Topband: Elimination of Treadmill RFI on 160 meters

[Jim Brown]

On Sat,1/31/2015 5:48 AM, Paul Christensen wrote:
> For a long time, Corcom has recognized this for use in their 
> power-entry EMI/RFI filters.  Their high-performance line filters use 
> a combination of common-mode and differential-mode components.

In name only. As Tom noted earlier, AC line filters do NOTHING to the 
Green wire -- it goes right through them. What the power industry calls 
"common mode" is voltage between neutral and Green. That's NOT common 
mode as the rest of the world would define it.

On Sat,1/31/2015 6:58 AM, Jim Garland wrote:
> We're just making the point that a common mode choke's effectiveness
> in blocking common mode currents depends on the impedance to GROUND of the
> termination of the conductors. I used a parallel line in my example merely
> to illustrate the point, since current flow in e.g., coax cable.

On Sat,1/31/2015 6:58 AM, Jim Garland wrote:
> We're just making the point that a common mode choke's effectiveness
> in blocking common mode currents depends on the impedance to GROUND of the
> termination of the conductors.

Not the impedance to GROUND, but the impedance of the common mode 
circuit as an antenna.

> I used a parallel line in my example merely
> to illustrate the point, since current flow in e.g., coax cable.

An antenna doesn't have to be grounded to carry RF current and radiate 
it.  As the extreme case, consider, for example, your parallel wire line 
that is a half wavelength long (free space). It's a near ideal length, 
and will radiate a lot of signal if the antenna puts common mode current 
on it.  A ham talkie has no connection to earth, but the chassis and the 
capacity-coupled body of the person holding it provides enough of a 
counterpoise for it to work. Likewise, a quarter-wave whip on a vehicle 
has no connection to earth, but the vehicle's frame provides the 
counterpoise (if we manage to remove or bypass the paint that prevents 
sections of it from making contact).

A common mode choke of the sort I've recommended (and measured) is, by 
definition, a brute force solution. Yes, the extent to which it reduces 
radiation from the unintentional antenna will be very dependent on the 
impedance of the common mode circuit and the effectiveness of those 
unintentional antennas as radiators. I've looked at it in NEC models 
with best and worst cases of feedline length with half-wave dipole fed 
varying degrees of off center, for example. But if the noise source is 
strong,  and/or if the noise source is close enough to us, antennas 
don't have to be ideal to give us a lot of grief.

The first work I did on this was around 2003, in Chicago, and the first 
noise sources I tackled were the wired Ethernet switch and computers 
connected to it. I made a pretty good dent in the birdies that I heard 
in my HF RX by winding 5-7 turns of the CAT5 through a #43 Fair-Rite 
toroid (I didn't yet have #31).  But it made no dent at all in the 
broadband noise I heard on 2M from those cables.

But the reason I was looking at this in the first place was that I was 
hearing lots of anecdotal reports of RFI to church sound systems from 
stations on the high end of the AM broadcast band, and wanted to both 
understand the causes and develop a fix. The SCIN paper that I 
referenced to you yesterday showed one of the causes. Another paper 
documented the Pin One Problem as part of the cause. Both causes were 
related to shield current on shielded twisted pair cables. The 
multi-pair mic cables running through a wood-frame church make great RX 
antennas on the broadcast band, and the otherwise very nice Mackie 
mixers not only had Pin One Problems at their mic inputs, but also had 
response from DC to daylight in the name of good audio phase response. 
When I swept 2 MHz into their inputs, I got 2 MHz at their outputs! In a 
field test across the road from WGN (50kW) 14 turns of the mic cable 
around a #43 2.4-in o.d. toroid killed the RFI. As I learned later when 
I studied chokes in detail, this wasn't even an ideal choke, but it was 
good enough. This was 2001-2005, and after I very seriously beat them 
up, Mackie redesigned their mixers to correct these issues.

> Obviously a parallel line not terminated in anything will have
> no common mode current because there's no place for the current to go.
> Introducing choke coupling and stray capacitance, Q, resonances, etc., into
> the discussion, while obviously important in real life, tend to obscure the
> basics.

Not true -- indeed, a common mode choke is most effective BECAUSE OF the 
high resistance around resonance that RESULTS from its own stray 
capacitance.  And because it's a very low Q resonance (typically 0.4), 
it's quite broad. The virtue of #31 material is that there are TWO 
resonances -- the resonance of the coil, and the dimensional resonance 
of the core. This is discussed in my tutorial. Dimensional resonance is 
discussed in detail in the first Snelling book, which is also referenced.

73, Jim K9YC