Twisted?

[FastEddy]

The question has come up in several posts here and elsewhere: Should I twist the wires?

Generally, where possible, the twisting of data, signal or power wires does accomplish some level external noise rejection. Usually the rule of thumb for audio applications is one full twist (360 degrees) per foot for metal conductor, small signal audio interconnecting wires. The same should also apply to any two or three power wires of longer length, say longer than one foot. (This does not apply to optical cables.)

The reason this does anything at all is quite simple. Outside interference strikes wiring in a plane (a wave front), usually laterally to the wires ... so each or the two or three or more conductors "sees" the incoming noise, equally. By twisting these wires, each wire is, alternately, slightly closer then slightly further from the noise source, and "out of phase" with the noise source due to their rotation about the central axis of the wire bundle (this is Michael Faraday stuff = complicated equations) ... this acts to improve what is called "common mode rejection", an actual electronic specification that is measured in db, relative to each conductor. At audio frequencies (less than 100 K Htz.) this does not really cause difficulties except in very small signals or when small signals are amplified by noisy power supplies ... when a combination of power and some unrejected noise passes through or around or near an input to a small signal amplifier ... Like a pre-amp.

Suffice to say it is good practice for similar wires of a length of more than a foot or so are run together, in parallel, then these should be twisted together to improve the common mode rejection ratio, reducing modestly, any outside noise.

[Leolabs]

[AndrewT]

Hi,
Cat5 are 4pair twisted.
The clever bit is each pair have a different rate of twist. I have never stripped one yet to measure the rate.

If you want some modicom of interference cancellation then go for twisting.
Some balanced cables rely on the twist only and no screen!

[anatech]

Hi Andrew,

Some balanced cables rely on the twist only and no screen!

True, but the good ones are screened properly.

I even specify screened cable for the 16 and 14 GA 70V lines I run for paging.

I should point out that running balanced is not always the best way to do things. This is against popular wisdom, but it's true.

-Chris

[FastEddy]

Yes?

[clem_o]

Originally posted by AndrewT
Hi,
Cat5 are 4pair twisted.
The clever bit is each pair have a different rate of twist. I have never stripped one yet to measure the rate.

CAT5 cable uses 3 to 4 twists per inch...

http://www.phptr.com/articles/article.asp?p=23104&rl=1


Cheers!

Clem

[FastEddy]

Yes, CAT5 is four twisted pairs. The "better" CAT5e (and some other varients) have pair twists staggered or with different twist ratios, relative to the other pairs to reduce "cross talk", improving CMRR. CAT5 (and varients) can have all manner of different insulation, PVC (poly), Teflon and several others and often the manufacturers of CAT5 lie about what they are using, designating buzzwords like "compatible" or "complient" as the IEEE does not have very many consumer advocates or policemen. (Almost all CAT5 of every varient is #28 AWG, copper, whether used for power or data.)

USB also has a single twisted pair of data wires and a twisted pair of power wires with the shield (screen?) being connected to power ground (black) at each end. Most USB cables are made of #28 AWG for data and #20 to #24 AWG for power. Almost all USB cabling is PVC covered.

FireWire 1394a has two twisted pairs of data wires with a power wire being twisted into the bundle and the shield (screen?) acting as power ground. All FireWire 1394a and 1394b cables use #28 AWG for data and usually #24 AWG for the power wires, all copper. The 1394a cables of 10 meters in length have #22 AWG for power. Almost all FireWire 1394a and 1394b is plastic covered, usually PVC.

FireWire 1394b has four twisted pairs (ala CAT5) with a #24 AWG power wire being twisted into the bundle and the shield (screen?) acting as power ground, even the 10 meter lengths.

[FastEddy]

anatech: " ... I should point out that running balanced is not always the best way to do things. This is against popular wisdom, but it's true. ..."

Too true. It turns out that, at least for long data cable runs, unshielded cables can transfer signals over about twice the distance as shielded cables. (The use of terms like Balanced as applied to cables usually implies shielding.) Anatech is correct however that unbalanced can run further, especially on those "industrial" 70.7 volt audio lines between transformer, taps, etc.

The unshielded cables have better impedence matching than shiielded, at least in the lower frequencies like commercial & industrial audio (elevator music, as they say).

Of interest: ... those long 70 V. audio lines are suceptable to noise, static, etc. from the outside world ... and the twists become very important in these kinds of installations = at least one full, 360 degree twist per foot is an almost absolute rule, not just a good idea.

[AndrewT]

Hi,
I see your source is Stallings,

CAT5 cable uses 3 to 4 twists per inch

I have gone and measured my Cat5,
blue - 11mm pitch.
green - 13mm
orange - 14mm
brown - 16mm

It appears mine is about half of Stallings figure.
62 to 91 twists/m or 19 to 28 twists/ft.

[FastEddy]

Politically incorrect diatribe:

Too bad the Scotts and Irish engineers don't appreciate the British measurements standards anymore. Many (old) US engineers and technicians still prefer the inch (in decimal tenths and hundredths) over the French system ... and of course those of us who understand farm impliments believe in the yard (cubic and otherwise), bushel and peck, prefering to ignore the unwieldly and often costly cubic centimeter and its long winded bretheren.

Personally I'll take the nautical mile as well, being into sailboats in a modest way.

All that palaver about the length of ruby red light waves being an "absolute" doesn't hold much water, either, considering that the monochromatic red laser light can just as easily be measured in inches .... and that comes to a more easily handled number of inches as well.
["The metre is the length of the path travelled by light in vacuum during a time interval of 1/299 792 458 of a second." ... which is 1/7.5 millionths of a second to the inch.]
Of amusing interest: http://www.bwmaonline.com/Metric%20Culprits.htm ... http://www.bwmaonline.com/indexALT.html ... http://www.angelfire.com/ok4/iaml/ ... http://www.metricsucks.com/

Anyway, here in the colonies, audio equipment installers still rely on feet and inches, circuit board designers must constantly deal with dual units ( 1/10" between DIP chips and 1.0 to 2.5 mm between SM chips. ) and interfacing with sheet metal chassis makers with fractions of an inch.

But it seems that the French "system" is on the acendence once more, prevailing at least in sports medicine ... condeming an athlete for having more than a certain number of CC's of testostrone in his blood, because their own French team has ... none.



" That Anglo-French masterpiece, Concorde, was built with imperial nuts and bolts in an American UNF thread form, whilst incorporating metric rivets! "