Impedance match

Thanks for the excellent input, George.

An additional point everyone should be aware of is that how objectionable a given low ratio is likely to be depends not only on the ratio itself but also on how much variation that ratio has over the frequency range.

For example, if the ratio is say 3:1 at the worst case frequency (i.e., at the frequency at which the ratio is lowest), but the impedances that are involved don’t vary much over the frequency range, and hence the ratio doesn’t vary much over the frequency range, the consequences would be a slight and inconsequential reduction in gain; an increased sensitivity to cable effects (especially if the low ratio is due mainly to a high output impedance of the component providing the signal); and perhaps a small degradation in the distortion performance of the component providing the signal.

However if the low ratio involves the kind of output impedance characteristic that occurs in the case of many tube-based components, where the output impedance may be a few hundred ohms in most of the spectrum but may rise in the deep bass region to a few thousand ohms at 20 Hz, the consequences of that same 3:1 ratio (at 20 Hz in this case) will include significant deep bass rolloff as well as frequency-dependent phase shifts in the bass region. Which are likely to be much more noticeable and objectionable than the effects described in the preceding paragraph.

Or putting it all another way, 3:1 may be fine in some cases, while 8:1 may be unacceptable in some cases, depending on how the impedances and consequently those ratios vary over the frequency range.

Regards,
-- Al

Tim's response is correct, of course. The following excellent article may provide clues to what might underly the dealer's misconception:

https://www.soundonsound.com/techniques/understanding-impedance

As the article indicates, the concept of numerically matching source and load impedances for transmission of analog audio signals originated in the early days of telephony. And it was necessary in that application, in pre-digital times, as a consequence of the fact that the distances involved (many miles) were a substantial fraction of the wavelengths of the signals.

That concept was also carried forward into the early days of pro audio, in part because of the characteristics of the microphones that were used in those days. But for the most part that concept no longer applies even in pro applications.

In modern home audio systems the only situations I can think of in which source and load impedances should be numerically matched are those in which RF frequencies are involved, such as in the transmission of AES/EBU and electrical S/PDIF digital audio signals. Matching is necessary when RF frequencies are involved in order to minimize or eliminate reflection effects that would degrade waveform quality.

Regarding power amp outputs, unfortunately some manufacturers confusingly refer to the 4 ohm and 8 ohm taps of their amplifiers as having output impedances of 4 or 8 ohms, when what they really mean is that those outputs are designed to perform in an optimal manner when loaded by 4 ohms or 8 ohms. That can be seen in the measurements of many tube amps John Atkinson has provided in Stereophile's reviews. Depending on the particular amp the output impedance of 4 ohm and 8 ohm taps can be anywhere from a small fraction of an ohm to several ohms, or even more in a few designs, and has no particular relation to the load impedance the tap is designed for.

Regards,
-- Al