Can you get lower output impedance from an OTL design?
Yes.
The question isn't if you
can, but if you
should. It requires lots of feedback in any tube amp to achieve a low output impedance. When you have those levels of feedback you also are pushing the envelope of phase margin in the amp. Reliability has to be considered; any amp that tests its phase margin isn't reliable!
When we came on the scene, the world was more convinced than analog vs digital or solid state vs tubes that OTLs were simply unreliable.
We made the first reliable OTLs; one reason was no feedback. Instead we relied on triodes, class A operation, fully differential balanced operation, a single stage of gain and no output transformer to create a low distortion circuit. It keeps up with tube amps using feedback in that regard, but obviously needs higher impedance loads (or speaker loads that have a flat impedance curve).
But Class D has sometime difficulties high midrange and high frequencies.
But Class D is subject to hiss (in the high frequencies) that is never quantified in the measurements.
The hiss comes in when the amp is unable to compensate for frequency drift of the triangle wave. If it shifts frequency, a quantization noise which sounds a lot like white noise is created. To get around that you have two choices: use a temperature controlled crystal oscillator and divide down to the desired switching frequency,
Or
Use a self-oscillating design. In our circuit, the opamps used in the input buffer creates about 90% of the noise of the circuit. So its quite suitable for use on horn speakers.
More feedback and lower open loop source impedance, the latter of which is expensive, the former of which is tricky. Not to speak for Ralph, but I don’t think low source Z is his goal. In the past, he’s argued for a constant power source rather than a constant voltage source. I disagree with him, but I understand his reasoning.
Being pragmatic is helpful in life preventing expensive mistakes! We found in blind tests early on that listeners preferred our amps on easy speaker loads and not on difficult loads and no amount of feedback would fix the latter. So we stuck to easier to drive speakers. Then along came SETs and the speakers people made for them (or maybe people made SETs for higher efficiency speakers, not sure which...) . Turned out no matter how expensive the SET, our OTLs could sound better in blind tests on the same speakers. With a lot more power, way wider bandwidth, several orders of magnitude lower distortion and so on. SETs have a high output impedance and the speakers on which they are used expect the amp to be a power source rather than a voltage source.
"Class AB has the added advantage of wide HF bandwidth and of not requiring an output filter"
Can this be one explanation of why some do not like class D in the HF bandwith? That the required low pass filter with limited bandwidth cause audible ringing and phase errors in the audible range?
I know that this is not the case with modern class D amplifiers, like Hypex, which I agree sound fantastic.
The Hypex is self oscillating, which means its using a lot of feedback. Its enough that its able to compensate for phase issues that the ear uses for recreating the sound stage.
Traditionally it was not possible to use that much feedback, so to avoid phase shift wide bandwidth was needed. IOW with a self oscillating class D amp you don't need a wide bandwidth and yet the 3D sound field is properly created.
I'm not sure why class D has to have, as often stated, "large amount of feedback" to be stable. If anything, amplifier can be unstable because of the feedback. Output stage is "unstable" to start with, since output Mosfets switch from open to close. It cannot be more "unstable" than that.
In a self-oscillating class D amplifier, so much feedback is applied that the phase margin of the amp is grossly exceeded and it goes into oscillation as soon as its powered up. The feedback network is designed so that it will only oscillate at one frequency: the switching frequency. In this way you kill two birds with one stone- how to get lots of feedback and setting the switching frequency. This in addition to the advantage I mentioned just above.
