If you measured the frequency response and/or the phase the difference would be apparent. Distortion measurement does not detect this type of error.
Really?
Let's play the devil's advocate:
First principles first:
A transmission line subjected to current and voltage stores energy (this property is used in high-power radar pulse generators for example).
This energy distorts the local space-time fabric, and as a result increases the constants µ0 and ε0
This increase is signal-dependent, and thus affects the linearity of the line impedance, and as a consequence the signal delivered to the load.
Second principles:
The line conductors are subjected to electrodynamic and electrostatic forces: when a pair of conductors is subjected to a current, their magnetic fields will cause a force trying to separate them.
When a difference of potential is present, the result will be an attraction.
If the separating medium is not infinitely stiff, this will cause movements, changes in the lineic constants, and ultimately a change in the signal delivered to the load.
Other lowly practical details:
The dielectric, often polyvinyl chloride loaded with a cheap filler is not perfectly linear: its permittivity depends on the applied electrostatic field, and will also modulate the line properties depending on the signal level.
Ohm's law is an approximation, based on the assumption that charge carriers in conductors (free electrons or holes) are available in infinite quantities, and resistance's only cause is collisions.
At "normal" current densities, the approximation is true, but it breaks down at extreme current levels (yes, I mean extreme, like shattering the conductors from electrodynamic forces, or vaporizing them from heat)
All of this is quite real, but has to be taken with a pinch of salt, of course....
I can see a good use of those (true) argues to help the sells of some exotic and high priced audiophile cables.
Last edited:
Elvee, those are the valid comments and real effects, no doubt, however... the influence of those effects on the sound quality is MUCH lower than the one coming from the speaker itself. A good cable can decrease the influence of the speaker's impedance non-linearity to some extent, or make the situation slightly worth if the cable is not so good, but anyway, acoustic non-linearity is going to play a major role in perception of the sound reproduction quality.
Hold it.
Everybody forgot "memory distortion" by the speaker cable.
Here it is:
The resistivity of Copper has a pretty strong temperature coefficient. Resistance of copper wires do increase with temperature. ( The effect is large enough so as to be used as a mean to measure copper wires temperature. For instance, in the wire enameling process, they use this to monitor the wire temperature ).
So, the speaker cable makes distortion because it's resistance varies with current.
So you do need an heatsink all along the speaker cable. I should write papers, start a company and make a fortune with these new speaker cables.
Everybody forgot "memory distortion" by the speaker cable.
Here it is:
The resistivity of Copper has a pretty strong temperature coefficient. Resistance of copper wires do increase with temperature. ( The effect is large enough so as to be used as a mean to measure copper wires temperature. For instance, in the wire enameling process, they use this to monitor the wire temperature ).
So, the speaker cable makes distortion because it's resistance varies with current.
So you do need an heatsink all along the speaker cable. I should write papers, start a company and make a fortune with these new speaker cables.
Actually, with normal insulated copper wire, 1000 amps per square inch of copper is the rule. The insulation can be 90 C rated, and it works well.
I noted someone mentioned free space mu and epsilon changing with current? Oddly, I've not heard of or measured that, so cannot confirm such. Granted, 30 kilo amps and 20/30 tesla isn't big stuff, I'd have to defer to the high end audio for new physics I guess.
Using water cooled copper works quite well, but while it keeps the conductor size small, it does tend to dissipate a lot more.
Jn
I noted someone mentioned free space mu and epsilon changing with current? Oddly, I've not heard of or measured that, so cannot confirm such. Granted, 30 kilo amps and 20/30 tesla isn't big stuff, I'd have to defer to the high end audio for new physics I guess.
Using water cooled copper works quite well, but while it keeps the conductor size small, it does tend to dissipate a lot more.
Jn
For some Northern countries, we can use an antifreeze coolant instead of water, ensuring smooth operation when the system is powered-on at the temperatures below zero Celsius (before 1000 amps warm it up)
I believe, a far-end Zobel helps to "normalize" the speaker's impedance at HF (1-100MHz), preventing unwanted "reflections" through the cable.
That is the intent. However, normal cables require between 120 and 150 ohms. 10 ohms only maintains speaker impedance to hf.
A 10 ohm zobel helps amp stability if the cable is a high capacitance type, and the load lets go in the hf region where the amp still has gain over 1. Without the continued load at hf, you lose phase margin.
Jn
Here are the explanations why the far-end Zobel makes sense:
Nelson Pass - Speaker cables
Rod Eliott - Loudspeaker cable characteristic impedance
Nelson Pass - Speaker cables
Rod Eliott - Loudspeaker cable characteristic impedance
It can make a big difference.
If the amp is "hot" (more than unity gain at very high frequency) and the user chooses low inductance high capacitance cables, the cable capacitance will present as capacitive to the amp if the load decouples at high frequency. If the amp has greater than unity gain and the capacitance causes phase margin to pass zero, the amp will oscillate.
A zobel at the amp end will not present to the amp the same way it does at the load end, so may not do the intended function.
Jn
I am quite familiar with both sites, and their discussion and conclusions are consistent with what I've stated, thank you for the links.
I don't care for Rod's writing style, as he refers to cable "3" but doesn't state that cable 3 is not mentioned until paragraphs later. (I've been editing papers of late, so have been keen on finding such inconsistencies). I note as well that his cable 1 has too low a Z, so suspect he chose a zip with very tight spacing between conductors.
The conclusions are the same, that to prevent oscillation issues especially with low capacitance cables, the line impedance has to be matched to the zobel.
In the discussion in this thread, the cable has been bog standard zip, so the cable impedance will be in excess of 100 ohms, a zobel at that value will suffice.
Edit: I suspect Nelson simply assumed that a zobel built into the load at load Z wouldn't affect performance with zip style cables but would aid should the customer choose a high capacitance cable. A reasonable design choice, but I was hoping he would elaborate that choice for all.
Jn
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.