The cable conundrum
- Thread starter DaveyF
- Start date
Apologies for the ambiguous style - "cables can sound different and can improve audio systems " should be understood in the sense that your system can sound different when you change cables, and the choice of proper cables will improve the sound quality of the system. IMHO the difference is real and is perceived as such.
I assume that the proper cable can improve the sound quality of a system. Sound reproduction is a compromise incorporating many non ideal contributors, cables can be an improvement factor in this mix.
Single person experiments have very limited value, and as such are only informative, but as I have used monoblock amplifiers with 20 feet XLR interconnect cables I have more than once tried direct wiring of the crossover to the amplifier output connector and it never sounded better than with a properly chosen cable.
Thanks for the response, ack.
This statement is in the MIT white paper: “Without a proper inductive component, current cannot be stored, and the cable will have difficulty transporting active power to the load.” That statement makes little sense to me. An inductor will try to maintain the same current flow through it, so I assume that is what is meant by “current storage”. I am not sure what “transporting active power” means. All wires have some degree of inductance, and I have the feeling the technical content was garbled trying to make the point.
Later sections make the point that, at low frequency, the inductance is small, but I fail to see how that causes great audible degradation. Here’s a quote:
“Most audio cables have problems in this area between zero Hertz
(DC) and 1 kHz. It is not uncommon for audio cable with poor
inductive characteristics to lose nearly all of the inductive phase
angle at low frequencies (<100 Hz). Such a deviation from the ideal
is shown in the non-ideal phasor diagram in Appendix B, Figure
3. This loss of inductive phase angle has tremendous implications
on audio quality.”
For maximum power transfer, you want PF = 1, assuming they define power factor the same as everyone else (real power delivered to the load / apparent power in the circuit). However, it looks like they are defining the PF of the cable itself, an interesting approach to prove their point. They then use this to determine that if the inductance is not “proper” then voltage and current are not in phase and thus the cable’s PF represents “badness” if not = 0 (purely reactive). I am not sure treating the cable this way is valid (neglecting the source and load and what happens at the ends where the power is generated and used).
Graph 2 in the paper is interesting, as it shows that most cables transport 0% active power at LF. This says all LF power is reactive; that is, voltage and current are out of phase. It seems like something is amiss with their application of PF and T-line theory, but I may be glitching because they are applying PF to a cable instead of the system the way PF is normally used.
Multi-gauge meaning multiple different-gauge wires insulated from each other and bundled together? Such a situation normally requires a 3D field solver to model, quite complicated, at least at RF. I assume they are using something simpler, but doubt they would say exactly how they are combining. Presumably a coupling model they have derived, empirically or theoretically, that matches their measured results.
Hmmm… Pretty subjective impressions and I am not quite sure what to do with them. The MIT white paper applies T-line theory at audio and claims the low inductance of most cables is a problem which they have solved with their special construction techniques. Adding series inductance will tend to emphasize bass and reduce highs, though how much change there is due to a cable is hard to say. The goal appears to be to eliminate loss, which can be achieved by reducing the resistance, and maximizing real (“active”) power through matching networks. Given the fairly fixed impedance at audio of most low-level sources and loads, and widely variant impedance of most loudspeakers, they would have to design the network to match the source and load to achieve maximum power transfer in the system. However, by using a combination of passive components, they could certainly emphasize (or e-emphasize) different frequencies that would provide a pleasing sonic result.
So, they want to create a more capacitive cable because capacitance (C) varies less than inductance (L)? The only way to reach -90.000 degrees is with a purely capacitive cable, i.e. no resistance and no inductance, or some network that achieves that goal. Note that is not the goal of most transmission lines; an ideal lossless line has zero R, infinite G, and sqrt(L/C) such that the ratio of L and C determine the desired cable impedance. Optimum power transfer occurs when the line’s impedance matches the load.
This statement is in the MIT white paper: “Without a proper inductive component, current cannot be stored, and the cable will have difficulty transporting active power to the load.” That statement makes little sense to me. An inductor will try to maintain the same current flow through it, so I assume that is what is meant by “current storage”. I am not sure what “transporting active power” means. All wires have some degree of inductance, and I have the feeling the technical content was garbled trying to make the point.
Later sections make the point that, at low frequency, the inductance is small, but I fail to see how that causes great audible degradation. Here’s a quote:
“Most audio cables have problems in this area between zero Hertz
(DC) and 1 kHz. It is not uncommon for audio cable with poor
inductive characteristics to lose nearly all of the inductive phase
angle at low frequencies (<100 Hz). Such a deviation from the ideal
is shown in the non-ideal phasor diagram in Appendix B, Figure
3. This loss of inductive phase angle has tremendous implications
on audio quality.”
ack said:
For maximum power transfer, you want PF = 1, assuming they define power factor the same as everyone else (real power delivered to the load / apparent power in the circuit). However, it looks like they are defining the PF of the cable itself, an interesting approach to prove their point. They then use this to determine that if the inductance is not “proper” then voltage and current are not in phase and thus the cable’s PF represents “badness” if not = 0 (purely reactive). I am not sure treating the cable this way is valid (neglecting the source and load and what happens at the ends where the power is generated and used).
Graph 2 in the paper is interesting, as it shows that most cables transport 0% active power at LF. This says all LF power is reactive; that is, voltage and current are out of phase. It seems like something is amiss with their application of PF and T-line theory, but I may be glitching because they are applying PF to a cable instead of the system the way PF is normally used.
Multi-gauge meaning multiple different-gauge wires insulated from each other and bundled together? Such a situation normally requires a 3D field solver to model, quite complicated, at least at RF. I assume they are using something simpler, but doubt they would say exactly how they are combining. Presumably a coupling model they have derived, empirically or theoretically, that matches their measured results.
Hmmm… Pretty subjective impressions and I am not quite sure what to do with them. The MIT white paper applies T-line theory at audio and claims the low inductance of most cables is a problem which they have solved with their special construction techniques. Adding series inductance will tend to emphasize bass and reduce highs, though how much change there is due to a cable is hard to say. The goal appears to be to eliminate loss, which can be achieved by reducing the resistance, and maximizing real (“active”) power through matching networks. Given the fairly fixed impedance at audio of most low-level sources and loads, and widely variant impedance of most loudspeakers, they would have to design the network to match the source and load to achieve maximum power transfer in the system. However, by using a combination of passive components, they could certainly emphasize (or e-emphasize) different frequencies that would provide a pleasing sonic result.
FWIW, I have measured several network cables (MIT and Transparent) many years ago, and they were flat within 0.1 dB (the resolution of a good Tektronix) between 20 and 20 kHz. I think that such a variation can not justify a "pleasing sonic result" per se alone.
Here is a nice paper with simulations and actual measurements of Monster cables and generic: http://dspace.mit.edu/bitstream/handle/1721.1/46225/41567257.pdf
It has interesting measurements like THD of cables! Never thought about doing that.
It has interesting measurements like THD of cables! Never thought about doing that.
I was really interested in the test set-up, not only the signal source. But already this is starting to look odd. The HP you cite is a component tester - that means it tests single port networks only. A cable is a two port network - to test a cable they'd have needed a network analyser.
Power factor would be the phase angle looking into the cable, but they only have an LCR meter so they can't measure this without a load being applied at the other end (or perhaps shorting the speaker end). Do they say whether they used any kind of load? Also, do they say why they need to measure separately? I'd like to see justification for this. Presumably they're talking about different sized wires in parallel - in which case why measure separately because when you put them close together there will be mutual inductance?
They didn't use any kind of simulated load so how could they calculate the effect on FR without knowing the load impedance?
Overall from the details so far given this does not look like particularly thorough work to me, and that's a typically British understatement
Typical idiot engineer - focusses 100% on a cable's differential mode performance, no attempt whatsoever to consider common-mode effects. Even though he records he was told by a couple of salesmen that cables are system dependent, he still didn't make the connection.
This was a great find Amir!
I read the entire paper, and the area around page 35 is particularly telling.
I read the entire paper, and the area around page 35 is particularly telling.
Last edited:
Apt holman preamp specs are easily located. http://kenrockwell.com/audio/apt/holman-preamplifier.htm
Begins with the assumption that the high end in general and and cables in particular are a rip off. Then sets out to prove it. Ignores anything contradictory.
Begins with the assumption that the high end in general and and cables in particular are a rip off. Then sets out to prove it. Ignores anything contradictory.
That pre looks to be an interesting design - several points of detail addressed, like variable gain stage for volume. But other points of detail missed, like minimizing input-stage derived distortion in non-inverting connected opamps (TL072 he uses).
Concur with your analysis Greg - reads more like a work of debunkery than of science.
Concur with your analysis Greg - reads more like a work of debunkery than of science.
I didn't post the paper for its conclusions. I post it since it has a lot of measurements in there.
I'm familar with this guy with photo gear reviews. I see him as sort of a loon. Conundrum ? Perhaps in finding the best match for one's system...as far as cables sounding different...those that don't believe need a hearing test, given respectable gear and setup..
The measurements are fine, made with decent kit. Just so happens though that they're the wrong ones if we'd like to explore why cables sound different.
So why did Mr Cooper decide to make those particular measurements and not others? I have some ideas myself but they pretty much boil down to intellectual laziness. What do others think?
I post it since questions were asked about frequency response of cables. He did that test plus some.
I don't think he was lazy at all. He ran tons more tests than any other report like it that I have seen. He tested dozen plus cables. He is testing using the same analysis that we use for audio equipment in general: frequency response, phase, noise, THD, etc.
Have you run the tests that you are talking about and can share the results? And what exactly do you want tested? If you can describe it and I can use my Audio Precision with it, I can generate the data for the cables I have.
Similar threads
 | Steve Williams Site Founder | Site Owner | Administrator |  | Ron Resnick Site Co-Owner | Administrator |  | Julian (The Fixer) Website Build | Marketing Managersing |