Loudspeaker Cable Resistance

When putting together a good quality audio system a question that often arises is, “What loudspeaker cable should I choose?” Answering this tends to become quite complex if you start reading ‘reviews’ in audio magazines that tend to make almost a holy mystery of the claimed effects of various expensive cables.

In practice, though, the main basic engineering requirement tend to be that the cable should have a ‘low enough’ series resistance. This isn’t because there is normally a worry that the cable may ‘fail’ or burst into flames. It is that a large cable resistance may interact with the loudspeaker to reduce the signal level or change the frequency response.

Many years ago the standard advice people followed was that given by experienced people like Peter Walker of QUAD. This is to simply ensure the loudspeaker cables have a series resistance which is less than a tenth of the impedance of the speakers you are using. However although this is good advice, these days I would tend to recommend going further than this and ensuring that the cable resistance is smaller than QUAD used to recommend. My own personal preference is to recommend that the cable resistance should be comfortably below 0·1 Ohms. So to ensure this I tend to aim at 0·05 Ohms or less. This should ensure that any tendency for the cable resistance to alter the response will be so tiny as to be inaudible.

Now the total resistance of the cables will depend on three factors:

The resistivity of the material used in the conductors.
The cross-sectional area of the conductors
The length of the cables

In general, people tend to use Copper for ordinary wiring, so I’ll assume that the chosen wires are made of Copper. (Some cables use ‘OFC’ copper or silver, but in practice these have a resistively which is more similar to ‘ordinary’ copper than the purveyors of expensive cables tend to imply!)

I will also ignore questions like, “Do I like the look of the cable?” as that will be a matter of personal taste once you have taken resistance into account! I’ll also ignore the possibility that the changes produced by a high series resistance might alter the results in ways you prefer.

To help illustrate what cables might or might not be suitable we can use as an example a set of types of loudspeaker cable sold by Maplin. The ones I have chosen are listed below

Label on graphs	Name	Part No	Strands per bundle	Strand Dia (mm)
A	Bell Wire	XS92A	1	0·50
B	Min. LS Cable	XR39N	7	0·20
C	LS Cable	XS47B	13	0·20
D	Heavy Duty LS cable	XR60Q	40	0·20
E	Hi Fi LS Cable	XR72P	79	0·20
F	High Quality LS Cable S12	XS37S	415	0·10
G	High Quality LS Cable S10	XS37P	1050	0·08

We can then ask the questions:

What is the length of cable (of each type) that would give a series resistance of 0·3 Ohms? (i.e. apply QUAD’s old “one tenth” rule and assume the speaker has an unusually low impedance that drops to 3 Ohms at some audio frequencies.
What is the length for a cable resistance of 0·05 Ohms? (i.e. a value that is very small.)

These lengths will then represent the maximum lengths which, for either ‘guiding rule’, a given cable would be acceptable in practice.

How much effect the cable resistance has will depend upon the loudspeaker’s impedance properties. The specifications for many loudspeakers quote nominal impedances of ‘8 Ohms’. However in practice if the speaker is measured we tend to find that the actual impedance varies with frequency and may dip well below the nominal value at some frequencies. (And may be much higher than the nominal value at other frequencies.) A detailed example of the effects this can have are considered on another page that considers the QUAD ESL57 and 303. In general, the lower the impedance at the ‘minima’, the greater any effect of cable (or amplifier output) resistance may be.

Here I will assume that the speaker may dip down as low as 0·3 Ohms. In practice, most speakers won’t have a value this low at any audio frequency, so this gives us a ‘probable worst case’.

A series resistance of a tenth of the load might alter the level or response by up to approximately 1 dB. A series resistance of 0·05 Ohms might produce changes of between 0.1 and 0.2 dB. It may be possible to hear changes at the 1 dB level, but it is doubtful if many people can detect a change of less than 0·2 dB. Even if they can, it seems likely that such a small change would not be significant given that moving your head or the speaker locations a few cm would probably have a far larger effect!

The graph above illustrates the results for the cables listed. The green line represents the weaker ‘QUAD’ situation, and the blue line the more demanding ‘0·05 Ohm’ situation. Note that the lengths of cable we can use increases linearly with the cross-sectional area. So the general rule is that the longer you need the loudspeaker cables to be, the greater the cross-sectional area of the copper you should choose.

To see how to make use of the above, lets consider an example where we want to use loudspeaker cables that are 3 metres long. At this length it will be easier to read a rescaled version of the above graph, so this is shown below. Here the green broken line is for a 3 metre length.

Looking at the blue line on the graph we can see that cables E-G would provide a resistance of less than 0·05 Ohms for a 3 metre length, so would pass the more ‘demanding’ requirement. The line also shows that a cross-sectional area of copper of at least 2 square mm is required.

Similarly, from noting where the solid green line crosses the 3 metre length we can say that C and D would also pass the weaker, QUAD, requirement. For this less demanding requirement we would only want a cross-sectional area of copper of around 0·5 square mm or more.

Hence I’d say that if you want cables that are 3 metres long, choose one of the cables E-G if you can, but C or D might be fine. Avoid cables A and B as they may have too high a series resistance and audible alter the sounds.

Having seen the above graphs, and how they are used, you can now use them for other cable lengths, and you can add other possible types of cable by finding out the cable’s cross-sectional area of conductor and making a suitable point on the lines shown on the graphs. More simply, for loudspeaker cables up to 3 metres in length, go for cables that have at least half a square mm of copper, and more like 2 square mm if possible.

Of course, it may be that you might prefer the sound if you deliberately used a long run of ‘thin’ (i.e. high resistance) cables as a result of the changes in frequency response. However if this is the case, bear in mind that you could probably get much the same effect by fitting series resistors. These would be cheaper than long runs of cable, and easier to add/remove.

In some cases the series inductance of the loudspeaker cables may also matter. This is less likely to have an audible effect in most cases, so probably isn’t worth worrying about unless either:

The cables have been built to deliberately have a high series inductance. You can usually recognise such cables as the conductors tend to be spaced much further apart than is normal for twin-flex types.
Your loudspeaker leads are exceptionally long. (e.g. much more than two or three metres)
Your loudspeakers have an unusually low impedance at high treble frequencies. (e.g. behave as a capacitative load at HF, like the QUAD ESL57)

In each case you would probably be aware of the special circumstances, and might wish to take them into account. Unfortunately, loudspeaker cable reviews and advertisements in the UK rarely give a value for the cable inductance. On this webpage I have therefore concentrated on resistance as this is more likely to have an effect, and easier to assess the cables. If you wish to know more about this topic you may find it useful to visit the Scots Guide to Electronics and read some of the webpages there on cables in the Analog and Audio section.

29th Dec 2005
Jim Lesurf

(A version of this article has been published in Living with Technology magazine)