Of course not. 
I know where your question is heading. Everyone should agree that testing is supremely sensitive to the areas being tested, and completely blind to areas not being tested.
An obvious example is mechanical hum. An amplifier can ace all electrical tests and still sound like a beehive.
Ed
I know where your question is heading. Everyone should agree that testing is supremely sensitive to the areas being tested, and completely blind to areas not being tested.
An obvious example is mechanical hum. An amplifier can ace all electrical tests and still sound like a beehive.
Ed
I struggle to place meaning to that group of sentences. I don't think the question needs anything more than a yes/no answer really...
That's a very narrow definition, Mark.
If your definition of an "amplifier" is a PCB ... then I guess it stands. In which case, you would presumably diagnose the situation where hum is generated because the input signal wiring is placed too close to the mains wiring ... as due to "the wiring" - rather than "the amplifier"?
(My own definition of an "amplifier" ... is: amp PCB + PS + wiring.)
Narrow definition means correct definition in this case. An amplifier is a circuit that amplifies a signal taking DC power in (well almost invariably DC power). With batteries you might not need a separate PSU and you certainly won't see mechanical (ie transformer) hum! As I said why conflate two different things which are clearly different and serve different purposes? After all we are talking about "circuit strategies" (i.e. amp topologies), not integrated amplifiers with built-in PSU's in this thread.
If the thread were about product strategies, say, then PSU's would be part of the scope.
If the thread were about product strategies, say, then PSU's would be part of the scope.
The idea that amplifiers and power supplies are separate, modular devices that don't interact in complex ways is IME not always a sufficiently accurate model. IOW, that the two functions are conceptually independent of each other except for a module interface specification may amount to an over-simplified system model.
Amplifier is a complex term that is usually sufficiently precise. And, as is the case here, this term is also differentiated - depending on requirements;-)
Another recourse: a follower. If you finally want to have and hear "high-end", you should take a look at these simple things. One example is Pavel's:
https://sound-au.com/project83.htm
Another recourse: a follower. If you finally want to have and hear "high-end", you should take a look at these simple things. One example is Pavel's:
https://sound-au.com/project83.htm
Some opamps have higher bandwidth and slew rates above 2000 V/us
So I would say they suck or depends what opamp you use.
Since a amplifier will just basically be a opamp anyways....who cares
Sky is falling. I agree with Markw4 on something.
My approach in design is that amplifier is an extension of its power supply and that there is no excellent amplifier with bad power supply. PSRR alone can’t help there.
However, as thread is about amplifier topology, PS discussion can be omitted, except if amplifier design relies on specific PS.
Strategy is not the wrong term to use in amplifier design. Once you have a rough idea how an amplifier works, you have to start planning for constraints, like how hot can my heatsinks get in a realistic use scenario? How much heat can I generate at idle? What source impedances must it tolerate? Any real use case benefits from planning to have a good level of gain, noise floor, power etc. There are differing ways to meet the same requirements and you can take it as far as optimizing the circuit topology down to the resistor for a given use case if you so desire. In this case strategy can take an amp that is sufficient and make it into an amp that is exactly what was needed. That is no small thing!
As an example, multiply your PSU impedance by your PSRR, and that curve will be the difference the PSU is making to your gain curve with a 1 ohm load. Not to mention any PSU loop areas radiating into the input and feedback loops. It makes a much larger difference on amps with low PSRR of course.
As an example, multiply your PSU impedance by your PSRR, and that curve will be the difference the PSU is making to your gain curve with a 1 ohm load. Not to mention any PSU loop areas radiating into the input and feedback loops. It makes a much larger difference on amps with low PSRR of course.
Last edited:
IMHO, PSRR is more of a figure of merit metric than a hard limit type of thing. It can vary with frequency, maybe a little differently if tested with a transient waveform, and it still just a ratio, not a hard floor. Of course it can be used for approximation purposes through.
Also, in some cases an amplifier can have a negative impedance characteristic at a power supply node in part of its operating range. A small example is AD797, which can oscillate with too little bypass ESR and big caps. Similar for the error amplifier in some voltage regulators, where output cap ESR can affect stability, maybe depending in part on pass element current. Just not convinced PSRR is necessarily an all inclusive metric. Also regulated power supplies are usually amplifiers too in that they have their own error amp, sometimes with a fair amount of bandwidth and gain. They also have an external power source which is usually the AC power line. So they need some PSRR from that too, don't they?
Also, in some cases an amplifier can have a negative impedance characteristic at a power supply node in part of its operating range. A small example is AD797, which can oscillate with too little bypass ESR and big caps. Similar for the error amplifier in some voltage regulators, where output cap ESR can affect stability, maybe depending in part on pass element current. Just not convinced PSRR is necessarily an all inclusive metric. Also regulated power supplies are usually amplifiers too in that they have their own error amp, sometimes with a fair amount of bandwidth and gain. They also have an external power source which is usually the AC power line. So they need some PSRR from that too, don't they?
Last edited:
I want good sound. On this topic I am more aligned with @tombo56 and others. Power supply design is an integral part of amplifier design, because at least at some points, and or maybe all the time, you will hear the power supply too, not just the amplifier board.
In a related vein, the idea that PSRR, FR, noise, and HD/THD, output Z, is all there is, is not complete. Its only an approximation. They are things that are commonly measured because they are very practical to measure, not because they tell the whole story. Yet many people come to believe that the models are the reality. It falls on the more experienced and with greater breadth of experience people to keep things honest on not be satisfied with the lowest common denominator models as fully representing physical reality. This relates to the human part you speak of.
And when I say experienced, I'm talking about you and everyone else who is experienced, sees the big picture, who knows limitations of models, etc.
In a related vein, the idea that PSRR, FR, noise, and HD/THD, output Z, is all there is, is not complete. Its only an approximation. They are things that are commonly measured because they are very practical to measure, not because they tell the whole story. Yet many people come to believe that the models are the reality. It falls on the more experienced and with greater breadth of experience people to keep things honest on not be satisfied with the lowest common denominator models as fully representing physical reality. This relates to the human part you speak of.
And when I say experienced, I'm talking about you and everyone else who is experienced, sees the big picture, who knows limitations of models, etc.
Last edited:
What is not involved with the amplifier and its design? The source at least is very passive in most respect, but the supply, the loudspeaker (and even the living room) is to some more or less degree part of it. Just cold figures does not count to human perception, that's for meters and robots. This triplet is the amplifier actually.
To perhaps sidestep semantic debates, I would expand upon my earlier statement: The sensitivity of the amplifier to the characteristics of the power supply is a characteristic of the amplifier. This then encompasses first order amplifier characteristics (e.g. at what voltages does it function), well known lower order characteristics (e.g. PSRR), and whatever else you'd like to include (e.g. sensitivity to inductive coupling).
Similar: The sensitivity of the power supply to the characteristics of the amplifier is a characteristic of the power supply. This then encompasses first order power supply characteristics (e.g. at what current/voltage demands does it function), well known lower order characteristics (e.g. CMRR), and whatever else you'd like to include (e.g. sensitivity to inductive/capacitive coupling).
And so on...
In my experience, exclusions does not benefit nor yield into a better strategies.
And so on...
In my experience, exclusions does not benefit nor yield into a better strategies.