[RFI] Wiring practices to minimize emissions

[Jim Brown]

On Wed,7/13/2016 7:27 AM, nm8rmedic via RFI wrote:
> Gentlemen,
>
> Thank you for well considered input.
>
> 1) Mention was made of using twisted pair. Is that practical for 40 horsepower motors?

What are the voltage and current?  Twisted pair to handle up to about 
40A can easily be formed by placing a pair of stranded conductors rated 
for the voltage and current  in a vise on one end and a drill in the 
other.  I'm using #10 THHN twisted pair in my small solar system.

> 2) I recall from past board discussions the necessity of minimizing the distance between, and area encompassed by, power carrying conductors to reduce emissions.  Would someone please expound on that.

The mechanism is the creation of a magnetic field by the conductors that 
form the current loop. The strength of any field is directly 
proportional to the area enclosed by the loop. Unintentional loops are 
formed by stray capacitance between the motor windings and it's frame, 
and also by capacitors added to bypass RF components to "ground," under 
the mistaken assumption that the earth is somehow a sump into which all 
trash may be poured.

Many years ago, I saw an excellent demonstration of the importance of 
the use of transmission lines to carry noisy signals at an IEEE EMC 
conference in Chicago, where I also happened to be a featured speaker.  
The setup was quite simple -- an HP audio generator (up to about 2 MHz) 
connected by RG58 to a load a few yards away. The generator and load sat 
on a large flat plate, and the return for both were bonded to the plate. 
Current probes were connected in series with the wire feeding the plate 
and also with the coax shield, and fed to a dual trace scope. The 
generator was swept from DC to a high frequency.

At low audio frequencies, all the current was in the plate, but as the 
frequency increased, still within the audio spectrum, more of it was in 
the coax shield. And by the time the generator hit 10 kHz, nearly all of 
the current was in the coax shield. The principle here is that current 
follows the path of the lowest IMPEDANCE, and when the frequency is 
sufficiently high, the impedance of a large loop is dominated by its 
inductance. When we provide a transmission line for a path, we confine 
high frequency components to the line, which in turn confines the 
magnetic field to the dielectric, and which also prevents radiation 
(because the field of the forward current is cancelled by that of the 
return).

In his 3-day EMC workshops, Henry Ott shows how a trace on a 2-layer 
circuit board where the second layer is a "ground" layer forms a 
transmission line with the ground layer, such that at high frequencies, 
return current is carried by a small region directly under the trace. 
This greatly suppresses crosstalk on the board AND radiation by the trace.

He then notes that a break in the "ground" layer (to add a trace that 
the layout artist forgot on another layer) interrupts that return path, 
causing return current to find whatever path remains, often the 
shielding enclosure or a path around the perimeter of the "ground" 
layer. This produces a strong magnetic field (because of the large loop 
area), resulting in crosstalk to other circuits, and, if the device is 
not well shielded, radiation by simple antenna action.

Ott also observes that in the near field of a source, the magnetic 
component of the field dominates, and at low frequencies, the near field 
can extend hundreds or even thousands of feet (it's a direct function of 
wavelength).
.
> 3) The motor controllers will be soft-starts rather than VFD's . Are there any special EMI considerations regarding soft-starts?

I plead ignorance of that technology.  Ott observe that an important key 
to EMC is to understand where the current flows, and that includes ALL 
of the current, including the unintentional components like the high 
frequency harmonics of switching, regulation, and digital signals.

It is well known, for example, that NEC prohibit bonding of neutral to 
ground except at the service entrance, or within a facility where a new 
"system" is created (for example, by a distribution transformer), where 
the neutral must again be bonded on the secondary.  At the fundamental 
frequencies of power (60 Hz and low order harmonics), a neutral to 
ground bond anywhere else would allow some return current to flow on 
ground, which can include building structure and other paths. This does 
two things. First, the field between the phase and neutral conductors 
only partially cancels (because neutral current is now less than phase 
current). Second, the part of return current on "ground" forms a strong 
magnetic field because of the large loop area. Now, here's the kicker -- 
when there is enough capacitance, either intentional or stray, between 
neutral and ground, it can create that bond between neutral and 
equipment ground.

Hope this helps.

73, Jim K9YC