[RFI] Inverse Square Law and RFI

[Jim Brown]

While this is all solid, there's a LOT more to it than that.  First, we 
must understand the difference between impulse noise generated by arcing 
on the power system, which is extremely broadband, and is radiated by 
the wires connected to the source (the arc). Because it's an arc, it's 
spectrum is theoretically infinite. But because the highest frequency 
components are most strongly radiated by the wires close to the source, 
the most effective way to find the source is at the highest frequency 
where it can be heard.  Lower frequencies are radiated more strongly by 
longer sections of the power line.

Noise is often attenuated by transformers, which helps localize the 
source a bit more.

Unfortunately, noise sources inside a house often DO go much further 
than the house. They are mostly radiated as common mode signals on 
wiring connected to the source.  These are NOT usually impulse sources, 
but rather electronic sources. They mostly fall into three categories: 
1) clocks in microprocessor systems, which tend to be quite stable in 
frequency and appear as a steady or pulsed carrier; and 2) switch-mode 
power supplies, variable speed motor controllers that rectify the power 
line, do almost no filtering of the DC, then switch that DC at high 
audio frequencies (10-20 kHz is common) to produce a square wave, again 
do only enough filtering to take the ripple out of the DC (and often 
regulate it). While it is POSSIBLE (even easy using smart circuit design 
and layout) to prevent the strong and rich harmonics of that square 
wave, it is rarely done (or done very well) for most consumer products.

In general, these power-handling devices are free-running and rarely 
frequency stable, and the poor filtering the DC produces an 
amplitude-modulated "hump" of noise at harmonics of the  switching 
frequency that drift (because they're not frequency controlled by a 
clock). This mush of  noise is conducted as a common mode signal to 
wiring to which it is connected, again as a result of poor design and/or 
construction, and is radiated by the wiring (just like any other RF 
current would do on any other antenna). THESE signals can often travel a 
significant distance (thousands of feet) or even miles for strong 
sources that produce large enough currents.

Now, noise CAN be radiated by wiring internal to products due again to 
poor design, but THIS component of noise generally does NOT travel very 
far because the wires are too short. Plasma TV sets are an example of 
this -- noise is produced by the switched currents driving the display.

I've written a lot more about noise, and it's on my website. 
k9yc.com/publish.htm

BTW -- what the EMC rules call common mode is the voltage between 
neutral and ground, which is NOT what hams call common mode, and it's 
not what radiates. Rather, common mode current can best be described as 
the algebraic sum of all currents on a cable, and it's that common mode 
current that radiates. This often happens when the "green wire" is not 
properly connected to the shielding enclosure. EMC line filters external 
to equipment won't do anything to kill this current -- it's on the green 
wire, and the green wire goes right past the filter, which is between 
phase and neutral and neutral and ground (green). The only way to kill 
this current is a ferrite choke. For the HF bands, we need multiple 
turns of the cable through a suitable core. Lots of details at the link 
posted above.

73, Jim K9YC

First, think On 6/11/2018 3:20 PM, David Eckhardt wrote:
> from a device located in a neighbors home over long distances before being
> attenuated?
>
> This is why its so difficult to isolate arcing or corona using RF
> emissioins and walking the length of the power line.  Generally sources
> inside a neighbors home don't go much further than the house.  However, if
> they are strong sources, they will.  The reason we do conducted emissions
> in regulatory testing is that lengths of conductors in housing are good
> radiators of RF energy below 30 MHz.  So, we measure conducted emissions
> that become radiated emissions once they couple onto the house wiring.
> Conducted emissions are measured and regulated from 150 kHz to 30 MHz for
> most home electronic equipment.  The generator is usually not a good
> radiator due to small size wrt wavelength.