

 Now, Andy's reply comments to my post:

> [Original Message]
> From: Andy <AI.egrps1+contesting at gmail.com>
> To: <rfi at contesting.com>
> Date: 9/10/2012 9:46:05
> Subject: Re: [RFI] just what everyone needs!
>
> Dale Svetanoff wrote:
>
> > For starters, I strongly suggest a visit to the URL link that David,
K1TTT,
> > provided in an earlier posting on this topic.
>
> That link was what got me started on this question.
- - - - - - - - - - - - - - - - - - - - - - - -
Good!  - Dale
- - - - - - - - - - - - - - - - - - - - - - - -
>
> >  One look at the diagram in
> > the spec will tell you that no, this is not coax, although it might look
> > like coax.
>
> But it sure looks like coax to me.  In fact it is coaxial; the
> question being how conductive that "Filter Layer" between the
> insulating layers is.  And whether/how it would be connected
> externally.
- - - - - - - - - - - - - - - - - - - - - - - - 
I understand why the product looks like coax, but no, it is NOT coax. 
There is no concentric shield/conductor over the center conductor (wire) to
complete the circuit.  The lossy ("absorptive") material must not be
thought of as a shield - it's not.  See my next comment below.
  - Dale
- - - - - - - - - - - - - - - - - - - - - - - -
   
>
> >  It absorbs RF rather well (at 30 dB/ft) at 1 GHz and above.  It
> > does not need an outer shield to do that ...
>
> But, but ... the spec does require a shield to get those numbers, doesn't
it?
- - - - - - - - - - - - - - - - - - - - - - - - - - -
Ah, this is the key to my argument: NO, the spec does NOT require a shield
to get the 30 dB attenuation numbers at 1 GHz and above.  In fact, read the
spec carefully.  The required performance (given in a table) is to be met
with the wire "as-is".  As I said earlier, I believe that the main reason
to put a shield over the loaded wire is to minimize radiated emissions
below 500 MHz that may couple onto the wire within whatever equipment it is
associated with, or to reduce pick-up of RF in the vicinity of the wire
when run within some system environment.  Keep in mind that MIL-SPEC-461
(and its subsequent versions of A thru F) impose stringent radiated
emission limits on units and systems over a wide frequency range.  - Dale
- - - - - - - - - - - - - - - - - - - - - - - - - - -
>
> This is what got me wondering whether maybe the Filter Layer itself
> WAS the required shield.
- - - - - - - - - - - - - - - - - - - - - - - - - - -
See above.  Answer: NO, as the absorptive material is NOT a shield, it is
an attenuator (with defined loss per unit length, not in cross-section). 
Yes, a true shield is also an attenuator, but to a much greater extent (80
to >100 dB, typical), and over a very much wider frequency range.  Do not
regard the absorptive material as a filter - it is not so.  Again, it is an
absorber or attenuator that converts RF to heat.  Filters (comprised of
bulk or distributed components) use a combination of reflection and
absorption to achieve desired results.  - Dale
- - - - - - - - - - - - - - - - - - - - - - - - - - -
>
> > - the material itself can be
> > considered as a "dummy load" for any microwave RF that happens to be
> > traveling on the wire.
>
> This is where I don't follow, especially the "how".  One way to get
> attenuation is if it is like a coaxial xmsn-line and the Filter Layer is
> in the signal's return path, and imparts a frequency dependent loss.
> But that requires "grounding" (let's call it "terminating") the Filter
> Layer on both ends.
>
> Or perhaps the Filter Layer acts similarly to a ferrite bead (like the
> distributed equivalent of a bead) which is not in direct electrical
> contact, but its presence around the wire alters the magnetic field
> and changes the impedance.  I am guessing by your reply that perhaps
> this is closer to what is really going on here?
- - - - - - - - - - - - - - - - - - - - - - -
Again, do not think of the absorptive layer as a filter.  In this
application, think of that material as being a distributed dummy load for
incidental UHF and microwave RF riding on the wire.  I don't know if you
are familiar with anechoic or semi-anechoic chambers, but if not, let me
simplify things by saying that the vast majority of mil-spec testing is
done in shielded test chambers.  Those chambers all have conductive
surfaces (all 6 sides), and thus would have RF ping-ponging all over the
place and making accurate measurements of a test article's performance all
but impossible.  So, such chambers are fitted with absorbing materials on 5
sides ("semi-anechoic" - no absorber on the floor) or all 6 sides
("anechoic") to allow measurements in a defined and controlled EM
environment.  The materials used to absorb the stray RF and kill
reflections is usually a combination of special ferrites (as tiles applied
to the surfaces of the test chamber) and carbon-loaded plastic foam (in the
form of pyramid-shaped cones or truncated pyramids, and applied over the
ferrite tiles).  In other words, "dummy" loads for impinging RF.  The
absorptive layer in this wire spec is simply a special material that has a
lot of loss at and above 1 GHz.  I suspect that the main reason for low
performance at lower frequencies is the cross-section of the material is
not sufficient to provide much attenuation at longer wavelengths.  The
cones used to provide a load at frequencies of 100 MHz and below get very
large (2 feet or more in thickness).  - Dale
- - - - - - - - - - - - - - - - - - - - - - -
>
> >  The advantage of having a shield over the absorber
> > material is to further reduce any emissions that still occur since mil
> > specs require emission control over a very wide frequency range (from as
> > low as 10 kHz to 18 GHz in many instances).  At that lower end, the
shield
> > will do all of the attenuation work, as the absorber would be useless.
>
> But note that the spec indicates low attenuation at lower frequencies
> (max of 0.015 dB/ft at a MHz), so it sounds like this wire is meant to
> pass those frequencies.  Also, there is no mention in the spec of
> anything about emission control.  It is all about insertion loss.
- - - - - - - - - - - - - - - - - - - - - - - - - - - -
The "insertion loss" is the performance of the absorptive material, not the
wire!  Remember, this is a wire, not coax.  However, as a wire, I assume
that the intent was for it to carry DC and/or low frequency AC signals for
control or similar purposes.  RF would be carried on standard coax.  - Dale
- - - - - - - - - - - - - - - - - - - - - - - - - - - -
>
> No doubt the need for control of broadband emissions is true.  I just
> don't think the special characteristics of this wire were created for
> that purpose.
- - - - - - - - - - - - - - - - - - - - - - - - - -  - - 
Go do a read on MIL-SPEC-461F and you'll see what I mean.  Look at the
radiated emissions requirements.  - Dale
- - - - - - - - - - - - - - - - - - - - - - - - - - - - 
>
> > Specifying performance by the foot is a way to "normalize" the required
> > spec ...
>
> Yeah I was aware of that.  I just used one foot as an example.
>
> > ...  Since this is single conductor wire (see the spec!),
> > it was not meant to carry GHz RF.
>
> But it clearly WAS meant to be used in situations where its ability to
> block >GHz signals is important.  Whether or not it was supposed to
> carry actual GHz signals, the whole idea here is how well it attenuates
them.
>
> I have a hard time imagining using any practical length of hookup
> wire, where measured performance at >GHz is important.
>
> Thank you for your informative comments.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
You are welcome.  I hope these additional comments help to better explain
the situation and application.  - Dale
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
>
> Regards,
> Andy
> _______________________________________________
> RFI mailing list
> RFI at contesting.com
> http://lists.contesting.com/mailman/listinfo/rfi