[Elecraft] Grounding

[Jim Brown]

On Sun,9/27/2015 10:37 PM, Fred Townsend wrote:
>
>
>   > Often the requirements of one system will be at odds with other >
> grounding requirements. For instance you generally want a high > 
> impedance for RFI issues and a very low impedance for lightning > 
> grounds. One system defeats the other so they really need to be 
> separate systems.

> NO, NO, NO, NO!
>
>    [Fred Townsend] Well yes maybe. How do you bond a ferrite choke? 
> Are not all chokes high impedance at their intended useful frequency? 
> I said 'system' not 'ground'.
>
> [K9YC]
>
> Huh? Chokes are used on SIGNAL lines, NOT on bonding conductors. A
> fundamental requirement of bonding is that it be LOW IMPEDANCE, so, in 
> general, bonding conductors should not be choked. Further, lightning 
> is NOT a DC event, it is an RF event. It's an impulse, with broad 
> energy peak around 1 MHz.
>
> [Fred Townsend] Correct on all scores
> Jim I think you have tunnel vision on bonding. Jim as others suggested 
> this was really an EMI issue and I was addressing this as a system or 
> EMI issue.

The bonding methods I have outlined in my tutorial are specifically 
designed to satisfy both EMI and lightning protection. Proper bonding 
for NEC is NOT in conflict with noise free operation of ANY system, 
whether it be audio, video, or radio.

> In this case chokes are appropriate for signal AND power cords 
> containing neutral. I don't believe anyone suggested chokes on bonding 
> conductors. However you suggested everything should be bonded and I 
> merely asked how you bond a choke.

I said that ALL GROUNDS and GROUNDED equipment must be bonded together. 
A choke is neither.

Signal cables and power cables are NOT bonding conductors.  Bonding 
refers to those conductors that provide a dedicated, robust, low 
impedance connection between all grounded elements of premises wiring. 
Bonding conductors are those which connect together all equipment in a 
station, all earth electrodes in a premises. This includes but is not 
limited to the EARTH connections for CATV, Telco, satellite dishes, 
mains power. It also includes those used for antennas where they enter a 
premises.  In simple terms, NEC (and the laws of physics) allow as many 
earth electrodes as we want, but they MUST all be bonded together.

> Also the NEC as well as other codes prohibit tying neutral and safety 
> (green wire) together.

Not quite!  NEC REQUIRES that neutral and green be bonded together at 
one, and ONLY one point in every SYSTEM, and, in general, requires that 
the bond be very close to the point where the system is established.  
[The word "system" here specifically refers to the wiring that 
distributes power in a facility, and to equipment connected to it.]  A 
"system" is established when the power enters a premises, and a bond is 
required there. In most premises, that's the only system. A system is 
also established by a transformer, as might be used in large buildings 
to run large loads, and by equipment like a UPS. In both examples, this 
does NOT require a new earth electrode, but rather a bond between 
neutral and green at that point.

The correct part of your statement is that NEC prohibits tying neutral 
and safety (green) together any place OTHER than where the system is 
established.

> Two cases I called out in my four different parts of the system. So no 
> you do not tie all grounds together.

You misunderstand the concepts. See the above.  All grounds MUST be 
bonded together.

Another important distinction -- a cable shield is not "grounded," and 
there is no virtue in doing so EXCEPT for lightning protection. The 
shield is properly connected to the shielding enclosure at each simply 
to make SHIELDING continuous, and thus make it effective. This is not a 
GROUND, it is a SHIELD connection. And in an unbalanced circuit (where 
the shield is the signal return), it is also signal return.

When we add a common mode choke to an antenna downlead, we are 
preventing common mode current on the cable shield. We bond the antenna 
shield to ground at the point of entry, and we bond that point of entry 
to a the building ground, but that is to protect the premises, NOT the 
antenna.  That cable shield is NOT a ground, it is a shield, and it is a 
signal return. That would work with a coax feedthrough, but it would not 
protect a rig connected to the antenna. To do that, we replace the 
feedthrough with a Polyphaser, which shorts the center conductor to the 
shield in the event of a strike.

Again, I strongly urge you and others who want to understand these 
concepts to study the links on my website, previously cited. Members of 
the AES Standards Committee Working Group on EMC spent YEARS in 
discussions that ultimately led to the understandings I've articulated. 
The members of the WG are all engineers, most with very broad background 
in everything from power systems to broadcasting to recording to live 
sound to RF systems. They included engineers from the BBC, ABC-TV, sound 
system consultants, equipment manufacturers, and contractors. Many 
are/were hams.

> Furthermore I think the whole issue of grounding is exacerbated by the 
> overuse and incorrect use of the word 'ground'.

I absolutely agree, and I've preached that for years. I wrote material 
on exactly that for the ARRL Handbook, and it was rejected.

> The Brits prefer to use the word 'earth' which helps a little.

Only a little -- they use "earth" and "earthing" in the same manner that 
we use "ground" and "grounding" in North America, and they talk about a 
"Protective Earth Conductor" which corresponds roughly to our "green 
wire." I serve on the AES Standards Committee with these folks. Good 
people, but they talk just as funny as we do.

> I would prefer to bury the word 'ground' and replace it with context 
> appropriate words like 'signal common' but that isn't common practice.

It should be, and I have been doing that for years. There are, at least, 
four common uses of the word "ground" to describe very different things. 
I have long preached that the words "ground" and "balun" should be 
erased from all literature, because they do far more to confuse us than 
enlighten us.

And this question remains. When I spend MONTHS writing a tutorial on 
stuff like this, and post a link to it in response to someone who 
obviously needs to study it, why is it that the person won't take the 
time to read what I have taken pains to write for their benefit? 
Instead, we burden the thousand or so readers of this list with a long 
response that should not be necessary.

73, Jim K9YC