This isn't an article per se as much as it is a further explanation of some of our DIY cable and interconnect projects. Between our egroup and personal emails, we have been receiving numerous questions regarding CAT 5 wire and the proper construction techniques for our DIY Triple T and the FFRC speaker cables.
Along with our DIY speaker cables, many of our readers and egroup members have been building interconnects out of CAT 5 cable. I've built several sets of these myself and found their sound to be very good.
Hopefully what is presented here will help to answer some of your questions. What follows is the questions and reply to one of our readers.
----- Original Message -----
From: John Little
To: Scott Faller
Sent: Saturday, September 22, 2001 8:39 PM
Subject: Speaker Cable Recipe - Mod Questions
I've read with interest all the CAT 5 DIY speaker cable recipes on the web (including TNT) but they frankly are confusing in their multitude and don't address the system I have. First, twisting the conductors trades mutual inductance for capacitance, requiring as short a length as possible be run. My problem/question is this: I have a vintage Design Acoustics sub/sattelite system in which cable is run from the amp to the sub/mass module (where the crossover network is located) and then to the satellites. This requires running considerable more than the 7' or 8' lengths I see commonly cited in the DIY stuff. How long a cable can I run....what is the impact of the additional runs? Everything I see does not mention 3-piece systems.
Also, which type of CAT 5 is best? I have a spool of Belden 1872A CAT 6 (MediaTwist Ultra High Performance Twisted Pair Multimedia Cable). Is this OK? Another concern is how many twists/runs in each cable.....I've see upwards of 27 pairs of 3 runs for each cable - is there a metric identifying the proper number of pairs/lengths?
thx in advance!
----- Reply -----
OK, you asked a few tough questions. These answers are going to be a bit lengthy but there's no other way to approach it. Ready for an education?
Let's take on the easy one first. The CAT 6 wire that you have is just fine, in fact it's better than the CAT 5 or CAT 5e.
There are a few things you need to know about any multi-strand twisted pairs. The 1872A that you have is comprised of four twisted pairs of wire. The premise behind the lack of an outer shielding is that signals (positive through one and ground through the other) traveling through the twisted pair will reject noise. They don't reject it as a copper foil or braided sheath does but it absorbs the noise then since the noise is introduced into positive and ground at the same time, each travelling opposite directions, they cancel each other out. This is what's referred to as Common Mode Noise Rejection (CMNR). In the case of the 1872A, the signal to noise ratio at 20kHz is unreadable (or greater than 110 dB). That's mighty impressive.
Let's address your concerns regarding trading mutual inductance for capacitance. In the case of the 1872A, the capacitance fairly reasonable, 15pF per foot. I believe that is a weighted average between the tightly twisted pair and the loosely twisted pair. In comparison to some of the "high end, high priced" cables, this is actually lower. Just to give a few examples, Kimber 4 TC is 44 pF/ft, 8 TC is 100 pF/ft, Cardas Quadlink 5C is 23 pF/ft and the Goertz starts at 500 pF/ ft and goes up from there !! That being said, I wouldn't worry about capacitance.
Now lets address the proper number of pairs and lengths. This is (lots) more complicated.
First there is an issue that we must address. If you take a short section and remove the outer sheath you will notice two pairs are tightly twisted and two pairs are loosely twisted. If you use both the tightly twisted pair and the loosely twisted pair in the same cable construction you will have an issue with "time". Obviously the tightly twisted pair uses a longer run of copper to cover the same one foot as the loosely twisted pairs. Is this really an issue? Some think so. Granted, in the communications field, they transfer information in Gigabit networks. Ultra-high speed transfer of information. Does this apply to audio? Doubtful, but it's worth mentioning. Since our cable runs are relatively short, the likelihood of hearing the audible difference is small. Some out there will argue that fact.
Here's a suggestion, to be on the safe side and since you have a spool of 1872A, when you build your cables, peel off the outer sheath and just use the tightly twisted pairs. Slide the bundle of wires into your favorite color heat shrink.
While I am talking about construction techniques, I need to mention a couple of things about unshielded twisted pairs. First and foremost. The positive and ground have to be run through the same twisted pair. If you don't, the CMNR of the twisted pair is negated. This will be a pain in the old sphincter muscle when you go to terminate the wires but it should be worth it. Second, when you build and place your cables, don't kink them. For that matter, don't turn them in tight radius's either. If you do, this will skew the twist of the pair, thus introducing a point where noise can enter the cable and also create the possibility of crosstalk. You ask, What about the ends of the cables where I make my terminations? Good question. Try some foil shielding there. I can't say for sure that is will work, but it sure can't hurt.
Wire Sizing. This is an education all unto itself. First, speaker cables and their damping factor. Yep, that's right, damping factor. In an audiophile system, the cable should be sized not only to reduce the effects of resistance (voltage drop resulting in loss in SPL's) but should be adequately sized to allow the amplifier do it's job in controlling the speakers vis-à-vis damping. If we undersize our cables, we tend to get lots of "overhang" at higher volumes.
Rather than me attempting to explain damping factor, I'll let our resident damping factor expert Dejan explain it in his own words.
Simple definition - damping factor is the difference between the amp's output impedance and the speaker's nominal impedance. Impedance means AC, not DC.
High damping factor means that the amp is able to control the speaker's Q factor properly, and therefore not only the quantity of its motion, but also its quality. While it is assumed a damping factor of 20:1 (26 dB) is sufficient, one must also realize that speakers drop below their nominal impedance as a general rule - therefore, that 20:1 should be applied to the speaker's lowest actual impedance, not its nominal impedance. Assuming drops to 2 Ohms, relatively pessimistic, the 8 Ohm damping factor should be no smaller than 80:1.
So how do we relate damping factor to cable sizing? Rather than doing real Math, I created (actually swiped from the internet and modified to suit our needs) a Microsoft Excel file that will do all the dirty work for us. Also in this spreadsheet is an aggregate wire gauge calculator. If you are building a set of CAT 5 speaker cables, you will need to know how many twisted pairs will be need to make, say, a 12 gauge speaker cable. This spreadsheet will do it for you.
Well, I think I've answered your questions about CAT 5. Hopefully this helps.
Scott Faller (and Dejan for the DF definition)
© Copyright 2001 Scott Faller - http://www.tnt-audio.com