Reference

The Cables' Electrical Properties were published in May 1996 (Vol.19 No.5):

Careful readers will have noticed that my article last December about speaker cable measurements was originally written to inform my recording and live-sound engineer colleagues, quite a few of whom recognize sonic differences between cables but are as confused about the subject as everyone else. When, in a recent issue of Stereophile, an advertisement for cable claims, almost in the same breath, that the wire both is neutral and gives warmth, I can rest my case about the dearth of clear thinking and explanation.

I did not cite individual R (resistance), C (capacitance), and L (inductance), figures for the speaker Cables Under Test (CUTs), as making these measurements takes time and increases cost; in addition, as one-dimensional, first-order details, they do not get to the point as graphically as the pulse-shape method that I demonstrated. Nevertheless, the figures are of interest, and I support the collation of such data. How best to measure is another matter.

The resistance—strictly DCR, the DC resistance—of any speaker cable should be well below 1 ohm, and can range down to the low milli-ohms. Accurate measurement demands good instruments and careful technique. For example, copper's resistance varies such that tests can't be casually continued "later"; a 7°C change in lab temperature overnight can change a given cable's resistance by as much as 6%. To take the following readings, a Datron 1061 digital voltmeter/digital multimeter was used. This is a classic, auto-calibrating reference DVM/DMM, made in the UK, which even true-blue US instrument companies have used to calibrate their own designs! Unlike most cheaper DMMs, which are increasingly made of plastic and are also increasingly thoughtlessly designed, the Datron pushes out a decent 10mA on the low-ohms ranges. (Many $1000+ DVMs cannot even test an LED because the manufacturer's idea of an ohmic test current is so pathetic.)

The DMM was switched on for 12 hours before tests began, and the cables were placed in the same room. Testing began when the room's temperature drift was below 1°C/hour, an hour or three after daytime heating had kicked in. To make the connections transparent, 4-wire sensing was used, with locally shielded ("guarded") leads. Just one conductor in each CUT was measured. The complete resistance is easily calculated by doubling, though measuring the other side might reveal slight ohmic differences caused by deviations in cross-sectional area (CSA) or purity. Connection was via the XLR connectors used for the original test left in situ; the 4-wire sensing leads were terminated at mating XLRs.

The meter was first zeroed by plugging these XLRs together and pressing the zero button. This was done several times, and the connections remade, then left, to check for drift and repeatability. In these tests, repeatability was better than ±150 micro-ohms (0.00015 ohms), giving a minimum test accuracy of 0.4% with the lowest-resistance cables, and up to 0.03% with the most resistive. At the start of the tests, the room temperature was 17°C—not bad with arctic winds gusting outside—and at the end, it was 18°C. These temperatures are &#177 .5°C, so the temperature may have varied by 2°C at most, indicating a worst-case additional total temperature-induced uncertainty of about 1.5%.(footnote 1).

The cables in Table 1 are identified by the names used in the original article. Each test length of cable was 18', 5.5m, long.

Table 1: DCR, "half-section" resistance

Twisted, 0.5mm CSA, solid 0.371 ohms
Zip cord 0.136 ohms
Sonic Link AC 0.108 ohms
Twisted, 1mm CSA, stranded 0.102 ohms
Monster HF section 0.098 ohms
Sonic Link Blue 0.066 ohms
Connectronics, 2.5mm CSA 0.059 ohms
Supra Ply, 2.5mm CSA 0.050 ohms
PVC AC, 4mm CSA 0.025 ohms
Test Y-lead, 4mm CSA 0.0026 ohms

In effect, this list measures CSA, factored by purity.

Like resistance, inductance may be calculated by measuring one conductor end-to-end. But the figure attained will not be representative. The real inductance seen by a passing current will be much lower, because it is partially or even largely canceled by the return current's "counter-inductance." To measure this effective, net inductance, one end of the cable was shorted with an XLR shorting plug. All cables were strung out and hung free of magnetic fields and ferro-active surfaces.

As with resistance, the net inductance of 5m lengths of speaker cables is toward the lower limits of most common test instruments, below 10µH. There is also some risk of error, as few instruments have guarding schemes, or 4-wire connections. Cables also have complex loss factors, and this led to false results when I attempted tests using my "trusty" Marconi TF2700 bridge. I turned to the decade-younger and lesser-looking, knob-free Thandar TC200 RLC meter. This proved trustworthy, and the following readings were made within a certainty of ±0.15µH, at 18°C:

Table 2: Loop Inductance

Twisted, 0.5mm CSA, solid 5.7µH
Twisted, 1mm CSA, stranded 4.6µH
Monster HF section 4.5µH
Zip cord 4.0µH
Sonic Link Blue 4.0µH
Connectronics, 2.5mm CSA 3.5µH
PVC AC, 4mm CSA, 3.2µH
Supra Ply, 2.5mm CSA 2.1µH

The test frequency was 1058Hz ±1Hz, meaning that Table 2 is a snapshot of the cables' upper-midrange inductances. Due to the complex action of skin effect, oxidation of any copper, and other factors mentioned in the original article, results would be different at other frequencies and the relative differences would likely no longer apply.

Capacitance was easier to test. The main caveat is that the capacitance of a cable that consists of loosely twisted wires will be rather uncertain. There is also the fact that with some dielectrics, hence constructions, changes both in temperature and humidity can cause unexpectedly large changes in the material's permittivity, hence capacitance. Again, a Thandar TC200 digital RLC meter was used, calibrated against known, aged parts of similar value. The following values were measured at 18°C, with an accuracy, resolution, and repeatability of ±5pF, or better than ±2% in most cases. (As with inductance, this is a mid-frequency snapshot, taken at 1052Hz ±1Hz.)

Table 3: Inter-conductor capacitance

Supra Ply, 2.5mm CSA 1460pF
PVC AC, 4mm CSA 626pF
Monster HF section 605pF
Zip cord 577pF
Connectronics, 2.5mm CSA 551pF
Sonic Link Blue 380pF
Sonic Link AC 290pF
Twisted, 1mm CSA, stranded 251pF
Twisted, 0.5mm CSA, solid 223pF

These results should set some standards for future wire tests.—Ben Duncan

---

Footnote 1: I have not anywhere seen the temperature coefficient (tempco) of copper mentioned in the context of speaker cables. In passing pulses of up to tens of amperes, even chunky conductors are bound to heat up sufficiently to affect the DCR.