Reference

Thomas J. Norton's "Questions of Impedance Interaction" in January (p.109) showed that a ghostly echo of a loudspeaker's impedance modulus can be imposed on its frequency response by virtue of an amplifier's source impedance acting as the top limb of a potential divider. Mr. Norton illustrated the effect with some specific graphics, but the data can be usefully generalized by means of a simple rule arising from the ohmic arithmetic.

Assuming a worst-case situation of very large impedance undulations, with, for instance, an LF peak reaching 10 times the value of that characterizing the lower-mid region, the rule runs as follows: To confine frequency-response changes within an amplitude band of 1dB, the amplifier's source impedance (plus the resistance of its connecting cable) must be less than an eighth of the speaker's impedance at the latter's lowest point, and less than a sixteenth for a band of 0.5dB.

These criteria also satisfy damping requirements for practical purposes, since the resistance of a speaker's voice-coil is effectively in series with the amplifier feed, so that once the latter falls below about one quarter of the coil resistance, there can be no worthwhile improvement. Perceived changes in bass "softness" between transistor and tube amplifiers are thus more likely to be due to LF impedance "response ghosts" than to the damping-factor, as such. Huge ratios for this parameter make impressive reading in technical specifications (like ultra-low distortion figures), but are so far along the relevant asymptotic curve as to be useless for audio purposes.—John Crabbe, Todmorden, Lancashire, England