Autoformer vs Speaker impedance Curve

Ralph, would I be correct in thinking that the low B+ voltages that must be used for the 6AS7G power tubes in your amps, compared to the much higher B+ that is supplied to the tube types used in most non-OTL amps, would tend to increase the degree to which output power capability is sensitive to AC supply voltage variations?

Thanks.  Best regards,
-- Al

Back in 2010 I asked Ralph about that statement in this thread.  Here are excerpts from our discussion:

Atmasphere: Power cords: a 2V drop across a power cord can rob a tube amplifier of as much as 40% of its output power! Cripes! You're trying to say you can't hear that?? [That comment was addressed to someone else].  So this is very measurable and audible as well. On lesser transistor amps, a power cord will be less audible as the drop across the cable is reduced, but a class A transistor amp will easily bring out cable weaknesses.

Almarg: Ralph, could you provide a technical explanation of why that would be so? I don't doubt your statement, but I'm interested in understanding why that would occur.... Re your other points, all of which strike me as excellent, I think that it should be stated that none of those points NECESSARILY mean, to cite an example, that a $2,000 power cord will outperform a $200 power cord in any given system.

Atmasphere: Al, the reason a power cord can have this effect is simple. If there is a 2 volt drop in a power cord, the filaments of the tubes will run cooler and the B+ will be reduced. Since this is a voltage, the result is we get less voltage output out of the amp. Less voltage=less power. Depending on the amp this can be pretty profound. and I have seen it with my own eyes. I do agree though that that does not justify a $2000 power cord, but it **does** justify one that has decent connectors and conductors that will not heat up at all. That has to cost something, probably not $2000 though. One thing about audio is that if there is a phenomena, there is also snake oil for it.

Almarg: Thanks, Ralph. That would also seem to say that the value of the ac line voltage at each listener's location can be a very significant variable in the performance of a tube amp (assuming it does not have regulated filament and B+ supplies). Which in turn emphasizes how easy it can sometimes be for extraneous variables to lead to incorrect sonic assessments.

Atmasphere: No doubt! But it extends to anything that can draw significant power- and bigger transistor amps can! Imagine the peaks just... not... making it.

This taught us a lesson... when we set up an amplifier for test, we test the AC line voltage from the IEC connector. The meter on the variac (a bit of test instrumentation) cannot be trusted. 

A 2 volt drop in a power cord of typical length seems like a lot, of course.  But if the amp draws say 5 amperes a resistance of 0.4 ohms, including the resistance of the connectors and the contacts as well as the combined resistance of the hot and neutral conductors, would do it.  I'd imagine that could happen in some situations.

Regards,
-- Al  

For both transformers and autoformers, softer on dips and louder on peaks, ***relative to the response that the particular speaker would provide if the amp were an ideal voltage source, meaning a voltage source having an output impedance of zero.*** However in the case of an autoformer (which would be used in a solid state design) the differences will be extremely small, and essentially negligible with most speakers. Not because of the difference between an autoformer and a transformer, but because of the differences in output impedance between tube amps and nearly all solid state amps. As I said in my post dated 8-16-2018:

A major factor contributing to this, and what is probably the most major factor in many cases, is not the output transformer itself, but the interaction of the output impedance of a tube amp with the speaker impedance variations that you are referring to. In contrast to nearly all solid state amps, most tube amps have output impedances that are a significant fraction of speaker impedance, usually somewhere between a large fraction of an ohm and several ohms. That in turn causes the voltage divider effect to have significant effects on tonality, to the extent that the speaker’s impedance varies as a function of frequency.

In the case of McIntosh solid state amps which use autoformers that particular effect is essentially negligible with most speakers, because as a consequence of being solid state their output impedance is much smaller than the output impedance of most tube amps. (Although that certainly does not mean that an amplifier having low output impedance is necessarily the best match for a given speaker, in terms of tonality). For example the MC302 has a specified damping factor of "greater than 40," which for the 8 ohm tap theoretically corresponds to an output impedance of less than 8/40 = 0.2 ohms.

Regards,
-- Al

+1 Eric (bdp24).

We are privileged to have highly experienced and accomplished designers of outstanding audio electronics, such as Ralph and Roger, sharing their knowledge with us here.

Best regards,
-- Al 

Does this mean Macs are voltage paradigm amps?

Yes. Which paradigm an amp fits into is determined by its output impedance. (And btw, that categorization is along a continuum, rather than being a purely black and white distinction. Especially in the case of tube amps, which vary widely in their output impedances).

If the output impedance of an amp is a tiny fraction of an ohm, or is at least a very small fraction of the impedance of the speaker at any frequency, it will behave as a voltage source. Which means that for a given input signal to the amp, it will output a voltage which essentially has no variation as a function of the impedance of the speaker at whatever frequencies may be present, as long as the amp is operated within the limits of its maximum voltage, current, power, and thermal capabilities.

And in the case of McIntosh solid state amps having autoformers, the combination of their solid state output stages, the autoformers, and what I’m pretty certain is the liberal application of feedback is most of their designs, results in a very low output impedance. The MC302 I referred to earlier being an example.

Tube amps, on the other hand, will just about invariably have relatively high output impedances, usually somewhere between a large fraction of an ohm and several ohms, as I mentioned earlier. That will bring just about all tube amps much closer to the power paradigm end of the spectrum.

Regarding your questions 2 and 3, it follows from Ohm’s Law (I = E/R) and the definition of power (P = E x I, for a resistive load), where I is current, E is voltage, and R is resistance, that if a constant voltage is maintained into a varying load (as it would be by a voltage paradigm amp) more current and hence more power will be delivered into low impedances than into high impedances (assuming at least that all of the impedances are mostly resistive). It also follows that a power paradigm amp will come much closer than a voltage paradigm amp to maintaining constant power into those varying impedances, for a given input voltage to the amp, rather than maintaining constant voltage.

It may help to clarify some of this, btw, if you take a look at the Wikipedia writeup on voltage dividers that I referred to earlier. In the first figure on that page, consider Z1 to represent the output impedance of the amp, and Z2 to represent the impedance of the speaker. And consider Vin to be the voltage the amp is "trying" to put out, meaning the voltage it would supply without a speaker or other load being connected, and Vout to be the voltage seen by the speaker.

None of this necessarily means, however, that frequency response flatness (which you appeared to be referring to when you mentioned "linearity") will be compromised if additional power is or is not supplied into impedance dips. Depending on the design of the speaker its efficiency (SPL out vs. watts in) may or may not vary in a manner that is consistent with its impedance curve. As Ralph (Atmasphere) has said in a number of past threads, a tonal imbalance is especially likely to result when the paradigms to which the speaker and the amp conform are not the same.

Regards,
-- Al

With typical output transformers (common with tube amps) speakers with wild impedance swings (and phase) present a challenging load to the amp and fidelity can suffer.

A major factor contributing to this, and what is probably the most major factor in many cases, is not the output transformer itself, but the interaction of the output impedance of a tube amp with the speaker impedance variations that you are referring to. In contrast to nearly all solid state amps, most tube amps have output impedances that are a significant fraction of speaker impedance, usually somewhere between a large fraction of an ohm and several ohms. That in turn causes the voltage divider effect to have significant effects on tonality, to the extent that the speaker’s impedance varies as a function of frequency.

In the case of McIntosh solid state amps which use autoformers that particular effect is essentially negligible with most speakers, because as a consequence of being solid state their output impedance is much smaller than the output impedance of most tube amps. (Although that certainly does not mean that an amplifier having low output impedance is necessarily the best match for a given speaker, in terms of tonality). For example the MC302 has a specified damping factor of "greater than 40," which for the 8 ohm tap theoretically corresponds to an output impedance of less than 8/40 = 0.2 ohms.

Can the same be said of autoformers? Or, are these electrically different enough from transformers as to not present the same issues?

While as I’ve said the effects you appear to be asking about are usually not due to output transformers themselves, autoformers do have significant advantages relative to output transformers in typical applications. One is that autoformers don’t have to be designed to handle significant amounts of DC, while the output transformers in tube amps must be able to do that. Another is that in typical applications an autoformer just has to provide a small transformation ratio between input and output voltage, current, and impedance, while the output transformer of a tube amp usually has to provide a much larger ratio. Those differences mean that a well designed autoformer will tend to have fewer sonic side-effects than a comparably well designed output transformer.

Regards,
-- Al