Skin Effect
Oct 14, 2001 at 2:07 AM Thread Starter Post #1 of 26

Tomo

DIY tube amps can be SHOCKING
Joined
Jun 22, 2001
Posts
1,067
Likes
14
Hello,

Electrons repell each other and generally distribute themselves at outer most layers. This depth sometimes is refer to skin depth, however, actually this is not so.

This effect is rather mild and you should not go and try to make your cables have maximum possible surface area per volume. However, if I were to build cables with diameter 1um, obviously this is not as conductive as my braided 22 guage interconnect.

I believe cables should be able to withstand use. For me this is reason enough to exclude the use of solid wire. Stranded wires can withstand much more numerous bending.

Tomo
 
Oct 14, 2001 at 3:32 AM Post #2 of 26
22ga is a good choice. At 20KHz, the center of a
22ga wire is conducting almost the same as at 1Hz.

On a long run, a single 10ga might just be affected enough to sound different from an equal
amount of braided 22ga.
 
Oct 16, 2001 at 11:54 AM Post #3 of 26
I have got great results in using the Heavy gage Coax used as transmission Line in ham transmitters for Loudspeaker Wire. i think it is somthing like RG6U or RG8u. it has a 13 ga. stranded Cent Cond. and is about 1/2" Dia.
 
Oct 16, 2001 at 4:22 PM Post #4 of 26
ppl,

I LIKE! Wonderful!

I would never of thought of coax for the _speakers_.

Interconects, yes. Big loads like speakers, NO.

Coax has wonderful shielding and loads of
bandwidth. It has to have less capacitance as
a lumped load than things along the Litz line.

Even such specialized coax has to be cheaper than
most "audiophile" cabling!

Where do you get it? Do you have part numbers?
 
Oct 16, 2001 at 4:26 PM Post #5 of 26
Yeah, that might be a good idea! I'd like to try it too.

Actually...

A W E S O M E !!

I've spent last two weeks discussing speaker cable options with friends and this comes at the right time. Part numbers please
smily_headphones1.gif
)
 
Oct 16, 2001 at 5:24 PM Post #7 of 26
If I remember right, 89259 is good for
interconnects, but it completely misses the point
of ppl's idea. ppl's coax has a center conductor
that you could use to jump start your car.
 
Oct 16, 2001 at 5:47 PM Post #9 of 26
Quote:

Originally posted by topher
If I remember right, 89259 is good for
interconnects, but it completely misses the point
of ppl's idea. ppl's coax has a center conductor
that you could use to jump start your car.


Yes I understand that but the dielectric is also important. Plus I have used multiple runs which increases the effective gauge. Anyway, you can still try his construction techniques (cross-connecting) with other coax cables as long as they have a good copper shield.

Quote:

As far as I am concerned, this is NOT how to use
coax, not for speakers or anything else.

http://www.geocities.com/jonrisch/FIGB.GIF


Sorry but I don't get this one
confused.gif
evil_smiley.gif
very_evil_smiley.gif
 
Oct 16, 2001 at 7:57 PM Post #10 of 26
The whole idea behind coax is that the fields are
COMPLETELY inside the outer conductor.

The "cross connection technique" is not beneficial.

You lose the immunity to outside field interference,
and you lose the BENEFITS of the "good" dielectric
inside and are in reality
using the ultra cheap pvc outside protective
coating as at least half of the "signal dielectric".

You can check this in E&M text books, but it is
very heavy reading.
 
Oct 16, 2001 at 8:42 PM Post #11 of 26
Hello,

Coax probably have very distinct freq response due to the fact that it is definitely a wave guide. So I think demensions are very important when you are using it to propagate numerous freqs at the same time. (I think you get low cut off ...) Some freq signal can travel without much loss but some can get attenuated. (refering to modes in waveguides)

Personally, I have been thinking that coax are the best choice to be used as interconnect between CD player to DAC. I do not know how it is with audio signal but I remember when I upgraded my system with braided teflon 22 guage cables there was audible improvement. ... (Compared to generic Coax)

Coax should sound very good given that you got the right freq responce. But note you cannot just use ANY coax. The best coax will have very specific dimensions.

Please note also that propagation speed varies with freq. So longer the cable, bigger the phase shift. SO keep cables short.

Tomo
 
Oct 16, 2001 at 9:07 PM Post #12 of 26
Tomo,
you seem to be thinking of microwave style wave
guide. That has no center conductor and does have
modes.

Coax with a center conductor handles everything
from dc up to a point where at which the
dielectric and maybe the outer "conductor" gets
lossy. It happens as low as tens of MHz in that
cheap coax with a really spars weave for an outer
conductor.

Possum,

It's not quite that bad. That response waited on
my screen for over an hour as hordes of people
stopped by my desk.

Work comes first...
 
Oct 16, 2001 at 9:54 PM Post #13 of 26
I remember my 4th year course on optoelectronics... The formulas for "modes" of electric and magnetic field in the optical fiber (which is essentiallly a waveguide as well) were the most complicated that I've ever seen. The only thing that came close was mathematics behind wavelets (e.g. JPEG 2000 image compression).

How about this one? It removes outside PVC layer and places heatshrink (good quality presumably) instead, but does the cross-connecting afterwards. What's the deal with cross-connecting anyway?

http://www.tnt-audio.com/clinica/ubyte2e.html

I'm thinking of scouring the town for these thick coaxes and just soldering them to bananas or spades and using main conductor only. Should the shield be connected to, say, amp ground at one side or just left alone?
 
Oct 16, 2001 at 11:37 PM Post #14 of 26
aos,
Do you want to make a speaker cable or an
interconnect? For a speaker cable, I really like
ppl's idea. I would love to find out what others
hear when they try it. I'll give it a whirl when
my speakers are done.

For interconnects, I can't see anything being
better than using the "Low Loss Satellite Cable"
that your Thorsten Loesch reference speaks of.

Just use it like any other interconnect. Both
conductors have to be connected. I would suggest
center to red and outer to black for those systems
that show signal and ground with that color code.

But, there doesn't seem to be any way to predict
what things will sound like aside from things
like high capacitance cables sounding bad when
they cause amps to oscillate.
 
Oct 17, 2001 at 4:27 AM Post #15 of 26
Quote:

Originally posted by topher
The whole idea behind coax is that the fields are
COMPLETELY inside the outer conductor.

The "cross connection technique" is not beneficial.

You lose the immunity to outside field interference,
and you lose the BENEFITS of the "good" dielectric
inside and are in reality
using the ultra cheap pvc outside protective
coating as at least half of the "signal dielectric".

You can check this in E&M text books, but it is
very heavy reading.


Absolutely true but this is much more important for the low voltages in interconnects compared to those in speaker cables. From my listening tests, the cross-connected cables sounded much better than simple coax cables. Again, that is my preference. Personally I now use Canare starquad speaker cable because I like the way it sounds better than the coax cables on the ASL waves/Wharfedale combo.
BTW - the outer jacket of Belden 89259 is FET not PVC and the dielectric is teflon.
 

Users who are viewing this thread

Back
Top