Cable nonsense -- article #4
John Dunlavy
The large number of recent postings regarding audiophile cables and
loudspeaker design is encouraging. Perhaps, it is indicative of a
newfound level of interest in the way cables work and perform. Several
posts raised questions and or proffered information that deserve
comment. Unfortunately, my cramped work schedule leaves little time
for writing individual replies to everyone. Therefore, I will try and
lump related answers together and attempt to cover as much important
territory as time allows.
For those who asked how impulse response, step response, amplitude
Vs. frequency response and phase Vs. frequency response are related to
one another, lets consider the following. The impulse-response of any
linear analog network, including amps, loudspeakers, cables, etc., is
important because it contains information about virtually all other
measurable and audible performance properties. Beginning with a
measurement of impulse-response, the frequency-response,
phase-response, cumulative-decay-spectra, step-response, energy-time
response, etc., may be rapidly and accurately determined by FFT
analysis, such as that provided by the now well-known, computer-based,
MLSSA measurement system. (We have three MLSSA systems running
full-time for R&D and production QC applications, in addition to
spectrum analyzers, distortion analyzers, vector-impedance analyzers,
complex waveform generators, etc.)
Further, in answer to another question posed on the NET, variations in
phase Vs. frequency within a linear system are the "first derivative"
of variations in amplitude Vs. frequency. And, variations of amplitude
in the "time domain" produce variations of both amplitude and phase in
the "frequency domain". Indeed, virtually all measurable performance
attributes of any linear system, whether it be an amplifier, a
loudspeaker, a cable, etc., are related to each other in relatively
simple ways that are easily treatable by mathematics - an extremely
powerful tool for those who understand and know how to use and apply
it.
Several posts seem intent on taking issue with what I said about
low-loss, low-impedance loudspeaker cables causing some
poorly-designed power-amps (with a slew-rate exceeding stability
limits created by an improperly designed inverse-feedback loop) to
oscillate. One recent post said: "This is the third time you have
ascribed high slew-rate amplifiers to the problem of cable interface.
This is misleading. It's also the third time I have contradicted you
on this point, which is why I'm sending this reply directly via email
this time (as well as to the ng)".
But in my post on the subject, I never directly related "slew-rate" to
oscillation without the caveat: "... created by an improperly designed
inverse-feedback loop". Indeed, the following text (exactly as I
posted it on the NET) is the relevant paragraph that seems to bother
this particular contributor:
"Most "seemingly" unexplainable, yet truly audible differences between
cables, can be explained if critically examined with respect to
equipment interface considerations. For example, a well-designed,
low-loss loudspeaker cable (with a relatively-low
characteristic-impedance of perhaps 6 to 8 Ohms) can cause many
expensive, well-regarded power-amps (with a slew-rate exceeding
stability limits created by an improperly designed inverse-feedback
loop) to oscillate at frequencies well above the audio range. This is
sometimes audible as a low-level, high-frequency "crackling noise"
(usually emitted by the tweeter as it's voice-coil is being
cooked). Such amplifier instabilities may also alter the "sound" of
the amplifier by creating an "edgy" quality on musical transients or
an exaggeration of high-frequency notes, etc.. But the amplifier, in
this case, is at fault - not the loudspeaker cable."
From the above, I fail to grasp how this person interpreted my
comments as inferring that I believe amplifier stability is directly
related to slew-rate - alone! Far from it, for some of the best
power-amps I have heard and/or tested exhibited very high slew-rate
performance - obtained by using proper high-frequency transistors in a
"minimalist circuit configuration with relatively little
inverse-feedback". I sincerely hope that the above comments set the
record straight and that I do, indeed, understand network/circuit
theory, transmission-line theory, amp design, slew-rate, stability
margin, inverse-feedback problems, etc.
One post on rahe recently noted that, "I've been following
Stereophile's analysis of time-coherence for a while now, and have
noticed that almost none of the speakers reviewed are time-coherent,
including those which received excellent ratings." Without attempting
to justify "excellent ratings" sometimes given by Stereophile for
loudspeakers that do not exhibit "time-coherent" performance (good
impulse, step, waterfall and energy-time responses), their reviews are
most often an amalgam of two different approaches for judging
"accuracy": 1) subjectively perceived accuracy (based upon listening)
and, 2) objective accuracy (determined by assessing a full-set of
accurate measurements). The best reviews, in my opinion, are those
that compare the results of both and attempt to resolve and explain
any lack of correlation that might exist. Subjectively determined
accuracy, taken alone, is an unreliable means for establishing the
acoustical merits of audiophile components. This is because even the
most honest attempt at determining accuracy by listening, is subject
to personal experience, preferences, whims, long and short-term
memory, program material, equipment interface problems, listening room
modes, etc. Also, one reviewer might consider a "warm, mellow sound"
to be most accurate while another might be attracted by a "more
detailed, analytical sound" and so forth. If a multi-member group
listens to a system and attempts to arrive at a consensus regarding
its accuracy relative to some "standard", the danger exists that the
strongest-willed member may, without consciously intending to do so,
inadvertently impose his or her choice on the other listeners.
Several individuals have inquired as to why we designed and sell our
own loudspeaker cables and interconnects. The answer is simple: we
believe that most audiophile cables are very over-priced, do not
perform as advertised and do not provide the technical properties
required to insure the best possible system performance (taking into
consideration system interface problems). For example, most
interconnect cables exhibit a sufficiently high capacitance (typically
in excess of 30 pF/ft.) to cause non-linear distortion at
high-frequencies when used with some pre-amps and power-amps. The
relatively inexpensive top-of-the-line Radio Shack interconnects are a
shinning example of an excellent performing, low-capacitance cable
(typically about 15 pF/ft.) that is very, very affordable. Our own
interconnect cable exhibits nearly half the capacitance but is a bit
more expensive - though very affordable for most audiophiles.
With respect to loudspeaker cables, we measured most of the best known
and most expensive audiophile brands and were shocked to find that
little correlation existed between selling price and measured/audible
performance. If you read back to some of my earlier postings on the
subject, you will discover that I covered the matter in a reasonably
thorough manner. We will continue to design and market our own cables
to meet a consumer and professional demand for cables offering
credible performance, based upon solid engineering criteria and
accurate measurements of all relevant performance parameters - at very
affordable prices. While we do so, we also tell audiophiles and
professional users that, especially for relatively short lengths of
cable, there appears to be no consistently audible difference between
most loudspeaker cables (including high-quality #20 AWG Zip-cord). The
same applies to most interconnect cables, regardless of their cost.
But, in my opinion, it costs no more to design and manufacture cables
that conform to the dictates of good engineering practice than those
cables whose properties and performance are very questionable. So, why
not do so - and give customers a break from all the flooby-dust,
buzzard-salve, snake-oil and hokum that surrounds the advertising of
too many of today's cables?
Best Regards,
John Dunlavy