Published 28 Sept 2009.

The myth of loudspeaker bi-wiring

I have seen many claims that bi-wiring of loudspeakers improves sound quality. Since I could not understand the rationale for this, I made a simple impedance test. First, the concept of bi-wiring may be followed by referring to this diagram.

Diagram

The ‘old-fashioned’, traditional method of connecting an amplifier to a speaker with a single cable pair is shown at the top. The bi-wiring technique shown at bottom involves connecting the amplifier separately to the woofer crossover with one pair and to the tweeter crossover with another pair. Some loudspeaker manufacturers cater for this by providing separate pairs of terminals, which if used in the traditional single-pair sense, are linked together at the loudspeaker.

I used the Praxis system to measure the impedance of an Energy Pro 4.5 loudspeaker, via a 5 metre length of monster cable. In the first instance only one pair of the monster cable was used and the loudspeaker terminals were strapped together. Secondly, the loudspeaker terminals were unstrapped and both pairs within the monster cable were used, thus simulating the bi-wired configuration.

See the results in the impedance graph.

Impedance plot

The solid lines are the magnitude (blue) and phase angle (red) for the single pair (crossover network inputs strapped together) case. Dotted lines are the equivalent plots for the bi-wired mode. These loads are what the amplifier drives in both cases and there is not a lot of difference. So little difference as to be insignificant. The shape of these plots and the slight discontinuities can be ignored for this discussion. Clearly the speaker has some mechanical resonances creating the impedance glitches, however these are the same in both configurations. The midrange driver resonant magnitude peak is reduced slightly from 16 ohms to 15.1 ohms in bi-wired mode but above about 5kHz, impedance magnitude increases by a small fraction of an ohm. The use of two pairs of wires does transform the speaker impedance very slightly differently to the use of one pair only but if this were to create any change in the amplifier transfer curve, then it would be a heroically misconceived amplifier indeed.

Some claims I have seen suggest that cable distortion must be less when bi-wired because the woofer and tweeter power is separated. Certainly the tweeter currents and woofer currents exist on separate pairs in bi-wired mode, but they do sum at the amplifier terminals. The full frequency range of drive voltage still appears at both crossover inputs. If one were to suggest that the isolation of these currents means less intermodulation, then you would also be suggesting that wires have non-linearity. However, the temperature coefficient of resistivity of most conductors, including copper, is very linear over a very wide temperature range and there is no practical non-linearity possible in a speaker cable and so mixing of signals of various frequencies in the same wires will not occur.

Finally, there is the question of back-emf. Conventional dynamic moving coil type speaker drivers, being a compromise as they are, may not follow the electrical stimulus exactly. Usually cone movement will lag the stimulus to a varying degree. Certain waveshapes could result in the driver momentarily acting as a motor making the creation of a back-emf (voltage) likely. The level of this back-emf depends on many variables including voice coil inductance, resistance, the moving mass of the driver, the magnetic flux and the rate of change of current in the coil. The back-emf is developed across the amplifier terminals. Output impedance of an amplifier is very low (for most solid-state designs) and damps the back emf considerably, at least over the range of frequencies where the feedback loop is effective. Some amplifier designs incorporate a pair of diodes at the output terminal; one to each supply rail as an extra precaution to ensure that any exceedingly fast emf spikes (should these occur) are bypassed to a low impedance point. So, the back-emf from the woofer is damped to a low value due to the amplifier and does not have a lot of influence on the tweeter. The bi-wire connection does cause a small amount of extra isolation between woofer and tweeter, only to the degree of the sum of electrical resistance of each pair of wires. The wire resistance should, by design, be kept very low and so the very small advantage will be insignificant.

I have to conclude that there is no value in using a bi-wired connection. The same results are achieved by using a single pair of equivalent loop resistance to that of the two pairs of wires in parallel.

Note here that this article is not about using bi-amplification. Bi-amplification has many advantageous outcomes. This has been discussed briefly in my article ‘penultimate amplifier project’.

Axino-tech September 2009