Manley SKIPJACK - Manual (Page 9)

9
One thing that the Skipjack can be used for is comparing cables. In practice, this is not as quite as easy as one might
hope. We can use it to illustrate some points though.
Lets say you want to compare two sets of good audiophile interconnects. At the source, you might use a pair of
simple cheap “Y” adapters. This feeds the two sets of interconnects, which feed the Skipjack. Then you need a third
set of interconnects to feed the preamp. Obviously, you might lose a little integrity in the “Y” adapter, and probably
an similar small amount in the Skipjack. Finally you need another set of interconnects to feed the preamp. Maybe
that is OK because you are only trying to perform a “relative” comparison, not absolute, and you are just looking for
audible differences. What if the degree of audible difference is significantly reduced by the test set-up?
A little deeper analysis of the set-up and cable phenomena will help. What are the possible issues in cables? The
commonly definable characteristics of an interconnect are the wire (signal conductor) material (metallurgy), size
(gauge), shape (geometry), the insulation (dielectric) and its electrical and mechanical characteristics, plus the shield
and all of its electrical and mechanical characteristics and how effective it does its job as a shield across a very wide
range of frequencies. All of that is assembled into a cable that is considered to have measurable resistance,
capacitance and inductance, which combined is termed “impedance” that more or less suggests a complex resistance
versus frequency versus length. And then connectors are attached, which brings in yet more complexity to the picture.
At the same time, we should be keeping in mind that after all it is wire that we are talking about and these electrical
characteristics tend to be very small and difficult to measure for typically short interconnect lengths. In fact, there are
many that suggest those measurements are so small that they cannot possibly be an audible factor, and that wire is
just wire, and if there does seem to be some audible difference then one is obviously just unfit for audio. Maybe it is
not quite that simple.
Back when the Ma Bell was still a young girl, a toddler in fact, and they were beginning to try to run longer
telephone lines and the first lines across the English channel, a scientist was given the task of reducing the
“distortion” that was occurring. This was 1890 and his name was Heavyside and the “distortion” was not so much
amplitude related as it was time domain related, an echo. In solving the problem, he provided us with the
terminology, formulas and most of what we still refer to as impedance and its relation to resistance, capacitance,
inductance, wires and termination. He wrote the book. The problem he was dealing with were cable reflections and
part of the solution is the 600 ohm terminations still in use today to provide maximum power transfer. The 600 ohm
standard was used in pro audio until the mid 70’s.
Today, impedance and signal reflections are one of the biggest problems for Moore’s law and ever faster & cheaper
computers, and a rather big issue for most signals over 1 megahertz where proper termination practices are essential.
The two most common techniques for looking at signal reflections are A) using a square wave generator and an
oscilloscope to look at the damage to the wave shape, B) using a swept sine wave and looking at the frequency
response. Notice we said “looking” and not “listening”. We can usually hear artifacts below –60 dB and that is 1000
times lower than the signal and practically invisible on a scope. Maybe the guys that say that cable reflections are
irrelevant at audio frequencies and cable lengths have based that opinion only on “looking” not “listening”.
Recent hearing research also suggests a range of values to describe our acuity in hearing in the time domain and these
times range from 10-20 microseconds at the most conservative to 1-2 microseconds from the most bold. It might be
noted that the tests are typically done with low fidelity gear, audiologist headphones and artificial test signals, not
music and high end systems and that 1-2 microseconds corresponds with a 500 kHz to 1 megahertz wave period and
it corresponds to our ability to detect pitch (.1% at 1kHz). In other words, this business of ‘time’ is worth
considering carefully.
Back to our scenario comparing two high-end interconnects. Our source is driving 2 cables in parallel, which means
that the total capacitance is the sum of both, and if they started out equal, then the capacitance is double. This could
conceivably mess up a carefully designed impedance. If we call one interconnect A and the other B, then we get
reflections from A, plus B and AB depending how low the source impedance is. Probably more importantly, we also
have that final set of interconnects from the Skipjack into the preamp. If that impedance is different, we are setting up
reflections (like light diffracted through a glass of water) that might be described as A, C and AC, and depending on
the source impedance again B, AB, BC and ABC. We should point out that the source impedance may be dependent
on the line driver and negative feedback through an op-amp. In other words, it too is complex, not exactly
instantaneous, and not necessarily flat and zero.
COMPARING CABLES PT.1