Bill Fitzpatrick
If it were approaching zero (as just about every amp manufacturer claims) then we would be able to reproduce the recorded stereo image from just about every cheap amp on the market. Since we know that this is not true, we must assume that we would not be measuring near net zero if we knew what to measure but you just gave me another idea on how to test so its all good.
Hello Charles,
" But for mid to high frequencies this parallel network would have influence on the voltage/current relation at the outputput of the amp as well deteriorating the positive effect of current drive (it's series resistor has to be smaller than Rdc of the voice-coil). "
I thought that the optimum RLC circuit should be detuned (Q altered) so that the width of the shunt bandpass characteristic is corresponding to the width of the driver resonance bandpass characteristic, and therefore have nil effect at frequencies above and below resonance.
Regards, Eric.
" But for mid to high frequencies this parallel network would have influence on the voltage/current relation at the outputput of the amp as well deteriorating the positive effect of current drive (it's series resistor has to be smaller than Rdc of the voice-coil). "
I thought that the optimum RLC circuit should be detuned (Q altered) so that the width of the shunt bandpass characteristic is corresponding to the width of the driver resonance bandpass characteristic, and therefore have nil effect at frequencies above and below resonance.
Regards, Eric.
Nania:
No so, methinks. Amps are pretty damn good. The real culprit in the reproduction chain is the speaker system. It's so obvious. Bi or Tri-amping helps but it's not nearly enough.
Rather than dicking around with power cables, interconnects, speaker wire and other such, some of the potential on these forums ought to focus on the real problems. Then, maybe we could get somewhere. I always say, "No sense messing with the icing when the cake won't rise." God knows the manufactures aren't going to do it - they seem to be following the same blind alleys as everyone else.
In an earlier post I mentioned a speaker system that required a helium tank to work. Now, that was real progress - best sound I have ever heard in my life - totally coherrent. Does anyone know what happened to that 25 year old idea?
nania said:
No so, methinks. Amps are pretty damn good. The real culprit in the reproduction chain is the speaker system. It's so obvious. Bi or Tri-amping helps but it's not nearly enough.
Rather than dicking around with power cables, interconnects, speaker wire and other such, some of the potential on these forums ought to focus on the real problems. Then, maybe we could get somewhere. I always say, "No sense messing with the icing when the cake won't rise." God knows the manufactures aren't going to do it - they seem to be following the same blind alleys as everyone else.
In an earlier post I mentioned a speaker system that required a helium tank to work. Now, that was real progress - best sound I have ever heard in my life - totally coherrent. Does anyone know what happened to that 25 year old idea?
When a company worker gets a good idea, it's been reduced to a mere meaningless sentence by the time it gets to marketing.
heehee😀
heehee😀
Bill Fitzpatrick
I appreciate both the naysayers and the supporters responses because they both create the environment to progress and ultimately a better way. The real ill is apathy because then there is complacency and satisfaction with the inferior result. In this case the inferior result is the amp spec sheet and the lack of a quantifyable reason to pair certain speakers with a particular amp to recreate the music as it was recorded. That should be the focus of this thread. That said, I am looking for other ways to test my theory. What I want to do is find a way to measure the voltage and current component of a given interval of music in the coil as it is dissipated in the coil. Then I suspect I will find the theory will hold up and that there is a difference in the "power profile" of different amps delivering a signal.
I know the inclination is to blame all the ills on the speaker since they appear to be less predictable than the static circuits we measure in the amplifier. There is audible evidence to show that the amps ability to maintain the integrity of the music signal in the speaker is also important. Why would two amps make a different stereo image through the same speakers if the amps aren't in some way responsible? To my ears that is the more obvious truth. I intend to find with the theory that it is actually the amps ability to deliver and maintain the audio signal integrity with current that doesn't degrade in the moving coils which maintains the integrity of the music signal. I believe that this variation from the true audio signal is quantifyable with a number that will be much more descriptive than the THD and IMD numbers we use today. Music is multilple simultaneous frequencies delivered with amplitude so it is futile to measure an amps ability to recreate it with measurements of single frequency inputs.
I appreciate both the naysayers and the supporters responses because they both create the environment to progress and ultimately a better way. The real ill is apathy because then there is complacency and satisfaction with the inferior result. In this case the inferior result is the amp spec sheet and the lack of a quantifyable reason to pair certain speakers with a particular amp to recreate the music as it was recorded. That should be the focus of this thread. That said, I am looking for other ways to test my theory. What I want to do is find a way to measure the voltage and current component of a given interval of music in the coil as it is dissipated in the coil. Then I suspect I will find the theory will hold up and that there is a difference in the "power profile" of different amps delivering a signal.
You'll most likely find in your tests that certain amps with the same gain will have reduced output voltage and thus reduced current since the speaker loads are too heavy for those amps to maintain at music levels. Main factors are losses in components and power supply degradation. Different amplifier designs take loads differently, though, at the same gain, any speaker will try to load the same power no matter what amp it's on. Class A amps will differ in loading from AB and again from B. I've often found that most AB amps are good at producing current well since they have good supply headroom. Though, a well designed class A could have much better response to the same load with a very large PSU. You are right in detesting the THD numbers since they really are quite inaccurate, but they do give some idea to how the amplifier performs, just not a very good idea though. Better amps you find usually come with distortion charts however, that have a graph showing distortion into some common load impedances or common brands of speakers. Also, some self powered speakers come with a distortion chart that shows how the amp performs with it's particular speaker which is a very accurate example of distortion figures since you can see how much distortion there is for any given frequency.
Hi Nania
I can basically agree with DUO but I think not many people are aware how linear and nonlinear distortion can affect the perception of "space" in a playback system.
Regards
Charles
I can basically agree with DUO but I think not many people are aware how linear and nonlinear distortion can affect the perception of "space" in a playback system.
Regards
Charles
Bingo, Phase_accurate! At least in part. Maybe it would be better to think of the amp. /speaker interface in these terms.
Linear and Nonlinear distortion are two totally different kinds of distortion.
Linear distortion is in simple terms Frequency/Time related distortion. This is generally caused by frequency response, group delay altering systems (Resistance, inductance capsitance). It could be coupling caps or transformers. Anything other than DC feed back will also affect the amp performance in Linear distortion.
Nonlinear distortion can't be discribed in simple terms as far as i'm concerned but any condition which causes the phase and amplitude of the output to be altered is nonlinear. THD is one measure of nonlinear distortion, IM and TIM are also. What causes these? Any combination of conditions which prevent the amp output signal from having a fixed ratio relationship to the input signal. It could be the way the amp reacts to a reactive load (hehe). Is kickback a problem? Did it current limit on some strange phase angle? Did you run out of voltage swing. NL distortion can be caused by all of these and more.
1. frequency response
2. group delay
3. IM (intermodulation distortion)
4. slu rate (sp)
5. powerfactor control
6. damping factor
The above list IMHO are the main factors that define the sound of any amp. If all of the measured values from above are similar in two amps they will sound very much alike. If any one or more of the values is different it will cause a difference in the "sound" of the amps.
If there was a way to give these measurments some weighted value then maybe there could be a single number to rate an amp.
I don't think so my self but good luck in your search.
Bruce
Linear and Nonlinear distortion are two totally different kinds of distortion.
Linear distortion is in simple terms Frequency/Time related distortion. This is generally caused by frequency response, group delay altering systems (Resistance, inductance capsitance). It could be coupling caps or transformers. Anything other than DC feed back will also affect the amp performance in Linear distortion.
Nonlinear distortion can't be discribed in simple terms as far as i'm concerned but any condition which causes the phase and amplitude of the output to be altered is nonlinear. THD is one measure of nonlinear distortion, IM and TIM are also. What causes these? Any combination of conditions which prevent the amp output signal from having a fixed ratio relationship to the input signal. It could be the way the amp reacts to a reactive load (hehe). Is kickback a problem? Did it current limit on some strange phase angle? Did you run out of voltage swing. NL distortion can be caused by all of these and more.
1. frequency response
2. group delay
3. IM (intermodulation distortion)
4. slu rate (sp)
5. powerfactor control
6. damping factor
The above list IMHO are the main factors that define the sound of any amp. If all of the measured values from above are similar in two amps they will sound very much alike. If any one or more of the values is different it will cause a difference in the "sound" of the amps.
If there was a way to give these measurments some weighted value then maybe there could be a single number to rate an amp.
I don't think so my self but good luck in your search.
Bruce
Hi HDTVman
A lot of mechanisms that come into work here are described in an interesting series of articles by John Watkinson in "Electronics World".
Some of this info can be found inshort form under www.celticaudio.co.uk/technical2.htm.
Regards
Charles
A lot of mechanisms that come into work here are described in an interesting series of articles by John Watkinson in "Electronics World".
Some of this info can be found inshort form under www.celticaudio.co.uk/technical2.htm.
Regards
Charles
Since time and phase discussions have recently come up again in this thread , I am performing the badly needed edit to the opening statement of the "Nania power Theory". Once again restated:
The Nania Audio Power Theory?
We have all been taught Ohms Law as gospel to explain voltage, resistance and current and we have diligently applied Ohms Law as DIYers in our persuit of audio bliss but just as Newtonian Physics was inadequate to explain lightspeed energy we must attempt to find another way to explain what is happening to our music signal as it interacts with a loudspeaker. This thread will introduce the Nania Audio Power Theory to the forum because I think it deserves the full scrutiny of this Forum. If it can hold up against the scrutiny and critical eye of the esteemed peers in this forum, it may prove to have value. Even if it gets shot down in flames I believe it may become an important step to better understanding what makes a great audio experience and that is what I am here for.
Simply stated, the Nania Audio Power Theory asserts that the amplifier is delivering music to the speaker in a "power profile". I offer that the audio signal is delivered as power and that power has a profile of frequencies times amplitude. A given time portion of an identical audio signal can be delivered with a profile of 2 amps and 8 mV or 8 amps and 2mV for a given frequency by two different amplifiers. I believe that the amplifier (and the interconnects to a lesser extent) are the determining factor of how the system controls the speaker drivers and reproduces a signal. In order to identify what I believe we are hearing when a stereo image collapses or fails to form, I make the following assertion. A high current balance in a "power wave" sounds like a tighter grip on the speaker drivers which results in less driver transient motion and a tighter more resolute image. Volts will slap the drivers into motion but lose control relatively quickly and the residual motion (transient) from the inertia results in loose control and a poorer image. To make an analogy, voltage is a punch and current is a grab and pull of the speaker drivers. This is how I can reconcile what I hear to the specifications touted in the sales literature of amplifying electronics.
I understand that what I propose may seem blasphemous to many but my ears and many amplifier auditions have forced me to look for a better way to explain what I am hearing and why I am hearing it. I am currently working on a mathematical proof of the Nania Audio Power Theory and am open to any ideas the members of this forum can provide as to what other kinds of electrical instruments (oscilloscopes, etc.) and testing techniques would validate the Nania Audio Power Theory. I am currently using a pair of tektronix 442's (35MHz/2mV resoultion) but I am always looking for new ideas on how to prove my theory.
The Nania Audio Power Theory?
We have all been taught Ohms Law as gospel to explain voltage, resistance and current and we have diligently applied Ohms Law as DIYers in our persuit of audio bliss but just as Newtonian Physics was inadequate to explain lightspeed energy we must attempt to find another way to explain what is happening to our music signal as it interacts with a loudspeaker. This thread will introduce the Nania Audio Power Theory to the forum because I think it deserves the full scrutiny of this Forum. If it can hold up against the scrutiny and critical eye of the esteemed peers in this forum, it may prove to have value. Even if it gets shot down in flames I believe it may become an important step to better understanding what makes a great audio experience and that is what I am here for.
Simply stated, the Nania Audio Power Theory asserts that the amplifier is delivering music to the speaker in a "power profile". I offer that the audio signal is delivered as power and that power has a profile of frequencies times amplitude. A given time portion of an identical audio signal can be delivered with a profile of 2 amps and 8 mV or 8 amps and 2mV for a given frequency by two different amplifiers. I believe that the amplifier (and the interconnects to a lesser extent) are the determining factor of how the system controls the speaker drivers and reproduces a signal. In order to identify what I believe we are hearing when a stereo image collapses or fails to form, I make the following assertion. A high current balance in a "power wave" sounds like a tighter grip on the speaker drivers which results in less driver transient motion and a tighter more resolute image. Volts will slap the drivers into motion but lose control relatively quickly and the residual motion (transient) from the inertia results in loose control and a poorer image. To make an analogy, voltage is a punch and current is a grab and pull of the speaker drivers. This is how I can reconcile what I hear to the specifications touted in the sales literature of amplifying electronics.
I understand that what I propose may seem blasphemous to many but my ears and many amplifier auditions have forced me to look for a better way to explain what I am hearing and why I am hearing it. I am currently working on a mathematical proof of the Nania Audio Power Theory and am open to any ideas the members of this forum can provide as to what other kinds of electrical instruments (oscilloscopes, etc.) and testing techniques would validate the Nania Audio Power Theory. I am currently using a pair of tektronix 442's (35MHz/2mV resoultion) but I am always looking for new ideas on how to prove my theory.
All right, kids, enough childish hair-pulling and face-scratching! Let's try and keep this discussion civilized. Bill, some of your comments are totally out of line... and nania, you don't need to encourage him with retorts! Sheesh, do we need to call in a moderater here?
OK... on to something constructive:
nania,
I've read your post a few times over, and just can't seem to figure out what specifically you're trying to say. You envision this "power profile", but I'm not sure just what you mean. If you were to plot this profile on a graph, I presume you'd have power on one axis, but what goes on the other axis??? Cone displacement? Frequency? I'm just trying to get a handle on what this profile represents... or is it a 3D or even multi-dimensional graph with more than two variables?
As far as ohm's law is concerned, of course it won't apply to most speaker loads, since it is a simplification for dealing with pure resistance only. But, in all cases, when you're dealing with any electrical impedance, the voltage and current are <i>strictly</i> related, and there is no way to "cheat" this physical relation. That is, for a given load impedance at a given frequency, the current to voltage ratio and relative phase is FIXED. There is no way to change this, it's impossible! Even in transient situations. The relationship may vary with frequency, but it exists and holds true nonetheless. Period.
In theory, it matters not what is imposing a given voltage waveform on a particular electrical impedance. The source of this waveform may as well be a black box. For voltage waveform X, current waveform Y results. End of conversation.
As far as back EMF effects go... well these can be modelled as part of the speaker's impedance. Since we're dealing with a real world Thing, a causal relationship applies to the voltages and currents present in any part of the circuit. The moving mass of the driver assemblies, compliance of the air volumes being compressed and rarified, even the resonant vibrations of the cabinet itself, and of course the momentum and resistive losses in all these elements can conceivably be included as additional resistive, inductive, and capacitive elements in the electrical model of the loudspeaker. You can take this as far as you like, even including the acoustics of the room the speaker is sitting in. Of course, diminishing returns apply to the additional complexity added to the model to account for all of these miniscule details.
BUT...
All of the above is, of course, a first-order approximation. Nobody ever said that the impedance of a speaker was constant! Enter the evil <i>Dr. Non Linearity</i>. As a voice coil changes it's displacement in the magnetic gap, there is almost certainly some non linearity introduced into the impedance of the driver. The same thing applies to magnetic hysteresis ocurring in the iron cores of crossover inductors and the uneven conduction found in electrolytic capacitors. There are plenty of factors which can influence the impedance linearity of a loudspeaker.
Am I getting warmer? Perhaps... Let me continue:
It is my firm belief (and one I think which is well supported), that nonlinearity is anathema to all things Audio. I've seen this have great effect on amplifier circuitry and music recording technology (anyone ever hear a tape recorder that doesn't have a functioning bias circuit?), digital processing etc. Nonlinearity can be detected by the human ear/brain in even the tiniest amounts.
Now here's where I start to pull back the curtain and reveal all (OK, maybe not "all", I'm not a God after all):
Nonlinearity in a speaker load should generally be minimal and quite well controlled, if the drivers or basic technology used are good ones. If steps are taken to minimize the non-linearity, the effect of a slightly nonlinear speaker impedance (IMHO) should be quite small, even negligible... so where's the culprit? May I turn your attention back to the black box. That's right - the amplifier.
Some others touched on some issues which I believe are close to the mark, but don't quite take it the whole mile...
HDTVman came close to the mark with some comments about NL distortion and the various mechanisms in an amplifier which can contribute to it. Speaker impedance compensation was mentioned (too lazy to search back through the thread and look up the poster's name), which will help prevent large phase angles between the voltage and current output of an amplifier from ever occurring, thereby minimizing the risk of such nonlinear effects.
But, these are just scratching the surface, IMHO.
Let me introduce "Chad's Theory". It is, of course, ever evolving and subject to change with further consideration, but was arrived at in a manner something like this:
One day, I was pondering the following topics:
<ul><li>amplifier stability & Zobel networks
<li>load reactance
<li>subjective effects of feedback, specifically that lower feedback supposedly sounds better
<li>trading amplifier linearity for open loop gain (and therefore greater feedback)
<li>subjective claim that damping factor does not significantly influence an amplifier's sound - the evidence being that tube amps have a low damping factor, yet still sound very good
<li>amplifier load invariance</ul>
...and a lightbulb came on inside my head. Getting the idea yet? No? Let me wave my magic wand, and bring all of these things together in one Grand Unified Theory:
<dl><dd>Amplifier sound quality comes down to three basic parameters: open loop linearity, open loop output impedance, and amount of negative feedback present.</dd></dl>
Now, this is stated rather boldly. In reality, these aren't the only factors, but after the usual stuff has been attended to, I think they represent the limiting factors in a majority of today's amplifier designs. Allow me to explain further:
It struck me that a typical high-feedback transistor amplifier can be so influenced by the introduction of a simple capacitive load as to throw it into fits of oscillation and instability. Even self-destruction. If one studies the effect of reactive loading on the prototypical opamp circuit, you'll see that the load reactance, in combination with the <b>open loop</b> output impedance of the amp circuit act to introduce phase shifts in the negative feedback which can therefore encroach on the phase margin of the amp and ultimately violate the boundaries of stable operation. But there are other implications here: what this means is that for an amplifier with a non-zero open loop output impedance, the load impedance variations with frequency can <i>directly</i> influence many of the open loop parameters of the amp including it's basic linearity, open loop gain, and also the phase angle of the feedback.
There is one other general philosophy which I often use as a guiding principle: one must treat the entire audio spectrum equally. That is, distortions which are present in the bass region should also be present in equal proportion in the midrange and treble. The Human ear seems tremendously capable of adapting to different conditions to ignore various distortions which are present in the sound. But, it is also equally capable of detecting when these distortions do not act uniformly on all sounds. If higher frequencies aren't distorted in the same way as lower frequencies, I think you hear it. When transients distort differently than steady state tones, I think you hear it. Any time there is a lack of uniform distortion, the ear seems to recognize this as "artificial sound", since these conditions do not occur in live acoustic situations.
With these two ideas in mind, I then began to see how the amplifier and load interact with far greater clarity than ever before...
I'll leave it as an exercise for my dear readers who have made it through my post this far to consider the above list of topics in the context of these last two ideas.
OK... on to something constructive:
nania,
I've read your post a few times over, and just can't seem to figure out what specifically you're trying to say. You envision this "power profile", but I'm not sure just what you mean. If you were to plot this profile on a graph, I presume you'd have power on one axis, but what goes on the other axis??? Cone displacement? Frequency? I'm just trying to get a handle on what this profile represents... or is it a 3D or even multi-dimensional graph with more than two variables?
As far as ohm's law is concerned, of course it won't apply to most speaker loads, since it is a simplification for dealing with pure resistance only. But, in all cases, when you're dealing with any electrical impedance, the voltage and current are <i>strictly</i> related, and there is no way to "cheat" this physical relation. That is, for a given load impedance at a given frequency, the current to voltage ratio and relative phase is FIXED. There is no way to change this, it's impossible! Even in transient situations. The relationship may vary with frequency, but it exists and holds true nonetheless. Period.
In theory, it matters not what is imposing a given voltage waveform on a particular electrical impedance. The source of this waveform may as well be a black box. For voltage waveform X, current waveform Y results. End of conversation.
As far as back EMF effects go... well these can be modelled as part of the speaker's impedance. Since we're dealing with a real world Thing, a causal relationship applies to the voltages and currents present in any part of the circuit. The moving mass of the driver assemblies, compliance of the air volumes being compressed and rarified, even the resonant vibrations of the cabinet itself, and of course the momentum and resistive losses in all these elements can conceivably be included as additional resistive, inductive, and capacitive elements in the electrical model of the loudspeaker. You can take this as far as you like, even including the acoustics of the room the speaker is sitting in. Of course, diminishing returns apply to the additional complexity added to the model to account for all of these miniscule details.
BUT...
All of the above is, of course, a first-order approximation. Nobody ever said that the impedance of a speaker was constant! Enter the evil <i>Dr. Non Linearity</i>. As a voice coil changes it's displacement in the magnetic gap, there is almost certainly some non linearity introduced into the impedance of the driver. The same thing applies to magnetic hysteresis ocurring in the iron cores of crossover inductors and the uneven conduction found in electrolytic capacitors. There are plenty of factors which can influence the impedance linearity of a loudspeaker.
Am I getting warmer? Perhaps... Let me continue:
It is my firm belief (and one I think which is well supported), that nonlinearity is anathema to all things Audio. I've seen this have great effect on amplifier circuitry and music recording technology (anyone ever hear a tape recorder that doesn't have a functioning bias circuit?), digital processing etc. Nonlinearity can be detected by the human ear/brain in even the tiniest amounts.
Now here's where I start to pull back the curtain and reveal all (OK, maybe not "all", I'm not a God after all):
Nonlinearity in a speaker load should generally be minimal and quite well controlled, if the drivers or basic technology used are good ones. If steps are taken to minimize the non-linearity, the effect of a slightly nonlinear speaker impedance (IMHO) should be quite small, even negligible... so where's the culprit? May I turn your attention back to the black box. That's right - the amplifier.
Some others touched on some issues which I believe are close to the mark, but don't quite take it the whole mile...
HDTVman came close to the mark with some comments about NL distortion and the various mechanisms in an amplifier which can contribute to it. Speaker impedance compensation was mentioned (too lazy to search back through the thread and look up the poster's name), which will help prevent large phase angles between the voltage and current output of an amplifier from ever occurring, thereby minimizing the risk of such nonlinear effects.
But, these are just scratching the surface, IMHO.
Let me introduce "Chad's Theory". It is, of course, ever evolving and subject to change with further consideration, but was arrived at in a manner something like this:
One day, I was pondering the following topics:
<ul><li>amplifier stability & Zobel networks
<li>load reactance
<li>subjective effects of feedback, specifically that lower feedback supposedly sounds better
<li>trading amplifier linearity for open loop gain (and therefore greater feedback)
<li>subjective claim that damping factor does not significantly influence an amplifier's sound - the evidence being that tube amps have a low damping factor, yet still sound very good
<li>amplifier load invariance</ul>
...and a lightbulb came on inside my head. Getting the idea yet? No? Let me wave my magic wand, and bring all of these things together in one Grand Unified Theory:
<dl><dd>Amplifier sound quality comes down to three basic parameters: open loop linearity, open loop output impedance, and amount of negative feedback present.</dd></dl>
Now, this is stated rather boldly. In reality, these aren't the only factors, but after the usual stuff has been attended to, I think they represent the limiting factors in a majority of today's amplifier designs. Allow me to explain further:
It struck me that a typical high-feedback transistor amplifier can be so influenced by the introduction of a simple capacitive load as to throw it into fits of oscillation and instability. Even self-destruction. If one studies the effect of reactive loading on the prototypical opamp circuit, you'll see that the load reactance, in combination with the <b>open loop</b> output impedance of the amp circuit act to introduce phase shifts in the negative feedback which can therefore encroach on the phase margin of the amp and ultimately violate the boundaries of stable operation. But there are other implications here: what this means is that for an amplifier with a non-zero open loop output impedance, the load impedance variations with frequency can <i>directly</i> influence many of the open loop parameters of the amp including it's basic linearity, open loop gain, and also the phase angle of the feedback.
There is one other general philosophy which I often use as a guiding principle: one must treat the entire audio spectrum equally. That is, distortions which are present in the bass region should also be present in equal proportion in the midrange and treble. The Human ear seems tremendously capable of adapting to different conditions to ignore various distortions which are present in the sound. But, it is also equally capable of detecting when these distortions do not act uniformly on all sounds. If higher frequencies aren't distorted in the same way as lower frequencies, I think you hear it. When transients distort differently than steady state tones, I think you hear it. Any time there is a lack of uniform distortion, the ear seems to recognize this as "artificial sound", since these conditions do not occur in live acoustic situations.
With these two ideas in mind, I then began to see how the amplifier and load interact with far greater clarity than ever before...
I'll leave it as an exercise for my dear readers who have made it through my post this far to consider the above list of topics in the context of these last two ideas.
ok, after reading this thread completly, I'll first give it a try to condense the things I understand (an maybe don't).
First of all the original Nania-thing: a power-profile:
IMHO Nania has got it right: as long as input is distorted on it's way to the output, we can try to measure output, try to define the way it is distorted, add the reciproke value of the distortion to the input of the amp (fffffeedback) and hope that this corrected input gives us an output that resembles the original (not corrected) input more than in the original (uncorrected) output. pffff.....(perhaps correction will induce other kinds of distortion as a side-effect cfr. some medication)
Secundo: this is not only amp related: capito, it is not, some amps match with some speakers, others don't : there should be some truth in this. Someone mentioned that motional feedback systems are superior (or at least could be) because sound amplification is measured here as the WHOLE process, not as electrical amplification AND electro-mechanical transformation as two different processes with both of them maybe very favorable distortion figures, but therefore not producing the sound quality one should expect from it.
tertio: amp-speaker interaction: the amp should have a ferm grip on the loudspeaker (sending current, voltage, counter emf comming back etc.) but indeed there's something called matched impedances, known in rf-circuits, maybe or maybe not applicable in lf-circuits like we're talking about. Fact is that high damping factors from the amp as a infinite current source alone are not the only beatific (soul-saving?) thing to meet the ultimat goal.
Enough of this s.....!!!
Solutions, THAT's what we're after.
Fact: Horn-loaded loudspeakers are mostly an easy 'load' for amps, even tube-amps like them.
Fact: Resistive loads are also 'easy'
Fact: most of them tube-amps use output transformers. (!!! I said MOST)
Fact: (this is a personal finding) Mc-Intosh amps (with output transformers) are less sensitive to different 'loads', they tend to 'like' whatever they should feed.
possible conclusion: good (ie. adapted) transformation is maybe a solution. A horn is the ideal mechanical-sound transformation, an electrical (impendance-) transformer could be beneficial before electrical-mechanical transformation.(look at it as kind as glasses to see the load better)
Who build a PASS-amp yet (no offence mr PASS) the way mc-Intosh does it (ie. class A with output transformers)... Could be interesting because IMHO the output-transformer does 'regulate' certain 'things' output-transistors don't like seeing directly, but seeing them 'uncoupled' from the load they tend to 'digest' uneasy loads more easily.
I think the main reason is that transformers (driven by high voltage) can induce more current/voltage (whatever) on the moment needed to compensate the 'weird' needs of a loudspeaker (with or without x-over network) following and equalising the amp's effort (and 'limited' capabilities!!!) to 'drive' the speaker.
The mentioned Zobel-network insertion is maybe just half of the answer....
Only other solution could be via DSP, but I think complete correction would give 'impossible' signals for the amp to send to the speaker.(phase rotations, high Q-factors, etc)
Note: I like high-efficiency loudspeakers because I'm persuaded that everything that is not translated into sound could be counted as 'distortion'.
With dsp-technology nowadays, speaker-builders should care more about good impulse-reproduction than about a frequency response within 2 db limits.
.....Shoot me, I'm ready...
First of all the original Nania-thing: a power-profile:
IMHO Nania has got it right: as long as input is distorted on it's way to the output, we can try to measure output, try to define the way it is distorted, add the reciproke value of the distortion to the input of the amp (fffffeedback) and hope that this corrected input gives us an output that resembles the original (not corrected) input more than in the original (uncorrected) output. pffff.....(perhaps correction will induce other kinds of distortion as a side-effect cfr. some medication)
Secundo: this is not only amp related: capito, it is not, some amps match with some speakers, others don't : there should be some truth in this. Someone mentioned that motional feedback systems are superior (or at least could be) because sound amplification is measured here as the WHOLE process, not as electrical amplification AND electro-mechanical transformation as two different processes with both of them maybe very favorable distortion figures, but therefore not producing the sound quality one should expect from it.
tertio: amp-speaker interaction: the amp should have a ferm grip on the loudspeaker (sending current, voltage, counter emf comming back etc.) but indeed there's something called matched impedances, known in rf-circuits, maybe or maybe not applicable in lf-circuits like we're talking about. Fact is that high damping factors from the amp as a infinite current source alone are not the only beatific (soul-saving?) thing to meet the ultimat goal.
Enough of this s.....!!!
Solutions, THAT's what we're after.
Fact: Horn-loaded loudspeakers are mostly an easy 'load' for amps, even tube-amps like them.
Fact: Resistive loads are also 'easy'
Fact: most of them tube-amps use output transformers. (!!! I said MOST)
Fact: (this is a personal finding) Mc-Intosh amps (with output transformers) are less sensitive to different 'loads', they tend to 'like' whatever they should feed.
possible conclusion: good (ie. adapted) transformation is maybe a solution. A horn is the ideal mechanical-sound transformation, an electrical (impendance-) transformer could be beneficial before electrical-mechanical transformation.(look at it as kind as glasses to see the load better)
Who build a PASS-amp yet (no offence mr PASS) the way mc-Intosh does it (ie. class A with output transformers)... Could be interesting because IMHO the output-transformer does 'regulate' certain 'things' output-transistors don't like seeing directly, but seeing them 'uncoupled' from the load they tend to 'digest' uneasy loads more easily.
I think the main reason is that transformers (driven by high voltage) can induce more current/voltage (whatever) on the moment needed to compensate the 'weird' needs of a loudspeaker (with or without x-over network) following and equalising the amp's effort (and 'limited' capabilities!!!) to 'drive' the speaker.
The mentioned Zobel-network insertion is maybe just half of the answer....
Only other solution could be via DSP, but I think complete correction would give 'impossible' signals for the amp to send to the speaker.(phase rotations, high Q-factors, etc)
Note: I like high-efficiency loudspeakers because I'm persuaded that everything that is not translated into sound could be counted as 'distortion'.
With dsp-technology nowadays, speaker-builders should care more about good impulse-reproduction than about a frequency response within 2 db limits.
.....Shoot me, I'm ready...
Hi All,
Lots of ever decreasing circles here.
Nania, do you have access to multiple amps to test? I would suggest a simple way to analyse their outputs and differences:
1) Choose a particularly demanding musical passage;
2) Make a simple differential amplifier that allows one of its inputs to be turned down (by a high impedance potentiometer volume control);
3) Connect an 8 Ohm resistive load to your amp;
4) Connect the input and output of the amp to your differential amplifier, feed in a pure 1 kHz tone, balance the input/output levels, and connect its output to a storage oscilloscope;
5) Feed in your musical piece, and watch the imperfections in your power amp as it drives a perfect load;
6) Replace the 8 Ohm resistive load with a nasty loudspeaker;
7) Watch the oscilloscope as all hell breaks loose;
8) Repeat for your set of different amplifiers;
9) Tell us what you have learned.
The above exercise will be very informative to anybody who thinks amplifiers are anywhere near perfect, and will give some interesting real figures to argue about...
Bill.
Lots of ever decreasing circles here.
Nania, do you have access to multiple amps to test? I would suggest a simple way to analyse their outputs and differences:
1) Choose a particularly demanding musical passage;
2) Make a simple differential amplifier that allows one of its inputs to be turned down (by a high impedance potentiometer volume control);
3) Connect an 8 Ohm resistive load to your amp;
4) Connect the input and output of the amp to your differential amplifier, feed in a pure 1 kHz tone, balance the input/output levels, and connect its output to a storage oscilloscope;
5) Feed in your musical piece, and watch the imperfections in your power amp as it drives a perfect load;
6) Replace the 8 Ohm resistive load with a nasty loudspeaker;
7) Watch the oscilloscope as all hell breaks loose;
8) Repeat for your set of different amplifiers;
9) Tell us what you have learned.
The above exercise will be very informative to anybody who thinks amplifiers are anywhere near perfect, and will give some interesting real figures to argue about...
Bill.
no offence, but again same mistake: measure the output of the WHOLE system and compensate (by means of a microphone or motion sensor on loudspeaker), not just after the amp.
Is it the perfect amp you're after, or is it the perfect sound????
Acoustic result is meaningfull, not whats in between.
I think this is what nania meant, try to get the right transformation formulas into the black box called amplifier.
That why Ohm's law is meaningfull but not the only factor!!!
Is it the perfect amp you're after, or is it the perfect sound????
Acoustic result is meaningfull, not whats in between.
I think this is what nania meant, try to get the right transformation formulas into the black box called amplifier.
That why Ohm's law is meaningfull but not the only factor!!!
nania etc
Woneil,
The problem with the balancing is that you need to balance not only the level but also the phaseshift (group delay) of the output vs the input. Now, if you have a simple (in phase shift terms) amp with a dominant pole, meaning a single roll-off point, you can balance the phaseshift, but if the amp has multiple roll-off points it gets VERY difficult. So, whatever difference signal you get with music input may well be due to different phaseshifts in the amp vs the balancing path.
Peter Walker of Quad used to do this at HiFi shows to show the (absence of) distortion in the QUAD405 I think it was, but this was a "simple" amp. I never heard it, so I don't know how succesfull it was.
Jan Didden
Woneil,
The problem with the balancing is that you need to balance not only the level but also the phaseshift (group delay) of the output vs the input. Now, if you have a simple (in phase shift terms) amp with a dominant pole, meaning a single roll-off point, you can balance the phaseshift, but if the amp has multiple roll-off points it gets VERY difficult. So, whatever difference signal you get with music input may well be due to different phaseshifts in the amp vs the balancing path.
Peter Walker of Quad used to do this at HiFi shows to show the (absence of) distortion in the QUAD405 I think it was, but this was a "simple" amp. I never heard it, so I don't know how succesfull it was.
Jan Didden
maybe loudspeaker manufacturers should formulate their needs to dsp-programmers and work out something similar to a RIAA-curve in vinyl: keep those things easier for the mechanical side that you can compensate by electronic or digital manipulation.
We can find such things allready for some time in active bass-boost circuitry (mmmmm ever heard of the b*sE eq on ()@... or sunfire's little bc-bomb).
We only need standardisation to know in wich way the amp should be non-linear, make a standard digital IC and....
ok, your turn.
We can find such things allready for some time in active bass-boost circuitry (mmmmm ever heard of the b*sE eq on ()@... or sunfire's little bc-bomb).
We only need standardisation to know in wich way the amp should be non-linear, make a standard digital IC and....
ok, your turn.
- Status
- Not open for further replies.