(Fwd) RE: [RFI] Common mode
Sat, 20 Nov 1999 03:56:21 -0800
Let me see if I can help a little. I'll insert comments after your
Juan, EA5RS wrote:
> going thru the RFI reflector archives, you seem to be one of the
> persons with the highest expertise in EMI.
> Could you perhaps shed some light on my doubts?
> I am really lost.
> Thanks and regards,
> Juan EA5RS
> ------- Forwarded Message Follows -------
> From: Self <Single-user mode>
> To: "EDWARDS, EDDIE J" <firstname.lastname@example.org>
> Subject: RE: [RFI] Common mode
> Cc: "'email@example.com'" <firstname.lastname@example.org>
> Reply-to: email@example.com
> Date: Sat, 20 Nov 1999 14:21:09
> thank you very much for your reply. Let me make a few questions as I
> may be confusing common mode with coax shield carrying RF current.
"Common-mode" is the net current flowing in one direction in a
transmission line, regardless of what kind of transmission line it is,
and is the algebraic sum of all currents flowing in the transmission
line. "Differential mode" current is that portion of the overall
currents that flow in opposite directions. If equal and opposite
currents flow, then the algebraic sum is zero and we have no common-mode
current. If current flow in one direction is not equal to the other,
then the difference is the common-mode component.
> 1. I understand that in a coax fed dipole, you actually have three
> different conductors. The coax inner conductor carrying a
> differential current vs the inner side of the coax braid (due to skin
> effect). The inner conductor cannot pick-up external EM fields as it
> is shielded within the coax braid.
IF, and this is a big IF, one had a perfect coaxial transmission line,
then this might be a reasonable analogy. However, the outer conductor
of a real coax is definitely NOT a perfect shield. Because flexible
coax shields are woven, there is a phenomenon called "porpoising" where
currents induced on the outer shield are carried to the inner surface of
the shield, and hence superimposed on the intended signal, when the
woven shield wires cross under each other. Note that this same
phenomenon also carries part of the intended signal to the outside of
the shield, resulting in RF leakage from the coax. Typical coax cables
do real well to have better than something on the order of 80 dB of
shielding. In addition, at broadcast frequencies, the skin depth might
be approaching the thickness of the braid, which would reduce the
shielding effectiveness even further...I'll have to go look up the
Semi-rigid or rigid coax with a solid outer conductor does somewhat
better in this regard, but there is always still some leakage around
connectors. For most purposes, we can ignore leakage through connectors
(characterized by a measurement call "transfer impedance"), and
connector manufacturers don't normally supply this data for RF
connectors. However, in this case connector leakage may be a factor.
> 2. This differential signal is the only signal I would expect to be
> amplified in the receiver and a very good filter has been fitted for
> this purpose (60+ dB notch at BC freq).
Are you sure this is enough? 60 dB down from a large signal may still
leave energy to saturate the front end of your receiver. HF receivers
typically don't have a lot of selectivity ahead of the first RF amp or
mixer stage (most of the selectivity is at IF), and a strong out of band
signal may simply overwhelm the front end.
> 3. The third conductor is the outer of the coax braid. On one end
> this conductor is connected to a leg of the dipole (no baluns or
> chokes used) and the other end to the rcvr chassis which in turn is
> connected to ground thru a length of wire having some inductance.
Ideally, equal and opposite currents would flow in the dipole, which
means that equal and opposite currents would be introduced into the
coax. Again, in most circumstances, it doesn't really matter very much
whether or not the antenna is really all that well balanced, but in your
case where the source is only a short distance away the geometry of the
antenna may be such that currents induced in the two legs by the
broadcast station are not really balanced, resulting in additional
common-mode currents in the feedline.
> 4. Being "suspended" in a very strong EM field, moderate currents are
> induced in this outer side of the braid. I am not sure whether this
> is a common mode current. This current has to return to the BC
> antenna somehow, and this would be thru ground and any other
> mechanism from the dipole leg? Any transfer of this current into the
> coax differential mode currents?
Yes, this is common mode current. Think of your dipole and feedline as
a single random wire antenna, with the feedpoint at your radio.
Currents will indeed be introduced into this wire. Depending on the
electrical length of the combination of the antenna, feedline and ground
circuits (including the AC fault protection ground and AC wiring), you
may have standing waves, and the radio may be at a current or voltage
> When the ground/earth connection of the rcvr is removed, the chassis
> becomes hot with RF. This disapears when the antenna is disconnected.
> I am not sure whether only the coax braid is hot, or both inner and
> braid are hot with RF (probably both).
See comment above about dipole/feedline acting as a random-length wire
antenna. This is exactly correct.
> But I always thought that the fact that the chassis (common of al RF
> circuits) were at a given RF potential above ground should be no
> problem whatsoever, unless the good model is the one that says that
> the differential mode voltages from the dipole antenna were balanced
> (symetric respect to ground). Sorry, I donīt understand.
If the radio chassis were in fact a perfect RF conductor and the case in
fact a perfect RF shield, what you thought would be true. What happens
is you have BC signal currents flowing from the coax outer braid through
the radio chassis to other conductors (power, ground leads, etc.), and
because the chassis/cabinet is not an RF shield, there are potential
differences (RF voltages) generated within the chassis/cabinet which can
capacitively couple to the internal circuits, or superimpose RF voltages
on wiring connected to different parts of the chassis/cabinet. Also,
the power supply connection is in all probability not a transmission
line, and RF voltages can enter the front end via the power supply
conductors, which penetrate the radio chassis. In addition, the
impedance of the RF path between internal portions of the radio is
typically not very well controlled, and there is no guarantee that
currents will stay balanced inside the radio.
Since we are dealing with unintentional effects and a fairly complex
geometry, I have never seen a computer model capable of dealing with
this kind of a problem.
> So again, I donīt know how the BC signal is going around the notch
> filter and being amplified (saturates) the rcvr.
> Can you perhaps further clarify my ideas, where on the above am I
> right or wrong?
I think there are a couple areas to look at further. First, you may
actually be getting RF coupling past the notch filter, either because 60
dB is not enough to adequately attenuate the signal or because of
leakage through the coaxial cables/connectors between the filter and the
What I think is most likely, however, is that you have a common-mode
current problem which is saturating the front end of your receiver,
because the receiver is not really an RF shield or at a single potential
inside. There are two general ways to fix common-mode problems:
1. Introduce a high impedance in the RF path, i.e., a common-mode
choke. In this case, I don't think this is a realistic approach because
of the large amount of inductance required to have a significant effect
at broadcast frequencies.
2. Break the common mode path.
Breaking the common-mode path can be accomplished by inserting an RF
transformer in series with the transmitter output (remember the old TVI
trick of inserting back-to-back 75 ohm to 300 ohm baluns in the coaxial
feedline to break the common-mode path). In this case, a broadband 1:1
RF transformer inserted between the radio and the low pass filter would
probably help a lot. Off the top of my head, I suspect an air-core
transformer would have better common-mode rejection than a toriod core
transformer, but either one should be worth a try. Make sure that there
is NO direct connection between the input and output, i.e., don't try to
install it in a metal case. The idea to break the coax for common-mode
BC signals, which still providing a coupling path for the intended
differential-mode signals. Don't worry about the coax being
"unbalanced" while the transformer is "balanced"; the transformer could
care less whether its input/output voltages are symmetrical or offset
with respect to some arbitrary "ground".
Finally, don't forget that any conductor connected to the radio can be
an antenna, including mic cords, headphones, keyer cords, power lines,
etc., etc., etc.. Once you have done as much as you can to break the
common-mode path from your antenna, you can try disconnecting other
conductors one at a time to see if it helps.
Good luck. This is a difficult situation.
> Thanks for your time, Ed.
> > From: "EDWARDS, EDDIE J" <firstname.lastname@example.org>
> > To: "'Tim Duffy'" <email@example.com>, firstname.lastname@example.org
> > Cc: "'email@example.com'" <firstname.lastname@example.org>
> > Subject: RE: [RFI] Common mode
> > Date: Fri, 19 Nov 1999 17:17:42 -0600
> > > -----Original Message-----
> > > From: Tim Duffy [SMTP:email@example.com]
> > > ------- Forwarded Message Follows -------
> > > From: Self <Single-user mode>
> > > Reply-to: firstname.lastname@example.org
> > >
> > > Can anyone provide a model of how common-mode signals enter and are
> > > amplified on a radio receiver via the coax shield of the antenna?
> > >
> > [Ed-K0iL] Simple, both conducters of a coax (shield and center
> > conductor) conduct the signal in the same "common" direction at the same
> > time interval as if it were one single conductor. The RF then enters the
> > radio chassis were it can be radiated into and onto several circuits inside.
> > You can get front-end overload or audio rectification at this point.
> > When this happens on a Stereo rcvr/amp, the speaker wires become
> > common mode antennas and (both wires)conduct the RF into the Amplifier were
> > it is feed back into the output and input of the amplifier circuit via the
> > audio amp feedback circuit. Then it is detected by the P-N junctions in the
> > amplifier's transistors. The detected audio is feddback and amplified by
> > the amplifier as designed. The detected audio is mixed with the intended
> > audio and is heard from the loudspeakers.
> > > This seems to be the entry mode of a VY STRONG BCI in a
> > > contest-station. 50+ dB notch filters on the BC frequency
> > > (differential mode filters) have proved useless.
> > >
> > > BC has 10kW out on 1584 KHz and antenna is only 100m away from 160m
> > > amateur antenna.
> > [Ed-K0iL] I would move. That's going to be very tough to
> > completely eliminate! Good Luck!
> > 73,
> > de ed -K0iL
Administrative requests: rfi-REQUEST@contesting.com