Cat5 network wire can make a good speaker cable but to get the most out of it you should untwist the pairs.

Cat5 cable consists of four twisted pairs of 24-awg wires where each pair is intended to carry a discreet digital signal. To minimize the cross-talk between the signals, each of the pairs has a different twist pitch, i.e., different number of twists per inch. The result is each pair has a different electrical path length. As long as each pair is carrying a discreet, digital signal, the difference in path length is inconsequential. But when multiple twisted pairs are used in parallel to carry an audio signal, the path length differences smear the signal.

Here’s what I did: It’s a lot of work but I think it was worth the effort.

First, I cut several lengths of Cat5 cable, stripped off the outer jacket, and untwisted the pairs.

Next, I trimmed each individual wire to the same length and twisted them in groups of four in a star-quad configuration. Because each wire starts out with the same length, every electrical path through the composite cable will have the same length regardless of how the wire is twisted.

Finally, I twisted four star-quad strands together to make a basic 15-awg cable. (http://internet.cybermesa.com/~jmlpartners/images/DIY/c2.jpg)

The basic cable can be doubled or tripled to make a 12- or 9-awg cable. (http://internet.cybermesa.com/~jmlpartners/images/DIY/sqc5.jpg)

Did it sound cleaner?
I was experimenting last week without unwinding, just using stripes with stripes etc. and it did sound smeared.
What you say makes sense.

Sort of what I suggested in another thread about keeping the colored pairs (Br-Or / Bl-Gr) together rather than separating them into solids and stripes.

Did it sound cleaner?

Yes, but I think "coherent" is the word that best describes the difference I heard.

One question. How did you terminate the ends?

Paul

Sort of what I suggested in another thread about keeping the colored pairs (Br-Or / Bl-Gr) together rather than separating them into solids and stripes.

Using the existing twisted pairs as a single conductor may reduce, but will not eliminate, path length differences. And, compared to the star-quad geometry, it will increase, rather than reduce, the overall inductance of the cable.

This sounds like the same thing described in length on TNTaudio. Even using the same name star quad.

One question. How did you terminate the ends?
I chose to terminate stripes with stripes and solids with solids. That gives the lowest inductance at the expense of higher capacitance. Another choice would be to use all four conductors of each quad-star strand as a single conductor and consider the basic 4x4 cable as the primary star-quad. That reduces capacitance at the expense of higher inductance. Looking at the cross-section of the basic cable, you can terminate it as four parallel star-quads or a single star-quad.


- + - +
+ - + -

- + - +
+ - + -

or

+ + - -
+ + - -

- - + +
- - + +

This sounds like the same thing described in length on TNTaudio. Even using the same name star quad.

"Star-quad" is a discriptive term used to identify a particular cable geometry. It is not propritary and is little different then the terms "twisted pair", "coax", or "zip cord".

Using the existing twisted pairs as a single conductor may reduce, but will not eliminate, path length differences. And, compared to the star-quad geometry, it will increase, rather than reduce, the overall inductance of the cable.
OK, but if [+] is a signal feed and [-] the return in the circuit, then for the minute difference in the variation in the number of twists per linear foot per pair, the path length difference concern is moot. We're talking speaker signals, not digital backplanes with critical time frame constraints for receiving information and for data settling before processing the information. Calculate in the path length difference as the signal passes through the various crossover components before it even reaches the speakers with their variations in voice coil winding lengths and... :scratch2:

OK, but if [+] is a signal feed and [-] the return in the circuit, then for the minute difference in the variation in the number of twists per linear foot per pair, the path length difference concern is moot. We're talking speaker signals, not digital backplanes with critical time frame constraints for receiving information and for data settling before processing the information.

I think you will agree, a very large path length difference will be plainly audible. So the question is: What is the minimum path length difference that is audible? Do you know what it is? I don't except that I hear a difference between braided Cat5 pairs (VenHaus style) vs. braided Cat5 cables (TNT style) vs. the method I described.

Yes, but I think "coherent" is the word that best describes the difference I heard.

Coherent strikes more truly.
I don't really understand the science of it all but I do notice sound differences. I've got some TNT recipe triples that I made a long time ago and while they are ok they do seem grainy in the high end.
In my office system which changes all the time, I use cat 5e double strands that are measured equally and sound much truer to my ears.
Anyway this is a good topic for discussion and Welcome to AK.

Calculate in the path length difference as the signal passes through the various crossover components before it even reaches the speakers with their variations in voice coil winding lengths and... :scratch2:

Excelent point. The speed of an electrical signal down a wire approaches the speed of light so the wavelength of a 20,000 cycle per second signal is a whopping 15000 Meters!

It would take a 1500 meter differential to acheive a 10% phase shift and a 1.5 meter differencial would be a .01% phase shift.

Could a human ear detect a phase shift in the order of that magnitude? How does it relate to the magnitude of phase shift accountable to other characteristics of the entire system? (such as the forementioned cross over networks.)

A secondary question would be "what is the characteristics of Cat5 individual strands that when unwound and then braided back together that would make them diferent than any other solid core wire"?

Steve

Steve

Could a human ear detect a phase shift in the order of that magnitude? How does it relate to the magnitude of phase shift accountable to other characteristics of the entire system? (such as the forementioned cross over networks.)
The phase shift produced by cross-overs is quite audible. Just ask any advocate of single-driver speakers. I hear it too, but for me, the shortcomings of single-driver speakers overshadow the benefits.

A secondary question would be "what is the characteristics of Cat5 individual strands that when unwound and then braided back together that would make them diferent than any other solid core wire"?
The cable construction I described combines the benefits of:

1) Litz wire (multiple, individually insulated strands) to reduce skin effect when compared to single core wire of the same gauge.

2) Star-quad geometry to reduce inductance when compared to parallel conductors of the same spacing.

3) Uniform electrical path length when compared to other Cat5 cable that uses intact twisted pairs.

4) The low cost and abundance of Cat5 cable.

5) The pleasure of creating something different and most likely better than any other cable that can be made or purchased for the same cost.

The result is a more coherent sound when compared to any other Cat5 cable.

T
The cable construction I described combines the benefits of:

1) Litz wire (multiple, individually insulated strands) to reduce skin effect when compared to single core wire of the same gauge.

2) Star-quad geometry to reduce inductance when compared to parallel conductors of the same spacing.

3) Uniform electrical path length when compared to other Cat5 cable that uses intact twisted pairs.

4) The low cost and abundance of Cat5 cable.

5) The pleasure of creating something different and most likely better than any other cable that can be made or purchased for the same cost.

The result is a more coherent sound when compared to any other Cat5 cable.

Thanks for you reply.
I agree that you point 1 - 3 will impact the signal in a positive way I do not believe it would be detectible by a human ear. But thats my opinion only virtually impossible to prove without a blind A/B comparison that would allow impartial (and possibly unsuspecting) people to listen to one cable immediately after the other. But I don't hink it would hurt!

Items 4 and 5 I think are the most significant and why i'll replicate what you have done for my speaker cables. I can get the stuff for free so why not?

Thanks
Steve

I think you will agree, a very large path length difference will be plainly audible. So the question is: What is the minimum path length difference that is audible? Do you know what it is? I don't except that I hear a difference between braided Cat5 pairs (VenHaus style) vs. braided Cat5 cables (TNT style) vs. the method I described.
With all due polite respect, you're comparing the taste of apples with oranges. Analog and digital signals are two different entities, particularly when using specifications designed for data cable and assuming the same issues exist for audio cable use.

The "twist rate" differences in any Cat X cable denotes the calculated data rate (speed) a digital signal can travel down a specific length of wire.

Cat 3 has a typical twist length of 7.5 to 10 cm.
Cat 5 had a typical twist length of 0.6 to 0.85 cm.

Now, in regards to the concern over signal path length, the Nominal Velocity of Propagation (NVP) refers to the inherent speed of signal travel relative to the speed of light in a vacuum (designated as a lower case c). NVP is expressed as a percentage of c, for example, 72%, or 0.72c. All structured wiring cables will have NVP values in the range of 0.6c to 0.9c. Similarly, if you know the physical length and the delay of a cable you can calculate the NVP. In most instances, length is derived from the shortest electrical length pair in the cable.

Terms and other things to keep in mind:

The speed of sound in air is approximately 331.5 m/s at 0 degrees C or around 1200 km per hour.

The speed of light is 299,792,458 m/s (or 1,079,252,848.8 km/h)

Attenuation – All electromagnetic signals lose strength as they propagate away from their source, and LAN signals over cabling are no exception. The loss of signal strength in the cable is attenuation. The more attenuation you have, the less signal is present at the receiver. Attenuation increases with both frequency and length.

NEXT – When current flows in a wire, an electromagnetic field is created which can interfere with signals on adjacent wires. As frequency increases, this effect becomes stronger. Each pair is twisted because this allows opposing fields in the wire pair to cancel each other. The tighter the twist, the more effective the cancellation, and the higher the data rate supported by the cable. Maintaining this twist ratio is the single most important factor in any successful UTP installation for digital data networks.

Delay Skew – Propagation delay skew (skew) is the difference between the propagation delay on the fastest and slowest pairs in a UTP cable. Some cable construction employ different types of insulation materials on different pairs. This effect, in addition to unique twist ratios per pair, contributes to skew.

Propagation Delay – Propagation delay, or delay, is a measure of the time required for a signal to propagate from one end of the circuit to the other. Delay is measured in nanoseconds (nS). Delay is the principle reason for a length limitation in LAN cabling. (That's LAN cabling, where digital data timings are critical as compared to analog audio signals.)


Because of delay skew, the length of the four pairs often appears slightly different. This is normal and no cause for concern with the exception of significant (over 10%) variances. This applies to digital data transmissions with limited set transmission frequencies as opposed to analog audio signals ranging from 20Hz to 30,000Hz on the same wire.

What you're splitting hairs over is exacerbated in the following scenario: You have two speakers 10 feet away from the receiver on either side. You buy a pre-made pair of 15 foot Monster cables to connect them to each speaker and thus have an excess of 5 feet of cable in the length of the signal path. Now, calculate the signal delay in nS (nanoseconds) and then contrast the audible difference attributed to the Cat 5 typical twist length variation of 0.6 to 0.85 cm between pairs for similar 15 foot lengths of wire to each speaker. Bonus points for factoring in the the effects of anticipated psychological expectations.

For for the boring formulas, plus the additional effects that various jacket coatings also has on things, I refer you to:
http://www.adc.com/Library/Literature/1317032.pdf

sleddogman,

You gotta admire a man who can kill a perfectly good argument. :whip:

So... How many dogs you got? :pawprint: What type?

sleddogman,

You gotta admire a man who can kill a perfectly good argument. :whip:
Just sharing what I've had to deal with professionally for years, plus trying to look at things logically. Hopefully I didn't piss anyone off or rub them the wrong way, but I probably did. Gomennasi mina... (sorry folks)

So... How many dogs you got? :pawprint: What type?
A black and white female Siberian Husky named Shasta with one blue eye and one brown eye (my David Bowie doggie) that was one-quarter Timber wolf.
Gentlest soul on four legs that talked instead of barked, loved pizza, and loved to pull kids on skis and sleds (via a shoulder harness) although she never weighed more than 40 lbs her entire life. Passed on at age 14 due to a misdiagnosis that eventually led to cancer speading through her lymph system and al brain tumor that inpared her ability to walk.

The hardest thing I've ever done to date was holding her in my arms and saying goodbye in her ear while the vet put her to sleep.
I felt she deserved at least that much...