When this occurs, thats when you need to check your FR and adjust accordingly...
//
Perhaps as digital as class-d. What happens first is half wave rectification. But I doubt there is numerical quantification... so not digital, right?
//
An other article I would hope for!
A complete reproduction chain discussion and its problems and that is possible...
//
True.
Here is a quote about modern sampling theory:
Compressed sensing (CS) [1, 2] is a new framework about signal acquisition and sensors, and the development of its theory and technology has a profound impact on the field of signal acquisition, analysis technology, processing methods, etc.
Compressed sensing is a new sampling theory, it is sampling randomly under the frequency far less than the Nyquist sampling frequency to obtain the partial information of the signal, and then restoring the whole signal through the nonlinear reconstruction algorithm. Compressed sensing is a new idea about signal acquisition, representation, and processing, and it not only makes people re-examine the existing signal processing technology, but also brings a wealth of new ideas about signal acquisition and processing, which greatly promotes the combination of mathematics theory and engineering applications.
https://www.sciencedirect.com/topics/engineering/compressed-sensing#:~:text=Compressed sensing is a new sampling theory, it is sampling,through the nonlinear reconstruction algorithm.
Yes, but we have driven the thread far off-topic. Not sure what is OP's intention. But it is a destiny of every thread on a public forum. It may be interesting, though, to brain-storm new themes, as the original one is usually quickly depleted.
In the "Hey, that's neat" side of the articles ... I must say that I'm fascinated by the statement below...
To me... that's neat! I'd love to learn more. It's something I'd have never considered. A myth to bust might be something like... things above 20kHz don't matter.
First... I can't believe that I'm alone in being lazy. I'm a hobbyist that, like some, started the hobby as an 'audiophile'. After assembling a number of kits / boards etc., my curiosity led me further down the road.
Some of the 'tone boxes' I built certainly sounded different to me. However, did they perform the way the designer had intended as a stand-alone amplifier with a dummy load on a bench? Did they perform well as part of a system?
I'd love to see more discussion around how equipment behaves 'at the edges'. What do great designers do to help dumb-dumbs like me not build an oscillator? Given the relatively low entry cost of some oscilloscopes today, what can a new hobbyist do to understand this phenomena? Are there simple breadboard experiments (or LTSpice sims) that a hobbyist may conduct to illustrate to the curious what to some I'm sure is a boring old topic.
I certainly don't want anyone (Jan) to redo what's already known. I certainly don't want "cables" to be the hot topic again.
However, that might be far too basic for the intended audience. It may also be "old news". I'm enjoying reading through the posts...
Not to put words into @PMA's mouth... but from the sequence of posts, does this imply that these oscillations at >1Mhz could affect the "realism" of the sound?
To me... that's neat! I'd love to learn more. It's something I'd have never considered. A myth to bust might be something like... things above 20kHz don't matter.
First... I can't believe that I'm alone in being lazy. I'm a hobbyist that, like some, started the hobby as an 'audiophile'. After assembling a number of kits / boards etc., my curiosity led me further down the road.
Some of the 'tone boxes' I built certainly sounded different to me. However, did they perform the way the designer had intended as a stand-alone amplifier with a dummy load on a bench? Did they perform well as part of a system?
I'd love to see more discussion around how equipment behaves 'at the edges'. What do great designers do to help dumb-dumbs like me not build an oscillator? Given the relatively low entry cost of some oscilloscopes today, what can a new hobbyist do to understand this phenomena? Are there simple breadboard experiments (or LTSpice sims) that a hobbyist may conduct to illustrate to the curious what to some I'm sure is a boring old topic.
I certainly don't want anyone (Jan) to redo what's already known. I certainly don't want "cables" to be the hot topic again.
However, that might be far too basic for the intended audience. It may also be "old news". I'm enjoying reading through the posts...
Oscillations change the sound and you never know how, because the cases and issues are individual. However, I would not speak about "added realism" due to oscillations. IME they always worsen the sound.
When an amplifier starts to oscillate, the oscillation grows until some non-linear phenomenon stops it from growing any further. That usually means that a transistor or valve in the amplifier is driven far from the intended operating point into some clipping mode, and that usually though not necessarily means that the amplifier distorts like crazy.
One basic physical mechanism (demodulation/rectification and resulting bias shifts) is already well known. There are app notes about it in opamps including by companies like TI (IIRC Scott Wurcer mentioned it too). Yet its true that some people who consider themselves objectivists still believe the "above 20kHz can't matter" myth.
Last edited:
@PMA Thank you. Apologies. I didn't mean to imply or state that oscillations could make anything better. It seems I interpreted your intention properly.
Oscillations could make the realism (or some other sonic aspect) 'worse'.
I'd have never considered that. I never considered that anything outside the generally accepted 20Hz to 20kHz band "mattered" to the sound. I had considered that it may result in poor longevity / performance. I had never considered that it may impact the sonics.
Again, thank you!
Edited to add - @MarcelvdG - Thank you also. I can grasp that concept. Much appreciated!
Oscillations could make the realism (or some other sonic aspect) 'worse'.
I'd have never considered that. I never considered that anything outside the generally accepted 20Hz to 20kHz band "mattered" to the sound. I had considered that it may result in poor longevity / performance. I had never considered that it may impact the sonics.
Again, thank you!
Edited to add - @MarcelvdG - Thank you also. I can grasp that concept. Much appreciated!
Thank you. I know it is hard to imagine that >>1MHz oscillations may affect audio band, in case you do not "live" with electronics circuits. We see something at the output, but it is the inner structure behaviour that is grossly affected. I can share a case of an audio preamp that oscillated at 253MHz + multiples. No simulation would tell you, loop analysis perfectly stable. Parasitic impedances and diamond buffer output stage. As a result, low frequency hum and distortion!
^ @PMA
To me.... that's neat!
I know a lot of this is "old hat" to seasoned EEs, and it may be boring for even novice circuit designers. So, I appreciate anyone that takes the time to teach. For the average solder slinger that has just barely learned to interpret a FFT, it's fantastic information.
Edited to add - A few posts snuck in. This was in response to #154.
To me.... that's neat!
I know a lot of this is "old hat" to seasoned EEs, and it may be boring for even novice circuit designers. So, I appreciate anyone that takes the time to teach. For the average solder slinger that has just barely learned to interpret a FFT, it's fantastic information.
Edited to add - A few posts snuck in. This was in response to #154.
«We believe that an amplifier should be designed for low distortion and wide bandwith without feedback. Negative feedback is then added to make an already good design perform even better; it is not used to "clean up" problems in the basic design.» Bart Locanthi.
It's a bit saying the obvious?
It's kind of a sentence without meaning in that way.
Yeah, you can either improve the distortion more, or clean up problems.
That literally means the same thing, as long as the element that's controlling the feedback is good enough.
Just the threshold of issues is lower, but they are problems nonetheless.
Otherwise you would have no distortion to begin with.
It's kind of a sentence without meaning in that way.
Yeah, you can either improve the distortion more, or clean up problems.
That literally means the same thing, as long as the element that's controlling the feedback is good enough.
Just the threshold of issues is lower, but they are problems nonetheless.
Otherwise you would have no distortion to begin with.