[RFI] Single-point grounding

Jim Brown jim at audiosystemsgroup.com
Sun Jul 1 12:52:08 PDT 2012

On 7/1/2012 7:28 AM, Pete Smith N4ZR wrote:
> I'm not sure exactly where I should post this, but suspect that this
> group probably has more relevant expertise than most, so please bear
> with me.
> In the last year, I have had two expensive episodes.  In both cases, my
> transceiver's RS-232 transceiver was fried during a lightning storm, and
> several knowledgeable people suggested a difference in chassis potential
> between computer and transceiver as the reason.  Their prescription for
> avoiding a repetition was simple - connect the chassis of all of the
> units to a single-point ground.

I'm a strong advocate of that, but there's more to it.  The fundamental 
problem is that we have many pieces of equipment that have unbalanced 
connections between them, and that are connected to ground at different 
points.  A lightning event induces current in those ground wires, with a 
resulting IR drop in the grounding conductor. The DIFFERENCE between the 
voltage on one chassis and another appears on the low voltage connection 
between the equipment and smokes it.  The difference between the power 
outlet and our shack ground can also smoke equipment,

Exactly the same thing happens with leakage current on power wiring, 
causing hum and buzz to be added to unbalanced wiring between 
equipment.  The solution is quite simple, and applies equally to your 
problem.  1) Bond chassis to chassis from all interconnected equipment 
using short, fat copper. I've generally recommended either braid 
stripped from transmitting coax or #10 stranded copper. 2) Get power 
from all interconnected equipment from outlets that are in the same 
electrical backbox, or that share the same green wire, or that have 
their green wires bonded together.  3) Bond from the collection of 
equipment chassis to the common point where your antennas are grounded, 
again by short fat copper.  4) Follow NEC requirements for bonding 
carefully. This includes bonding EVERYTHING  to the power system ground 
by short, fat copper -- your station grounds, CATV, satellite, telco, 
cold water (if conductive), building structural steel, and lightning 
rods, if any.

> My solution - which I'm asking people to critique - was to fabricate an
> L-shaped, roughly 3x5 foot ground bus made from 3/4" copper tubing,
> which I mounted on the back of my L-shaped operating desk. All joints
> are silver-soldered. I then connected each of the affected units to the
> bus with very short and heavy stranded wire, and connected the end of
> the bus to my grounded shack entry panel (in a double-hung window).
> The DC resistance of the ground bus is very low, but the length
> approximates a quarter wave on 10 meters,

IEEE research says that the energy in lightning is very broadly centered 
around 1 MHz, with lots of energy at least a decade of frequency above 
and below.  At these frequencies, impedance is dominated by inductance, 
not resistance. Thus short is better.

> and with the units connected
> near the ends of the bus, I wonder if I'm feeling a false sense of
> security about the likely behavior of the bus during a nearby lightning
> event.

What you've done is very good, but I would ADD the short fat bonding 
between interconnected equipment. Contrary to popular myth about "ground 
loops," the added bond REDUCES noise and the ightning voltages between 
equipment that causes destructive failures.

> Would I be better off (or no better) running heavy conductors
> from each unit to the entry panel, even if they would have to be
> similarly long? Or should I just give up and plan on disconnecting the
> RS-232 connection whenever weather approaches?

We MUST connect every AC outlet to the entry panel -- it's the "green 
wire" -- and it MUST be bonded to the chassis of all connected equipment 
that is required to have a 3-wire plug.  The only AC-connected equipment 
that does not require that bond are products that are double insulated 
to prevent the possibility of shock. This equipment will have a 2-wire 
plug.  Lots of wall warts and computer power supplies have this 
exemption.  BUT -- the chassis of equipment powered by these power 
supplies still needs to be bonded.

73, Jim K9YC

More information about the RFI mailing list