Is a stereo amp, when bridged to mono, by definition differential?

The short answer is no.

Push-pull means the output stage is has complementary devices that alternately control about half the signal by sourcing and sinking current -- not voltage -- between the positive and negative voltage rails. How much of the signal each control depends on the biasing so both Class A and Class A/B amps are push pull. 

The term differential is usually applied to the input stage, where two transistors share a common emitter load or current. Each transistor is fed a signal (source and feedback, e.g.) and the difference is amplified and then converted to current and sent to the voltage amplification stage. The VA stage then sends the voltage gain to the output stage where the push pull takes place (the OS is a low impedance load that has unity voltage gain and current amplification). The current is then sent from the positive (red) terminal to the speaker with the voltage referenced to ground. The current is in phase with the voltage.

When the two channels of the output stages are bridged, one channel receives the input signal in the normal fashion and the other channel receives the signal inverted. Both channels still operate in their differential input and push-pull output stages as they would, however the output voltage is mirrored (i.e., fully balanced). But the difference is that the inverted channel's red positive output is connected to the negative terminal of the speaker.

The speaker is no longer referenced to ground. It is referenced to the negative voltage of the inverted signal channel. Since each channel is out of phase, the peak to peak voltage is doubled. Therefore the power is quadrupled (power supply willing and able, of course).

The reason it is not push pull is because the current is not sourced between the two channels through the speaker load. The current is simply doubled through the speaker and in phase with the positive (non-inverted) channel.

So, let me throw in a question about that Audiopax amp- the 88, if memory serves, which consisted of two mirror image amp circuits that were combined in a single channel with controls that allowed you to vary their performance in relation to each other. I had these amps for a year or so, and it was fascinating to essentially 'dial in' their performance.

Howard (Hodu), when I wrote my previous posts I had overlooked the reference in your original post to push-pull. As I indicated, my responses pertained to the relations or lack thereof between bridged, differential, and balanced amps. Thanks to Cleeds and GS5556 for addressing push-pull.

Also, Ralph’s (Atmasphere’s) post reminds me that the statement in my previous post referring to "two differential receiver stages" would also be applicable to a design in which the balanced input is routed to a single differential receiver stage which provides both inverted and non-inverted outputs.

Bill (Whart), I took a quick look at reviews of the Audiopax 88 in Stereophile and Soundstage, which I see is a single-ended pentode design. The Stereophile review states that its two amplifier circuits are utilized in series, and I certainly don’t know what to make of that. But the controls you refer to, called "Timbre Lock," apparently simply vary the bias of each of the pentodes, thereby changing the amp’s distortion characteristics. The stated intent being to allow those distortion characteristics to be adjusted by the user so as to be as complementary as possible to the distortion characteristics of the speakers that are being used. Although at the same time John Atkinson’s measurements make clear that for several reasons one’s choice of speakers that would be suitable matches for the amp is particularly limited.

Best regards,
-- Al

Thanks for explaining, Al. They were interesting to use. 

Here's why I posed this question in the first place. I was reading (in 6moons) a review of the McCormack DNA-500 and read the following from designer Steve McCormack:

"In the context of its power rating, how I can achieve this size and weight factor while still delivering this type of power is because unlike previous DNA designs, the DNA-500 is a fully differential/balanced amplifier from input to output," McCormack explains. "Speaker drive is thus differential or push-pull as well. We combine two complete amplifier sections per channel, each of which is driven from an input phase-splitter. (The incoming signal is split into a pair of opposite-polarity waveforms, each of which drives one section of the amplifier pair.) In most conventional amplifiers, the speaker is connected to a single positive output and ground. In the DNA-500, the speaker is connected across a pair of positive outputs of opposite polarity, thus driving the speaker in push-pull mode. This arrangement delivers a degree of speaker control that is not possible otherwise.

"Having experimented extensively with this design approach over the years, it has come to be my favorite method for producing high-performance amplifiers. It creates what is often referred to as a bridged amp. There are four independent amplifier channels, two per side, each channel pair in bridged/balanced configuration to double the voltage over a conventional amplifier. An ordinary 100-watt stereo amplifier in bridged mode will net you a 400-watt amplifier if the power supply can process the current. That's the principle of how the DNA-500 works.

"Many audiophiles feel that such a bridged-style amplifier design is a bad approach. Their prejudice is based on past experiences with stereo amplifiers that were switchable into bridged mono. These amps often had poorly designed input phase splitters and were not optimized to handle the reduced load impedance in bridged mode. The result was often compromised audible performance, instability, overheating, blown fuses and more than a few damaged amplifiers. It's no wonder that bridged amps have acquired such a bad reputation. However, the DNA-500 demonstrates that this need not be the case."

This all got me to thinking, and wondering. As I'm nobody's electrical engineer (or any other type of engineer, for that matter), I may well have been confused by some of the terminology, perhaps conflating this with that and the other thing. 

Thanks to all for your thoughtful and considered responses.

-- Howard