john curl said:
No, I'd had it figured out (the only way you could get any energy storage outside of a capacitor would be to use an inductor). In fact, I seem to recall Steve Bench having written about this on his website some years ago.
I just wasn't interested in getting too terribly involved with this in case the moderators decided to pitch it over to Texas in spite of your "hypothetical design" qualifier. You know how touchy they can be about certain subjects.
se
Ah, found it.
Just scroll down a bit to "No R No C Amplifiers and Power Supplies."
Steve's Tube Pages
se
Just scroll down a bit to "No R No C Amplifiers and Power Supplies."
Steve's Tube Pages
se
What is important is the CONCEPT of having a battery free, transformer free, capacitor free, AC powered amp. Yes, it is possible, however not very practical. It means that transformers and/or capacitors are unnecessary in the absolute sense. This means that more must be going on than caps storing energy or transformers are providing energy. This is what is important to note.
Going back a few posts...
Yeah, the measurement & instrumentation folks may have a good chunk of the picture, although T&M is, in and of itself, quite a broad (and fascinating!) field. Certainly, the focus on noise, shielding and grounding goes a long way, as well as the knowledge related to how one handles very delicate signals without degradation. I've picked up some excellent knowledge by studying the service manuals of some of the old test gear I own. Great way to learn, BTW - get some old, cheap equipment off ebay, then fix it up. Some of that gear still has no modern equivalent, and most of it is every bit as good - I have a nixie tube capacitance meter that reads to 1 fF resolution. One of my fav's. Cost me $40. Anyway, I digress... T&M is a good field to study, but there are still substantial elements of the picture which they are missing.
For example, the peculiarities of human psychoacoustics - which types of artifacts and anomolies in an audio signal are or aren't important to the perception of music. I don't think there is any other field of electronics in which this is a factor. Not even telephony (which is, strictly speaking, part of the audio discipline), in which the factors under consideration are superficially similar, but actually quite different - the fidelity bar is set far, far lower - in fact, for cost reasons the bar is deliberately set as low as it can be without risking lost sales... "how much can we degrade the signal before average people will complain?" kind of stuff. Or how about the large-signal linearity of power transistors? Offhand, I can't think of another EE discipline which cares about this - mostly they're concerned with switching losses, or perhaps frequency response.
The point may be academic (pun intended) anyway - I wager you'll have a very hard time finding a school with a good, comprehensive T&M instrumentation design curriculum. It's kind of too specific, like audio, for a university to teach - that's not what they're for. A EE degree is merely the foundation upon which to build an expertise in one's field.
So, I tend to agree with Grey - A EE degree will get you the basics, if you paid attention and didn't just sail through on rote memorization. Then, there is a great deal of audio-specific knowledge one must acquire through experience and interaction with other engineers, studying the designs of others, and isn't easily obtained all from one place. I have a EE degree and 15 years of audio design (most of it as a hobby, admittedly) under my belt, and almost as many years of industry experience, and... still learning new things in audio all the time. Many a time I've come back to a schematic I once disliked, only to see it in a new light, finally understanding what made it special - that which I did not recognize the first time. By comparison, designing a computer motherboard (my day job) is no sweat - a heck of a lot of work, but not what I would call technically challenging.
So let's put this power thing to rest. Power flows in from the AC line (where the heck else is it going to come from?). Therefore what goes out to the speaker must come in through the power cord. Simple enough, right? If you want to call the Hot phase "input", then Neutral is the "return".
Between the amp circuit and the mains, we have the power supply - transformer, diodes, and caps. Power supply filter caps are exactly that - a filter. For a conventional supply, high frequency AC currents on the load side will be mostly attenuated before they appear on the mains side (that's the idea anyway, but it's never ideal of course). Low frequencies less so. This is design dependent. If you're willing to sacrifice some power, one can do something like a balanced class-A design, wherein the AC currents null to zero at the power supply caps (power delivered to the load is offset / balanced by power dissipated as heat). Then you can have a steady DC drawn from the PSU, and thus the mains, but you also have to deal with the thermal implications - including power dissipation in the output devices which varies with the signal, and the difficulty of cooling it all.
Unless your amp circuit fits this description, you'll always have some degree of signal power flowing back out to the AC line. And, in all cases, you still have AC line disturbances flowing forward into your circuit - again, attenuated by the PSU filter. After that, it's up to your circuit's PSRR to deal with such noise.
The bottom line is, AC line quality matters. Power supply design matters. Circuit design matters. And they're inter-related.
And umm.... what was the argument again? I forget.
PS - Steve Bench has some cool stuff on his site... definitely a good one to peruse on a lazy Tuesday evening. Speaking of which... I'm off. Might be back later, but don't count on it.
KSTR said:
Yeah, the measurement & instrumentation folks may have a good chunk of the picture, although T&M is, in and of itself, quite a broad (and fascinating!) field. Certainly, the focus on noise, shielding and grounding goes a long way, as well as the knowledge related to how one handles very delicate signals without degradation. I've picked up some excellent knowledge by studying the service manuals of some of the old test gear I own. Great way to learn, BTW - get some old, cheap equipment off ebay, then fix it up. Some of that gear still has no modern equivalent, and most of it is every bit as good - I have a nixie tube capacitance meter that reads to 1 fF resolution. One of my fav's. Cost me $40. Anyway, I digress... T&M is a good field to study, but there are still substantial elements of the picture which they are missing.
For example, the peculiarities of human psychoacoustics - which types of artifacts and anomolies in an audio signal are or aren't important to the perception of music. I don't think there is any other field of electronics in which this is a factor. Not even telephony (which is, strictly speaking, part of the audio discipline), in which the factors under consideration are superficially similar, but actually quite different - the fidelity bar is set far, far lower - in fact, for cost reasons the bar is deliberately set as low as it can be without risking lost sales... "how much can we degrade the signal before average people will complain?" kind of stuff. Or how about the large-signal linearity of power transistors? Offhand, I can't think of another EE discipline which cares about this - mostly they're concerned with switching losses, or perhaps frequency response.
The point may be academic (pun intended) anyway - I wager you'll have a very hard time finding a school with a good, comprehensive T&M instrumentation design curriculum. It's kind of too specific, like audio, for a university to teach - that's not what they're for. A EE degree is merely the foundation upon which to build an expertise in one's field.
So, I tend to agree with Grey - A EE degree will get you the basics, if you paid attention and didn't just sail through on rote memorization. Then, there is a great deal of audio-specific knowledge one must acquire through experience and interaction with other engineers, studying the designs of others, and isn't easily obtained all from one place. I have a EE degree and 15 years of audio design (most of it as a hobby, admittedly) under my belt, and almost as many years of industry experience, and... still learning new things in audio all the time. Many a time I've come back to a schematic I once disliked, only to see it in a new light, finally understanding what made it special - that which I did not recognize the first time. By comparison, designing a computer motherboard (my day job) is no sweat - a heck of a lot of work, but not what I would call technically challenging.
So let's put this power thing to rest. Power flows in from the AC line (where the heck else is it going to come from?). Therefore what goes out to the speaker must come in through the power cord. Simple enough, right? If you want to call the Hot phase "input", then Neutral is the "return".
Between the amp circuit and the mains, we have the power supply - transformer, diodes, and caps. Power supply filter caps are exactly that - a filter. For a conventional supply, high frequency AC currents on the load side will be mostly attenuated before they appear on the mains side (that's the idea anyway, but it's never ideal of course). Low frequencies less so. This is design dependent. If you're willing to sacrifice some power, one can do something like a balanced class-A design, wherein the AC currents null to zero at the power supply caps (power delivered to the load is offset / balanced by power dissipated as heat). Then you can have a steady DC drawn from the PSU, and thus the mains, but you also have to deal with the thermal implications - including power dissipation in the output devices which varies with the signal, and the difficulty of cooling it all.
Unless your amp circuit fits this description, you'll always have some degree of signal power flowing back out to the AC line. And, in all cases, you still have AC line disturbances flowing forward into your circuit - again, attenuated by the PSU filter. After that, it's up to your circuit's PSRR to deal with such noise.
The bottom line is, AC line quality matters. Power supply design matters. Circuit design matters. And they're inter-related.
And umm.... what was the argument again? I forget.
PS - Steve Bench has some cool stuff on his site... definitely a good one to peruse on a lazy Tuesday evening. Speaking of which... I'm off. Might be back later, but don't count on it.
http://www.jensen-transformers.com/apps_wp.html
the generic seminar handout is pretty good, as well as the rest
the generic seminar handout is pretty good, as well as the rest
john curl said:
I don't see that it means that at all. All you've really done is to point out that inductors also store energy.
What I've said in this discussion has not been to say or imply that there's nothing else going on. It was simply to say that you had got something very basic and fundamental, wrong. The other things going on don't change the fundamentals. And unless you understand the fundamentals, there's nothing to build upon.
se
hifiZen said:
Well, it wasn't a "power" thing, it was a "current" thing, specifically as it relates to the current "return path" of the amplifier. And in that context, the current that goes out to the speaker isn't the current that comes from the power cord, nor does it return to it.
There is a current loop (i.e. "send" and "return") formed between the secondary of the line transformer up on the pole and the primary of the power transformer. Then there is a current loop formed between the power transformer's secondary and the power supply's capacitors which is responsible for keeping the capacitors charged up. Finally there is a current loop formed between the supply's capacitors and the loudspeaker which is responsible for driving the loudspeaker.
The salient point here is that each of these three loops are essentially distinct with each having its own "send" and "return."
Actually the main supply caps would more accurately be called "reservoir caps." While they do function as a filter by their nature, their fundamental role isn't one of filtering, but rather energy storage. You want to keep the average voltage across the caps some substantial percentage of your peak rail voltage (i.e. keep power supply ripple reasonable). And that's done by way of energy storage.
That won't give you a steady DC draw from the mains. You're still going to draw current from the mains in short pulses. Current's only drawn from the mains when the transformer's secondary voltage exceeds the voltage across the power supply caps. And even with balanced class A, the signal will cause the supply caps to be drawn down to various degrees which means the duration of the pulses will vary.
se
Actually, in this case, inductors smooth the AC waveform to make it into DC. Since AC always has some percentage of time where there is no output, you have to store some energy somewhere, just to not have a dropout in the audio. Still, it points out that transformers and caps are NOT necessary for audio reproduction, even IF you don't use batteries. However, IF you use a battery, you might just have to connect the return lead to the battery, or else it 'might' not work, and if you looked across or even inside a battery when music was playing, using it as a source, you would find return energy from the loudspeaker that needs to be damped out.
john curl said:
Say WHAT!?!?!?
Now you're claiming that inductors are RECTIFIERS and convert AC into DC!?!?!?!
Are you serious?
Um, John, with a continuous waveform like a sinusoid, any point on that waveform, including the zero point, is instantaneous. What percentage of time is "instantaneous"? How much time does a 60Hz sinewave spend at zero volts?
About all you've pointed out so far is that you seem to lack any sort of understanding of some very basic stuff here.
The "return lead" being what exactly?
Damped out? What, are you talking about some sort of resonance?
se
Fellow designers, I'm afraid that I have asked too much, and I just can't do a programmed learning version of what I was trying to convey. I DO skip over what I consider 'obvious' preconditions, just because I suspect that most of you can fill in the blanks. However, it hasn't worked so far, and I should try another approach, if I have something new to convey.
- Status
- Not open for further replies.