846 Chapter 25
a good thing, in some instances it can merely serve to
improve the matching and coupling of the RF source to
the resonance.
Carried to an extreme, even a console frame consti-
tutes a big fat resonant tank at a surprisingly low
(mid-VHF) frequency while frame resistance, however
heavily it may be constructed, cannot be disregarded
and cannot be treated as a universal ground path.
For the purposes of practical design, these consider-
ations perhaps become a little better defined. The reac-
tive elements of capacitance and inductance with the
attendant effects of resonance and filtering are
concerned with less obvious aspects (such as electronic
stability and proneness to radio demodulation), while
resistance gives rise to most of the horrors usually
lumped under the collective term grounding problems.
25.8.2 Earth Ground
The closest most of us get to earth is the fat pin on an ac
power plug. Fortunately for most purposes, it is
adequate, provided just the one point is used as the
reference. Other points are likely to have slightly
differing potentials due to dissimilar routing and resis-
tances. Compared to a technical earth ground (e.g., a
copper water pipe or, alternatively, a fortune in copper
pipe hammered into the earth), conventional earth
grounds can have a surprisingly high potential, a volt or
two, considering it is principally a safety facility not
ordinarily carrying current. Any potential implies resis-
tance in the earth path, which is bad news about some-
thing intended as a reference while also detracting from
the safety aspect.
Practically, though, it does not matter too much if
everything is waving up and down a bit provided every-
thing, including even unrelated things in proximity, are
waving up and down in the same manner. The potential
is usually small, meaning that the ground impedance is
reasonably low to the extent it may be considered
insignificant.
25.8.3 Why Ground Anything to Earth?
With all our component system parts tied together by a
reference ground and everything working as expected,
the question arises as to why it is necessary to refer our
ground to earth. If the internal grounding is completely
correct, our system will operate perfectly, quietly, and
tamely regardless of to what potential (with respect to
earth) it is tied. If not tied, it will derive its own poten-
tial by virtue of resistive leakages, inductive coupling,
and capacitance to things in its environment. For an
independently powered system (i.e., batteries), these
leakages and couplings will be of very high impedance
and, hence, easily swamped by human body impedance
to earth.
If, as is most often the case, most of the system is
powered off the ac lines, this floating ground potential
becomes of far lower impedance and consequently is
much more capable of dragging current through a
human load. That’s you or me. (It’s the current that kills,
not the voltage.) A telltale sign is a burring, tingling
feeling as you drag a finger across exposed metalwork
on something that is deriving its own ground potential.
The mechanism for this lower impedance is fairly
straightforward. Power transformers are wound with the
optimum transfer of energy at 50–60 Hz and very high
flashover voltages (several kilovolts) in mind; the finer
points of transformers such as leakage inductance, inter-
winding, and winding imbalance capacitance are all but
disregarded.
Being far greater in scale than ordinary ambient reac-
tive couplings, they primarily dictate the floating ground
potential to be anything up to 240 Vac above ground or
whatever the power lines happen to be locally.
It used to be that some units were fitted with bypass
capacitors from each supply leg to chassis ground,
partly in the fond hope that this would help prevent any
nasty noises on the ac mains from entering the hallowed
sanctum of audio within. Ungrounded, this guaranteed
the chassis floating at half the supply rail from a fairly
low impedance. Ouch. But it gets worse. With the near
universality of switch-mode power supplies, and with
nearly everything containing digital electronics to some
degree or other, there is the imperative of ensuring any
nasties that the supply/digits generates don’t find their
way out of the box and up the power lead! Some
reversal of fortunes there. This is required not so much
from the altruistic desire to not pollute but that the box
likely wouldn’t pass emissions testing (FCC Part 15, CE
and the like) and wouldn’t be able to be sold. Typical
supply-side filtering as a minimum in such supplies is a
S filter of common-mode inductors and parallel capaci-
tors—including capacitors to chassis ground.
The result is that if the chassis is not directly earthed,
it rides at (in the case of both lines having tied capaci-
tors) half the line voltage. The capacitor values grossly
swamp transformer and ambient leakages and give the
chassis floating potential an uncomfortably (literally)
low impedance. The chassis tingle changes from
“Mmm—interesting” to vile oaths with attendant
flailing limbs.