Partly hate myself for necro'ing this as my first post (what a thread to start off hah)- but I felt the need to ask (going into a diy hp/pre):
I didn't find any responses to the
comments made by Paul MacGowan addressed outside of the (possible) echo chamber within the blog's comment section & the included response from a "cable manufacturer" (hopefully I've lurked well & this doesn't clog the thread without expanding)
- Thanks guys. And of course there are as many opinions as people. My friend who is a cable manufacturer wrote to me the following, too long to reprint in Paul’s Posts, but ok here.
”
Skin-effect is the effect resulting from a delay in the change of the magnetic field relative to a change in current. The faster the change in the current (higher frequencies or a single fast transition like lightning), the more significant the relative delay in change to the previously existing magnetic field. The old magnetic field is in the way, forcing the new current to go around the existing magnetic field, therefor there is only 100% current density on the surface or skin of the “conductor.” Skin-effect is why you’re safe from lighting in a car or on a plane … Lightning is one-way, being in that sense DC, but the pulse is very high frequency, and so the magnetic field inside an airplane prevents lightning from traveling except on the skin of the airplane.
More skin-effect means less current density away from the surface of the conductor … the opposite of what you wrote. All conductors have 100% current density at all frequencies at the surface, and nowhere else (if the conductor is the universe of concern, of course no conductor is so perfect, whether due to dielectric involvement or proximity to other conductors). Skin-effect does not describe properties at the surface, it only describes properties away from the surface.
You are correct that it is higher frequencies which are more affected by skin-effect. The notion of some strands for bass and some for treble is pretty much BS, though I do recognize MIT’s point other electrical values could make the “treble strands” part of a high-pass network.
The common insanity in our world is that “we” are as wrong as the naysayers. When some engineer with poor use of language says/writes that “there’s no such thing as skin-effect at audio frequencies,” they are of course dead wrong, but they almost always mean “there is no relevant loss of amplitude at audio frequencies at the lengths I use in my home.” That second version is totally factual, while the first was as wrong as denying the earth goes around the sun.
“Our” insanity is that the normal enlightened response is to cite that no, even an 18 AWG (1.02mm) conductor has a 1/e, or 74% reduction in current density at 20kHz at the center. True, but so what, if the conductor was a donut with no center at all, it could still carry enough current to kill you and blow up your speakers at the same time.
What makes skin-effect a very real problem for analog audio is that there is a 1:1 rise in inductance along with the rise in impedance away from the surface of the conductor. This causes smearing in time: there is less and less current at progressively greater distances from the surface, that current is farther and farther behind the correct information at the surface of the conductor. The phase delay would be trivial if it were group delay, but as a smearing, the information arriving spread over time, it totally messes with our brain/computer’s expectations and algorithms. Depending on the particular AWG size, this causes a combination of seeming loss of high-frequencies at the same time as a subjective amplitude roller-coaster. The loss of bass weight, the very tight but constipated mid-bass, the seemingly projected upper midrange and “rolled-off” highs are all phenomena created in the brain … there is no relevant change or loss of amplitude at audio frequencies! … and of course not at 50 or 60Hz.
Much of the “brightness,” or “treble tilt,” or pity the stupid reviewers who write about cable having a “rising top end,” of a stranded cable is due to surface interaction between strands. Copper oxides are semiconductors, in an environment in which the conductivity of the connection between strands is dynamic, and the contact pressure (due to rule of rights) is also dynamic. This mostly explains the typically hashy sound of bare stranded conductors.
Silver-Plated conductors sound horrible for analog audio, while being an inexpensive way to improve the performance of a digital or video cable. At audio frequencies, the phase change from incrementally more conductive silver on the surface to less conductive copper suffering slightly higher inductance underneath the silver, creates in our brains that obnoxious tweeter-in-your-face effect. If calling this a “rising top end” were accurate, the world wouldn’t need fossil fuels for cold fusion, all the world would need is bad-sounding wire which provides more current coming out than went in!
A group of strands is a single conductor, subject to skin-effect. If the strands are anything other than concentric, and as the path of least resistance is towards the surface of the conductor, current is “forced” to jump from strand to strand — though the impossibly complex distortion mechanism of variable-conductivity variable-contact-pressure connections in-between strands. Simple fact is, stranded or solid, conductors have more audio frequency distortion the more they are bigger than 20 AWG. Only the imperfectness of the conductivity between strands prevents the skin-effect of a bare stranded conductor from essentially equalling the skin-effect of a same-AWG solid conductor
If the strands are magnet wire, the hash is pretty much eliminated, but the magnetic confusion remains, causing a typical out-of-focus effect, maybe more clearly revealed without the hashy-bright flaw to distract. As a litz cable (not the German word litz=strand, but the English) has to have an inner geometry such that over length every strand has the same impedance as every other strand at all frequencies, the strands have to occupy all possible positions within the conductor in order to achieve this average quality … which ensures maximum magnetic confusion for an analog audio signal, which is pretty much not a problem for a modulated signal feeding an FM broadcast antenna, or a radar installation, or whatever. (I have a sample of 6” diameter litz cable used for NORAD).
In part because the draw of AC by a power supply is modulated by the draw of power by the active circuitry, an AC cable exhibits (mis)behavior similar to that of a speaker cable. However, independent of AC cables being in some ways a full range audio cable, most of what goes wrong in terms of AC power transfer in a stranded cable is simply due to the 50 or 60Hz modulation.”
If taking this at face value, for signal connections - I should use solid 20awg (ptfe) uncoated[?], & 18awg for my power connections? [in this case my fuse is 10a so I guess I should go 16ga for safety to match]. Though I take it that the difference between a solid / stranded cable is negligible, is there any practical application for litz as a hookup wire (internal connections) for the purposes of our listening? Or is it more likely to introduce higher frequency distortions/harmonics?
Another
quote which added confusion
- In practical terms, the use of litz wire has been found to be a useful tool when designing wound components that operate from about 10 kHz up to 3 MHz. Below 10 kHz and above 3 MHz, the price vs. performance evaluations tend to favor other winding methods such as employing flat, edge wound conductors or evaluating foil based winding techniques.
...
One challenge to keep in mind when looking to design with litz wire is that the diameter of the twisted construction will be significantly larger than the diameter of a solid conductor of the same copper area. Due to the increased diameter, which can be in excess of 15% larger than a comparable solid conductor, the total number of turns and layers within the winding window as well as the copper fill factor must be evaluated carefully.
Reading around on RF sites it seems like the consensus there is that it only would have an impact on very high frequencies (potentially broadcast frequencies) - and in audio circles this seems to be a topic avoided dearly. If taking it as being 15% greater volume & run distance comparably to a solid core (or stranded) cable solution, it seems as though logically it would not be a worthwhile sacrifice (at least with regards to pre/power amplifiers, I don't know enough about digital conversion to speak to upper frequency possibilities when also considering digital clock frequencies)
Anyway - if you have some input on these questions/comments I'd be very thankful! Happy holidays, hope everyone is staying warm & well.