Optimally designed High-End cables can be expensive, so it is important for our
customers to understand what benefits these optimized cables can provide.
If most high-end cables were compared to inexpensive, but good quality cables
(such as those from from Radio Shack) on a single-speaker monophonic system,
virtually all humans would find them to be very close. The high-end cables may
have slightly better high-frequency clarity due to the use of Litz-wire or pure
silver conductors and less lossy dielectric insulations, such as Teflon or air-filled
Polyethylene or Teflon. These materials are more expensive.

Significant differences however, will be obvious in stereo. Properly designed high-
end cables will generally have lower phase shift at high frequencies, resulting in
a more clear, more focused stereo soundstage image. Humans can discern very
small phase shifts and frequency modulation when they are in the "sweet-spot"
of a two-channel stereo system. The reason that these are more obvious in stereo
is that both speakers are required to properly recreate a single sound source. If
the signal from one of the two speakers is not in-phase with the other because
its contribution is low frequency (no shift in phase) and the other is providing the
high-frequency (changing phase) component (such as room echo), then the image
of this source will be "fuzzy" and difficult to pinpoint. It is the difference in signal
from left and right that humans can discern. If one moves out of the sweet-spot,
this difference does not change, only the relative position of the image. The lab at
Empirical Audio has measured music bursts above 8 kHz and seen significant phase
shifts at these frequencies with inferior speaker cables. Our cables do not exhibit
these phase shifts.

Surround-sound makes it possible to place audible cues more easily, because there
are more speakers. However, this does not relegate the need for precise imaging.
If left-to-right or front-to-back pans are to sound natural, without obvious jerking,
a precise image is necessary. High-end cables can provide this precise image, adding
to the realness and excitement of surround music and movies. It is possible for the
center channel to "disappear" when playing surround program material much in the
same way the left and right speakers can disappear on two-channel program material,
but precise imaging is a prerequisite for this to happen.

High-end speaker cables will also typically use large gauge conductors or sufficient
numbers of Litz-wires to be equivalent to a large gauge, such as 8-10 AWG, reducing
the resistance and inductance of those cables. Lower resistance and inductance will
generally result in improved dynamics and "slam" depending on the capacity of the
amplifier to deliver current when needed and the efficiency of the loudspeakers. This
is simply ohms law at work.

Finally it is important to note that having high-quality source, preamp and amplification
equipment is critical to getting the most out of high-end cables. Any component can
limit the performance of the system and prevent a "live" sound from being achievable.
However, properly designed high-end cables can improve, to some extent, the sound of
virtually every system. To get the biggest bang for your buck, the best thing to do is
try them. You may discover that your $5K system sounds like a $30K system with $2K
worth of high-end cables installed. The live sound that we get on our Empirical Audio
reference system rivals and even beats systems costing 10X more and it is primarily
attributable to our cables!

Minimize Capacitance
Interconnects transfer analog voltage signals between components. The voltages
involved range from microvolts to a volt or so, but the currents involved are always
extremely small. The currents are small because the load that the interconnect
drives is generally between 10-100K ohms. This is the input impedance of the
component being driven. There is virtually zero power transfer with interconnects.

Because there is basically zero power transfer, it is not necessary for the driving
component to be capable of driving much power. As a result, most components are
designed with an output impedance of between 7 and 200 ohms. Lower is better
because the driver is less "sensitive" to the load. However, the load is actually
comprised of a resistive part and a capacitive part. This capacitance is caused by
the integrated circuit or transistor packaging, the printed circuit board traces and
the silicon itself. This capacitance presents a load to the driving component. If
the capacitance is too large, the high-frequencies will begin to attenuate or
decrease due to the loading on the driver. The input capacitance of a component
is generally never characterized (not in the specs), but this is actually as important
as the resistance. The interconnect also adds to this capacitance and can actually
contribute more to the total capacitance than the receiving component. It is
therefore an object to minimize the capacitance in an excellent interconnect.

The capacitance of in interconnect is a function of its length. The longer it is, the
higher the capacitance. This is why interconnect length should generally be minimized.
Interconnect capacitance is also function of geometry and dielectric material.
Capacitance is minimized by spacing the two conductors apart as much as possible
and by avoiding parallelism. It is also minimized by using low dielectric-constant
materials between the two conductors.

How does Empirical Audio minimize interconnect capacitance?

1. Air dielectric is used between the two conductors where possible
2. Where air is not possible, Teflon or other low-dielectric constant materials are used
3. Conductor parallelism is avoided by geometry
4. Conductors are spaced apart

Minimize skin-effect
Skin-effect occurs when the high-frequency currents flow on the outer "skin" of
the conductors whereas lower frequencies have more uniform current distribution
across the conductor cross-section. This happens when too large a gauge is chosen
for the conductors. The effect is that the impedance (primarily inductance and
capacitance) is different for low frequencies than high frequencies. This difference
in impedance can cause attenuation and phase shifts in high-frequency passages
relative to low-frequency passages, causing a smearing effect to the music. If a
sufficiently small gauge is chosen for the conductors, all frequencies are "forced"
to flow more uniformly in the conductors, effective eliminating skin-effect. Skin-effect
is also a function of conductor material.

How does Empirical Audio minimize skin-effect?

1. Careful selection of conductor gauge and stranding to insure optimum low and
   high-frequency response.
2. 99.99% Pure Silver conductors

Minimal use of conformal coatings
Conformal coatings (insulation) on conductors create a non-uniform dielectric medium
around the conductors. This dielectric material stores energy from the conductors in
the form of charge. Similar to a battery, the dielectric material prevents the conductors
from discharging immediately and completely when the music waveform demands this.
The result is that latent charge is still present in the dielectric material to be released
when it is not desired. The technical term for this effect is Dielectric Absorption. This
effect is more pronounced in less expensive cables that use PVC for insulation rather
than Teflon or other low dielectric-constant materials. This has two detrimental effects:

1. Latent charge can change the amount of energy required to charge the dielectric,
   drawing less current with some passages than others from the driver.
2. Latent charge can appear on the conductors when it should not be there.

Either of these effects can conceivably cause "smearing" or dispersion of the audio signal,
particularly between left and right channels, where this can become audible to humans.

Speaker cables transfer high-bandwidth power in a low-impedance environment. The
power amplifier output impedance is usually on the order of a few tenths of an ohm
and the speaker load typically varies from 3-30 ohms depending on the signal
frequency, particular speaker and crossover. Transient currents of 30 amps or more
can occur during dynamic, high-level music passages when driving low-efficiency
speakers. Because power transfer is the primary purpose of a speaker cable, resistance
and inductance are more important than capacitance. The low output impedance of the
power amplifier is generally capable of driving relatively high capacitance, whereas
significant voltage drops can develop across the speaker cable and connectors due to
resistance and inductance, particularly when high-current transients occur. Extremely
high capacitance has been known to cause instability in some amplifiers, but this is not
typical.

Minimize Resistance
Empirical Audio speaker cables minimize resistance by using a sufficient number of parallel
runs of wire to equal the equivalent of 11 gauge. We have determined empirically that a
minimum of 11 gauge is required to work effectively with a broad range of amplifiers and
speakers. We do not offer a cost-reduced version of this for more efficient or low-budget
systems, because it will invariably be installed in low efficiency systems, resulting in
voltage drops and an audible loss of dynamics. We also offer only spade lug terminations
on our speaker cables. We have determined empirically that banana plugs sound fine in
the most efficient systems, but the vast majority of high-end resolving systems are less
efficient, so voltage drops will occur across banana plugs causing a loss in dynamics.
This is why we do NOT offer banana plugs on our cables.

Minimize Inductance
Empirical Audio speaker cables minimize inductance by grouping the conductors as multiple
twisted-pairs. The pairs are connected so that the current in first conductor of each
pair runs in the opposite direction from the current in the second conductor. The magnetic
field coupling between the twisted-pairs reduces the inductance, making it lower than the
self-inductance of the wires themselves. Most other cable manufacturers use large gauge
conductors to make the self-inductance as low as possible, but they do not take
advantage of the magnetic field coupling that can actually make the inductance even lower.
Some other manufacturers actually run parallel small gauge wires for each of the two
conductors, but this actually increases inductance, so it is a bad idea.

Minimize Skin-effect
Skin-effect occurs when the high-frequency currents flow on the outer "skin" of the
conductors whereas lower frequencies have more uniform current distribution across the
conductor cross-section. This happens when too large a gauge is chosen for the conductors.
The effect is that the impedance (primarily inductance and capacitance) is different for low
frequencies than high frequencies. This difference in impedance can cause phase shifts in
high-frequency passages relative to low-frequency passages, causing a smearing effect to
the music. If a sufficiently small gauge is chosen for the conductors, all frequencies are
"forced" to flow more uniformly in the conductors.

Empirical Audio minimizes skin effect by careful choice of conductor size to optimize for
low as well as high frequencies. This insures that the current distribution is relatively
uniform at all audio frequencies.

Minimize multiple-conductor interaction
Multiple conductors are required to minimize inductance and minimize skin effect, but if
the geometry is not carefully designed, crosstalk between these can negate many of
the positive effects. Empirical Audio designs virtually eliminate interactions between twisted
pairs.

Minimize Dielectric Absorption and Dissipation Factor
Placing each twisted-pair in a separate tube forces air dielectric around the pairs, which
helps to lower the effective dielectric constant. A lower dielectric constant is desirable
because it results in lower dielectric absorption and lower dissipation factor.

A number of audio cable manufacturers have sprung-up over the last few years that are
intent on proliferating myths about the technical aspects of interconnects, power cords
and speaker cables. Since the staff at Empirical Audio is technical-based, we would like
to dispel some of these myths for our customers here.

Power noise and Power Cords
There are a lot of expensive high-tech power cords being sold in the marketplace these
days. Many of these claim to improve the delivery of AC to components by: Shielding
the conductors, providing very fine stranded conductors and other magical treatments.
Also, some audio power outlets are made of exotic materials and have heavy-duty contacts.

The reality is that a power cord made from 12-14 gauge solid copper is pretty good. The
problem with this is that this wire is not UL approved for cords and is very inflexible indeed.
Most electrical Romex runs to the outlet in question are 20-40 feet in length. The power cord
adds an additional 6 feet or so, so this is a small percentage of the entire run. It turns
out that typical "rubber" stranded copper power cords have significantly higher inductance
than the Romex in the wall, even at the same wire gauge, so these are not recommended.
Empirically, stranded rubber cords have been demonstrated to limit transient high-power
currents (dynamics) compared to solid copper conductors when supplying power to typical
audio power amplifiers.

It is fairly easy to build a serviceable cable that will minimize power cord inductance. A
simple 3-conductor twisted cable from 12 AWG solid THHN from Home Depot yields a very
high quality power cord, although it is so stiff that it must be bent to the desired shape. It
is actually superior to the Romex in the wall because the twisting and close proximity of
the insulated conductors will reduce the inductance by magnetic coupling between the
conductors. In the optimum configuration, the Hot and Neutral are twisted together and
then the ground wire run beside or wrapped around them. The trick is to design a flexible
version of this cord with the same characteristics. This is why some of the expensive
cords are actually good designs, although more than $500 is unreasonable to achieve a
good design. The Magnum2 power cord represents a good tradeoff between cost and
performance. It's inductance is extremely low, but it does not incorporate any exotic
materials. It is flexible enough to replace a rubber cord as well. It sounds significantly
better than the "Home Depot" cord described above, which is not bad. The Grand Slam
power cord takes this to another level, incorporating Bare Wire Technology, improved
dielectrics and ground-wire filtering (see the article below on RFI).

Unfortunately, most components do not have power supplies with sufficient energy
storage and fast enough response times so as to not benefit from low-inductance cords.
If the power supply in a given component has enough energy storage built-in with a
low-inductance path to provide current to the electronics, then an improved power
cord will have little or no effect. It is therefore primarily under-designed and inferior
power supplies in audio components that will benefit from improved power cords. From
experience, however, we have found that virtually all power amps benefit from a
low-inductance power cord.

Power Cord Shielding
Shielding a power cable is unadvisable. It will add significant capacitance to the cable
with minimal positive benefit. If you really need this, then the shield needs to be
spaced well away from the conductors (large diameter) to minimize capacitance and
avoid constraining the magnetic field lines that should couple between the conductors.
Empirical testing has shown that standard shielded 14 gauge stranded power cord
sounds less dynamic than unshielded 14 gauge stranded cord when used with audio
components that benefit from improved cords. The impedance of the AC electrical system
is extremely low and susceptibility to magnetic and RF fields is extremely low for power
cables so the benefit is questionable at best. Unfortunately, some of the commercially
available shielded cords appear to make some systems sound better, but are actually
"tone controls" for taming badly matched or designed components. There is some benefit
to shielding if you are trying to protect unshielded nearby unshielded interconnects from
the fields generated by the cord itself.

Better Power Outlets
Superior power outlets are another matter. Outlets that come stock in a home are
usually cheap ones with push-in wire connections and 15 amp contacts. These are
relatively resistive contacts. It is advisable to upgrade these to the screw-on wire
types with 20 amp contacts. Hospital-grade accomplishes this, albeit at higher cost.
Other improvements include high-copper-content outlets with silver or gold plated
contacts.

It is always advisable to run dedicated heavy-gauge (8-12AWG) copper runs to power
amplifier outlets, particularly in new construction.

Litz-Wire
Litz-Wire is created when a larger gauge solid or stranded conductor is split into a
number of smaller gauge conductors, each being insulated. This configuration improves
skin-effect. Skin effect means that current density of the high frequencies is greater
on the surface of the conductor than in the center. When a conductor is too large,
the current density tends to be essentially uniform from DC to mid frequencies, but
skin-effect occurs at high frequencies. This is why many small conductors are paralleled
in many cables, their combined cross-sectional areas equaling the area of a larger gauge
conductor. Skin-effect is detrimental in that the impedance of the conductors changes
with frequency instead of being constant across all frequencies. Skin-effect has been
shown empirically to occur in both interconnects and speaker cables.

Implementation of Litz-Wire can vary greatly from one product to the next. In speaker
cables, for instance, Litz-Wire is often seen running in parallel in close spacing with other
conductors whose currents are all running in the same direction, such as the
8-conductor Cable A below:

Because the currents are running in the same direction in adjacent conductors in Cable A,
the magnetic fields that couple between the same color conductors cause the inductance
of Cable A to increase. Since inductance is an important thing to minimize in speaker cables,
this is a bad thing. In Cable B, the conductors are alternated so that adjacent conductors
have current running in opposite directions. This tends to reduce the inductance of Cable B,
which makes Cable B the superior design.

RF resonances and "pollution"

Some cable manufacturers would have you believe that RF can easily sneak into interconnects,
speaker cables and power cords. Here is the reality:

Interconnect Shielding
Even in high-RF urban environments, shielding of interconnects is prudent, but not usually
necessary. The shielding need only have coverage such that the shortest wavelength is
attenuated by 50 dB or so. If the offending RF is television or radio, the size of the openings
in the cable’s shield need only be about ¼" in diameter, which is a very sparse shield. If the
offending RF is cell phones and other 800+ MHz RF, then holes in the 1/8" range should be
sufficient. Most common shields are much more dense than this, having a minimum of 90%
coverage. Shields add capacitance to the interconnect, so they should be used only when
absolutely necessary. A 90% coverage shield should be more than adequate for audio
interconnects, unless you live next to a transmitter. Most folks can successfully use an
unshielded cable, such as Kimber without any audible noise being picked-up.

Interconnect RF Resonances
RF resonances are possible on a shielded cable where the shield is not connected at one
end. This has nothing to do with the shield coverage, but with the length of the "stub"
antenna that is created by the un-terminated shield. If the length of the un-terminated
shield is equal to ½ or ¼ wavelength of a nearby RF transmission, a small AC voltage may
develop over the length of the shield. If the component driving the cable has a high
enough output impedance, the shield voltage could be induced onto the conductors in
the cable, which are a similar in length between the discontinuities of the RCA connectors
at each end. To eliminate this possibility, a high-frequency capacitor can bridge the gap
at the un-terminated end of the shield, behaving as a short at RF frequencies, but an
open circuit at the highest audio frequencies.

Speaker Cable Shielding
Shielding of speaker cables is a waste of money and will probably compromise their
performance. Speaker cables are driven by extremely low impedance drivers in the amplifier
to a very low impedance speaker load. In this low-impedance environment, coupling of
low-level high-frequency magnetic or electrical fields will be miniscule and insignificant.
Shielding speaker cable can also cause an adverse performance impact by increasing the
capacitance of the cable. Better not to do it.

Power Cord Shielding
Shielding of power cables serves no useful purpose. Shielding will add significant
capacitance to the cable with minimal positive benefit. If you really need this, then
the shield should be spaced well away from the conductors (large diameter) to minimize
capacitance and avoid constraining the magnetic field lines that should couple between
the conductors. Empirical testing has shown that standard shielded 14 gauge stranded
power cord sounds less dynamic than unshielded 14 gauge stranded cord. The impedance
of the electrical system is extremely low and susceptibility to magnetic and RF fields is
extremely low for power cables so the benefit is questionable at best.

Use of Ferrites to stop RF

Several companies offer clamp-on and slide-on Ferrites. Some audio manufacturers
claim that their Ferrites stop RF currents from being "picked-up" by power cords and
other audio analog and digital cables. Ferrites are routinely used on computer internal
and external cables to block RF. What is really happening here is that these are
blocking radiated emissions from the computer so that the computer will pass FCC
and foreign emission standards (CISPR, CSA). Their purpose is NOT to prevent RF from
being "picked-up" by the signal wires. In some cases they have been added to internal
computer cabling to attenuate radiation as a band-aid after the design is complete.
These Ferrites will "round-off" the signal edges, removing much of the high-frequency
content. The energy that Ferrites absorb is turned into heat as they are lossy elements.

Ferrites on Interconnects
What happens when you put one of these devices on your interconnect? It adds
inductance to the cable causing it to be a low-pass filter (passes only lower frequencies).
The problem is: if it is a large ferrite, or the composition is not correct, it can roll-off the
high audio frequencies. Bad idea. Better to get a shielded cable if RF is suspected to be
a problem. Some very small Ferrite beads, however, can be useful in taming some
unshielded cables, such as the Kimber PBJ, but the ferrite should be installed on one
conductor, not clamped across both.

Ferrites on Speaker Cables
What happens when Ferrites are installed on speaker cables? This is a more interesting
question. Ferrites, with the right composition and size can be helpful for optimizing a
speaker/cable/amplifier combination. I would avoid using the large clamp-on Ferrites
used for EMI (Electro-Magnetic-Interference), since these generally add too much
inductance. The Image Clarifier offered by Empirical Audio is a device that works for
audio because the composition and size of the Ferrites is specifically chosen for audio
cables. The Ferrite cores should be installed on one wire, not across both wires. The
reason that Ferrites can improve the performance of speaker cables lies in their ability
to add inductance and loss to the cable. We believe that this inductance reduces the
natural resonances in a cable, which we believe can become audible through secondary
effects. The ferrite creates a low-pass filter at very high frequency.

Power Filtering and Conditioning

Much attention has been given recently to both power filters and power conditioners.
In general, the power line voltage should be sufficiently filtered by the input
transformer and filter capacitors in a well-designed audio component so that an AC
filter will be of little benefit.

AC Power Filters
Power line filtering may protect a component from damage from a lightening strike
(because it contains Thyristors), but in general, power filters will insert inductance
in series with the power feed, which may limit transient current when the component
needs it. Limiting transient current can become audible by limiting dynamics. This
should be audible only with component's with poorly designed power supplies.
Unfortunately, many superior audio components have inferior power supplies. These
components will likely suffer limited dynamics when a power-line filter is used with
them. As far as filtering-out power line noise and distortion, these filters do prevent
high-frequency noise voltages from getting to the transformer of the audio component,
but the transformer itself usually does a good job of filtering these as well, although
some toroidal transformers may pass relatively high frequencies. In this case, any
high-frequencies that do get through will be swamped-out by the capacitive filters in
the audio component power supply anyway. Extremely little high-frequency noise will
actually get to the electronics. This is why the reviews on power filters vary so much.
It is very dependent on the design of the filter and the audio component power supply.
It is possible that an AC filter could be beneficial if the AC noise is significant in your
area, the AC filter does not limit transient currents and your audio component has an
inferior power supply.

Active Power Conditioners
Active power conditioners are another matter. These can clean-up even the low-
frequency distortion in the power voltage waveform, making the power conversion
in the component’s transformer more efficient. They may also change the phase of
the current in an optimal way. Reviewers have noted that 80-90 Hz is an
improvement over 60 Hz in most audio components. This is probably due to the
fact that the losses in the transformer are reduced at higher frequencies or that
phase changes in the current make the voltage conversion process less lossy.
Alternately, the output voltage from the transformer may actually be increasing,
giving the component more headroom, but at the same time shrinking the voltage
safety margin for the transistor or tube devices. This can be dangerous for the
devices and their reliability. Whichever of these is really at work in a given
conditioner, it is clear that they can be beneficial, particularly in the case where
the power supply design in the audio component is inferior.

Ground Loops
A system without ground-loops is surprisingly unsusceptible to hum and RFI pick-up.
Unfortunately, most systems have them. Ground-loops are created when components
have their grounds connected both through their AC cords and with single-ended
interconnects. This creates one or more loops of varying diameters. Ground-loops
provide a means for strong magnetic fields, differences in ground potential of the
power system or strong RFI to interfere and cause audible noise. Lets look at each
case individually:

Magnetic Fields
If a changing magnetic field (such as 60 Hz) intersects one of the loops, then a
like current will be induced on the ground loop. If the resistance of the ground wires
of the interconnect that is part of the loop is relatively high, a voltage will develop
across this ground wire. The resistance of the interconnect ground wire is generally
higher than the power cord ground wire. This voltage will appear on top of the signal
that is being transferred across the interconnect.

Power System Potential Differences
Each separate circuit in house wiring has a separate ground wire that is usually
connected to all other ground wires at the panel. Currents are not supposed to
flow on these ground wires. However, in practice differences in AC potential can
develop between grounds of different circuits. Therefore, when two or more of
such circuits are used to power a single system, this ties together the grounds
with different potentials through the power cords and interconnect ground wires.
Currents between the different grounds can result, which can cause voltage drops
across the ground wires in the power cords connecting between the different
circuits. This is not in itself a problem, but the same voltages are developed across
the ground wires of the interconnects between the components that are plugged
into the two circuits. These voltages, usually 60 Hz hum, will appear on top of the
signal that is being transferred across the interconnect.

RFI Susceptibility
Strong Electromagnetic fields can excite one or more of the loops. This happens
when the size of the loop is fractionally related to the wavelength of the RF,
usually 1/2 or 1/4 wavelength. The loops can act as loop antennas, actually tuning-
in the RF signal by resonance of the loop-antenna. The RF signal can cause currents
to flow and a voltage to build-up on the entire loop, including the ground wires in
the power cords and the ground wires in the interconnects that make up the loop.
The resistance and inductance of the interconnect ground wire will cause a voltage
to develop that will appear on top of the signal that is being transferred across the
interconnect.

The recurring theme in the above three cases is the resistance of the ground wire
in the interconnect which causes a voltage to build-up. Therefore, one would
conclude that making this wire very low in inductance and resistance would help to
eliminate the noise problems. This is true, however the realization of this is more
difficult than it appears at first blush. Making the interconnect ground wire very
low inductance and resistance conflicts with the need for the same interconnect
to be low capacitance, which is the most important factor for interconnect sound
quality. Therefore, one can conclude that it is fairly straightforward to make either
a low capacitance interconnect or a low-inductance interconnect, but not both.
Strategies to accomplish this include using multiple small ground conductors (to
minimize skin-effect). In effect, the only strategy that achieves a low inductance
without significantly increasing interconnect capacitance is to rely entirely on the
self-inductance of the ground wires, which equates to multiple separate ground
wires that a intentionally decoupled magnetically from the signal wires.

Another strategy that can help eliminate at least the above RFI problem is to apply
filtering to the power cord ground wire. This creates a low-pass filter in the loop
comprising the power cord ground wires and the interconnect ground wires, but
does not create a filter in the interconnects for purposes of signal transmission.

Other strategies include strategic system placement and selection of single-circuit
outlets. Minimizing the physical diameter of each ground-loop by careful cable routing
can minimize the susceptibility to magnetic fields and RFI. Using a single AC circuit for
all components can minimize the chance of AC hum due to ground potential differences.

What is Dielectric Absorption?
Dielectric absorption is also referred to as "soakage" or "voltage retention. After
a capacitor is charged, it retains part of the charge, even after being discharged
and even if the conductors have been shorted together. Cables, like capacitors
behave as if they have an additional series of RC networks in parallel with the
primary capacitance, and it is these small distributed capacitances that retain
charge due to the high series resistance. To measure dielectric absorption, the
capacitor or cable is charged to some voltage for one minute, and is then shorted
for two seconds. After a one minute delay, the recovered voltage is read using a
very high impedance voltmeter. In some instances, a significant voltage "rebounds"
from the capacitor or cable. Dielectric absorption is calculated by dividing the
recovered voltage by the charging voltage, and is expressed in percent. Teflon,
polystyrene, and polypropylene dielectrics will yield the lowest dielectric absorption,
while PVC and vinyl will yield the highest. To insure that the audio waveform is not
altered by secondary "rebounds" of charge and the high-frequency "fine inner detail"
is preserved, it is prudent to use dielectric materials that have low dielectric
absorption in audio interconnects and speaker cables.

What is Dissipation Factor?
Dissipation factor is important for AC power applications, which includes audio power
transmission to loudspeakers. Dissipation factor is approximated by dividing Equivalent
Series Resistance (ESR) by the difference: capacitive reactance (Xc) minus inductive
reactance (XL), and is expressed in %.

DF = ESR/(Xc-XL)

Capacitive Reactance is calculated by:
Xc = 1/(2*p *f*C) = wC

and Inductive Reactance is calculated by:
XL = 2*p *f*L = wL

Reactance can be thought of as "AC resistance".
Dissipation factor is a function of age, frequency, and temperature. Dissipation Factor
is a combination of conductor losses and dielectric losses.

What is Q?
All capacitors (and cables) have an inductive and a capacitive component. At very
low frequencies, the cable appears primarily inductive and at higher frequencies
becomes primarily capacitive. Q or "quality factor" is a measure of how abruptly the
change from inductance to capacitance takes place. At the point of transition, the
cable is in resonance, so it appears like a pure resistance. Resonance is when it's
resistance is equal to the Equivalent Series Resistance or ESR. Resonance will occur
at precisely one frequency. Q can be calculated as:

Q = 1/DF

We believe that the Q of a speaker cable is important. If the resonance point can be
"tuned" to the right frequency, the cable becomes more like an ideal resistance and
the phase response becomes more linear as well. High-Q cables seem to sound better
and have better focus and clarity in stereo.

Direct-Immersion Cryogenic Treatment Study
Empirical Audio^® has been monitoring the dialogue on the web by reviewers
and audiophiles concerning the use of cryogenic cooling to treat cables and
circuit boards. We were curious as to the effect of direct immersion in LNO2.
It is well known that many metal tools are treated by slowly cooling to LNO2
temperatures and then slowly warming back up which changes their characteristics,
some metals in the same way as heat-treating. The tool process obviously requires
very special equipment, often capable of lower than LNO2 temperatures. The
accounts that we have read with interconnects indicate that only immersion was
done and in some cases treatment with only dry ice was claimed to make a positive
improvement, albeit short-lived. If there is an improvement, then obviously do-it-
yourselfers can treat their cables at minimal expense, although the hazards of
working with LNO2 must be taken seriously.

To investigate these claims, we conducted some of our own experiments.
First, two identical pairs of 1m cables were assembled, one of which the bare
pure silver conductors were immersed in liquid nitrogen for 12 hours (the insulated
container also had dessicant in it to dry the air to prevent ice buildup on the
conductors prior to immersion).

Results:
The first tests were listening tests. These tests were conducted as single-
blind A/B comparisons with the Empirical Audio staff only using our reference
system. The results from these were unanimous: that the treated cables
sounded like listening through a cotton-filled tunnel compared to the untreated
cables. Significant loss of image size and depth was noted. Also the dynamics
seemed to be constrained or compressed.

Secondly, we made empirical measurements, including R, L, and C at 1KHz, 10KHz
and 100KHz using an hp 4263B LCR meter. These measurements are listed below:

This result adds fuel to the controversy over whether L-R-C measurements are
sufficient to characterize interconnect cables. Based upon this result, we
believe the answer is no.

The third set of measurements was a transfer response test, sometimes
referred to as a TDT or Time-Domain-Transmission. This is essentially a
time-domain measurement of a voltage step with a high slew-rate.
The step was generated by an 11801B Tek TDR and the output was monitored
by the same instrument. What we are looking for here is differences, not
absolute measurements. The resulting waveforms are shown below:

The green waveform is the step voltage emerging from either of the two 1m
non-treated cables. Both waveforms for the non-treated cables were virtually
identical, so only one is shown here. The red waveform is the step voltage
emerging from the 1m treated cables. The two treated cable waveforms were
also virtually identical so only one is shown here.

The most obvious difference is that the red waveform has a lot more high-frequency
ringing and spikes on it and generally looks more ragged. The waveform of the
untreated cable (green) is smoother indicating fewer reflecting boundaries.

There are at least two explanations for this, but here are some possibilities:

1. The surface finish of the conductors has a lot of micro-cracks that are causing
   reflections at high frequency. If this is the cause, then there is a strong
   case for polishing the surface of the conductors.
2. The crystal lattice has a lot of microcracks, causing reflections at high
   frequencies. If this is true, then significant damage was done to the metal.
3. The cable bandpass has increased and is causing higher reflections due to
   more significant impedance mis-match to the 50 ohm instrument cable.

The effect that the direct immersion treatment had on the cables may also help
explain why ultra-pure metals and single-crystal metals seem to make superior
conductors for audio transmission.

Empirical Audio is in the process of obtaining metallurgical photos of sections of the
treated and non-treated conductors to gain more insight into the results. Updates to
this article will appear when we receive additional information.

S.F.N.
Copyright © 1998-2002 Empirical Audio^®. All rights reserved.
Information in this document is subject to change without notice.
Other products and companies referred to herein are trademarks
or registered trademarks of their respective companies or mark holders.