[Elecraft] Elecraft: RFI Problems

[Stephen W. Kercel]

Martin:

I have a few thoughts on your EMI problem.

First, keep in mind that an antenna produces three different kinds of 
fields, the induction field, the near field and the far field. The 
induction field vanishes within inches of the antenna. The near field, the 
cause of many emi problems, has non-trivial strength out to about 1/6 of a 
wavelength. The far field is the one that "gets out," and behaves like a 
nice predictable travelling wave. The near field turns out to be extremely 
sensitive to boundary conditions, and simple acts like opening or closing a 
door can change the boundary enough to change the pattern of the near 
field. This is one reason for the seemingly "magical" properties of emi.

My reason for making this disclaimer is to make you appreciate the fact 
that any of the very sound advice you get on the reflector may or may not 
help you to overcome your particular emi problem. I would encourage you not 
to give it up; your problem (or set of seemingly unrelated problems) can be 
solved.

Second, you have not mentioned what bands give you the emi problem. You 
should check for all the symptoms on all the bands. There is a good 
possibility that part of your house wiring is perversely resonant in one of 
the ham bands, and is acting like an antenna that captures and reradiates 
your near field. Of course, if the resonance occurs on 80 meters, you might 
also be resonant on all the harmonically related ham bands. Anyway, it is 
worth checking which bands cause which bad effects.

Third, having no RF ground is a big nono. The problem is that the RF that 
is on the chassis of your rig has to go somewhere. If you do not provide it 
a low impedance path to ground (where RF is converted to heat) it will find 
its own path, and the one that it finds will not make you happy. The power 
system ground has very high inductive reactance (A straight wire has an 
inductance of 10 nanohenries per inch. A 10 foot length of wire has an 
inductive reactance of 26 ohms at 3.5 MHz in the absence of mitigating 
capacitive effects. ) In other words, from your rig to the power system 
ground rod the overall impedance is probably on the order of hundreds of 
ohms, and virtually none of your RF energy is finding its way to the ground 
rod. If you're lucky (in your case you're not) the energy will be turned 
into heat in the losses of the wiring. Otherwise, it gets coupled into your 
other equipment.

Since you are physically at ground level, I would strongly recommend that 
you install an 8 foot copperweld ground rod as close as possible to the 
rig, and feed through the shortest possible number 6 solid wire. (Solid has 
lower inductance than stranded or braid.) This is not an expensive 
solution, but it may solve your problem. Going from the shield of your 
cable direct to the ground rod may solve your problem and it may not. The 
emi problem may not be caused by radiation emanating from the outside of 
your shield. To avoid the creation of ground loops, and new emi problems, 
it is considered best practice to use a single point RF ground. Connect the 
rf grounds of your auxiliary equipment to your antenna tuner ground if you 
have one or to your transmitter ground otherwise. Make sure the ground 
connections are done in a star configuration (no loops in the ground wire), 
with short direct connections of solid wire (braid has higher inductance; 
avoid it) with your rig or tuner ground being the center of the star. To 
that single point at the center of the star connect the lead to the RF 
ground. This is a compromise configuration that tries to balance many 
conflicting tradeoffs, but it is the one that typically results in the 
fewest emi problems.

The 8 foot copperweld ground rod is not really that good an electrode at 
HF, but it is better than nothing and vastly better than what you have now. 
In fact, there is no good ground electrode at HF. All of them have 
non-trivial inductive reactance. The least electrically bad ground 
electrode is a long solid copper wire (I prefer number 6; it has good 
physical strength and is not too hard to handle) at least a half wavelength 
at your lowest frequency and buried just deep enough that your lawn mower 
does not hit it, but preferably less than an inch deep. Every 8 feet or so 
along the wire you should connect the wire to a 12 inch ground rod. The 
wire does not need to be in a straight line. It can meander all over the 
yard, turning corners as necessary, but no corner should be tighter than 90 
degrees. Obviously this is not practical to install during the winter. 
Nevertheless, when you can install such a ground, it is worth the trouble 
doing.

BTW, a cheap trick that often works in place of an RF ground is to use an 
MFJ (or equivalent) artificial ground to a length of wire running along the 
floor as an RF counterpoise. The idea is that the reactive elements in the 
box tune out the reactance in the counterpoise and create the effect of a 
low impedance RF ground. You can find artificial grounds pretty cheap on eBay.

You note that "these are all in different rooms of the house and all at 
least 15 feet away from the rig." Since the emi is by your own observation 
being coupled into the house wiring, distance from the rig is irrelevant. 
If you're lucky, the losses in the power company's equipment where your 
house connects to the system is attenuating the emi, and probably none of 
the energy is going out on their transmission line and into other peoples 
houses.

Personally, I would be looking for a systemic cause and cure rather than 
cleaning up individual symptoms. You might clean up everything you have, 
but the next new piece of gear that you bring into the house may cause you 
a new set of problems.

Some comments on your proposed fixes:

"Make a coax choke balun in the feeder to reduce out of balance currents
(quick and cheap)
Fit a commercial balun (heavy and pricey)"

Unless you're feeding the dipole off center, or your transmission line is 
severely off perpendicular to your antenna, or you have some very gross 
asymmetry in the boundary conditions (like your house having a metal roof), 
your configuration does not look like it is especially vulnerable to the 
problem. Unless there is something causing you have RF on the outside of 
your coax sheath, the baluns probably will not buy you anything.



"Construct an RF ground in the garden and connect to feeder sheath at entry
into the house. (time consuming)"

Time consuming, but probably worth it.  However, as I already noted, you do 
not appear to have a compelling reason to believe that you have a line 
radiation problem. I expect that you'd be better off using the single point 
connection scheme I described above.


"Fit a ferrite ring choke to the PA mains lead"

This is a cheap trick, but it might work. There are a couple of gotchas. 
First, in the absence of an RF ground connection, you have no low impedance 
path that takes the emi energy to a place where it is converted to heat. 
You're simply redirecting a bit of bad energy, and the practical result may 
be to trade one emi problem for another. However, you may find that you 
need both the ferrite and the RF ground working in concert to solve the 
problem. If by a ferrite ring you mean those little rectangular jobbies 
that you get from MFJ or Radio Shack, be aware that they do not provide 
much attenuation, about 3-5 dB per turn at HF. You need to go 7 times 
through the core to get significant attenuation, maybe 7 turns through one 
core, or a stack of 7 cores on a straight wire, or some other combination.


"Build an inductively coupled ATU to avoid connection of feeder sheath to
the mains ground."

You'd only avoid a DC connection. If you do not provide an RF ground, and 
if you do have feedline radiation, the system will probably find some other 
coupling mechanism to couple energy from the feedline to your house wiring.

"Fit ferrite ring chokes to the affected equipment"

Again, it might work. However, those rings are about four dollars each, and 
you might need quite a few on each piece of equipment. The price could grow 
in a hurry. You're probably much better off trying to find the mechanism 
that's coupling RF into your power system and breaking that up.


"Fit a proprietary RFI filter to the telephone line (to cure the modem
problem)"

This will only work if the problem is that RF on the phone line is being 
coupled into the modem and being demodulated by the modem to create the 
local emi effect. You already know that you've got emi on the power line, 
if your emi is coming in through the power lead, the telephone filter would 
not do you any good; you'd be locking the wrong door. You might want to put 
the scope on the phone wires and see if you have RF there. Those telephone 
filters are about 11 dollars each. You do not want to make a big investment 
in them unless you have reason to believe they will actually solve your 
problem.

Good luck with your problem, and please keep us posted on your progress.

73,

Steve
AA4AK