On Fri, 18 Jun 2004 04:44:59 -0400, Tom Rauch wrote:
> For the HF bands, Fair Rite #31 and #43 are the weapons of choice.
>
>Not true at all. If you use a core, you want one that has the highest
>possible impedance in the operating range. If you do not establish a
>filtering system with a bypass capacitor so you have known impedances in the
>system, you should use a core with very low Q. You want a core with a very
>low Q so the impedance is largely resistive.
Before you make this statement, you should measure chokes built with some of
these cores. I have done so, and in a disciplined way. In general, the
resistive component of the core's permeability (what Fair Rite calls u'')
causes the resistive component of the choke impedance to increase at lower
frequencies with a greater number of turns.
>Otherwise you can make the problem worse by adding series XL near the
>RF-floating
>consumer equipment terminating the line.
The obvious problem I have seen with XL is that the equipment DOES complete
the circuitto "ground" at RF and XL causes a relatively short cable to
resonate, increasing the current. Thus I am in agreement with your comment
that you want the choke to be more resistive than reactive. But multi-turn
chokes CAN have very significant resistive components at frequencies FAR
lower than single turn data would suggest.
As an example of this, the Ethernet switches in my home office throw out
strong birdies (S9) at several frequencies in the 80 meter, 30 meter, 20
meter, and 15 meter ham bands. I am able to achieve 2-6 S-units of
suppression of these components with multi-turn chokes wound around some big
clamp-on #31 ferrites. Too few turns will produce enough XL to increase
current, but a few more turns will reduce XL and increase R enough to provide
very effective suppression! [I'm looking at three receive antennas -- two of
them dipoles running parallel to the Ethernet cables about 30 ft above them,
and one a vertical less than 10 ft from one of the offending cables. Most of
the longer runs in my home office are in grounded conduit, but the shorter
runs are exposed.]
>You really want independent series impedances in each conductor to suppress
>differential mode at the same time it gets rid of common mode problems, and a
>simple cheap bypass cap (or better yet one on each conductor of the line to
>ground)
> makes the system predictable and much more effective than relying on the
> parasitic
> shunt capacitive reactance of the consumer device as part of the filter
> system.
Of course. If you are designing equipment or building a filter, that's
exactly the way to do it. But such a filter would have to be a dedicated
product (connectorization is an issue, it needs a package) and it needs
effective termination of the parallel capacitors. All of that increases the
cost (you've got to buy/stock one for each specific product/application).
Several points. My objective is "after the fact" suppression in the field
(that is, the equipment/system is in place and I'm trying to solve a
problem), not equipment design. Certainly differential mode and common mode
filtering (and shielding) should be part of the design. But when operating in
the "after the fact" mode, the last thing I want to do is modify equipment. I
would like to fix problems in the simplest possible manner, and with the
minimum of labor. In other words, parts cost is only part of the equation. I
don't want to reinvent the wheel, I don't want to turn each RFI problem into
a "science project."
Second, many products and systems have have common mode coupling mechanisms
components that are much more significant than their differential mode
mechanisms (or the differential mode mechanism is suppressed at the
frequency of the interference). The coupled RF will be the algebraic sum of
that coupled by all of those mechanisms. Conversely, there are products that
suppress common mode but ignore differential mode.
Third, most detection of RF in equipment is square law -- that is, the
strength of the interference increases as the square of the coupled energy.
If the suppression is able to reduce the current by 10 dB, it will reduce the
audible interference coupled by that current by 20 dB. That may be enough to
make the interference inaudible (or invisible).
Fourth, my research has clearly shown that common mode current induced by
antenna action on audio cables is a MAJOR cause of RF interference to real
world audio gear, and especially on the shields of those cables. I've also
shown that differential mode is also important, and you can't ignore either
one.
There is an excess of "mystique" in the specification and pricing of ferrite
materials and parts in the engineering community. Try to find pricing on any
of these products if you are not an OEM and want to buy 1,000 pieces. All I'm
able to find is single lot pricing from distributors like Newark and DigiKey
on the few parts they choose to put in their catalogs. Try to find detailed
technical data on any manufacturer's product other than Fair Rite if you're
not an OEM. I'm seeing VERY sharp engineers who don't have a clue how these
components behave. My objective is to break down this "mystique" and make
these products more useful components for the audio guy and ham working in
the field.
Jim Brown K9YC
_______________________________________________
RFI mailing list
RFI@contesting.com
http://lists.contesting.com/mailman/listinfo/rfi
|