Curl, ur a trouble maker! 
In fact, I think some of your equipment uses funny wire in it that ain't copper... HA! Proof!! ... you think because your name is on that Parasound pre-amp and it looks nice that anyone here is going to take you seriously?? That is one sweet looking power transformer in it. Is that what makes it sound good? Someone told me that if it has little switches on the rear panel, then it is really good. Why aren't they gold bats on the switches? I mean, really now!!
Just for the record, personally I am not saying that anyone should dispense with their output inductor. My position, and I am sticking to it, is that one can definitely hear those silly things. And, you don't have to go too far to do that, imho.
Now can't we all just get along?
_-_-bear
In fact, I think some of your equipment uses funny wire in it that ain't copper... HA! Proof!! ... you think because your name is on that Parasound pre-amp and it looks nice that anyone here is going to take you seriously?? That is one sweet looking power transformer in it. Is that what makes it sound good? Someone told me that if it has little switches on the rear panel, then it is really good. Why aren't they gold bats on the switches? I mean, really now!!
Just for the record, personally I am not saying that anyone should dispense with their output inductor. My position, and I am sticking to it, is that one can definitely hear those silly things. And, you don't have to go too far to do that, imho.
Now can't we all just get along?
_-_-bear
Hi Hartono,
Prior to Keith Howard's article, I made some simple calculations on thermal effects in coils of drivers and submitted them to Mike Renardson. He suggested to publish them. I am sorry to be currently unable to find them in my computer.
The conclusion was : considering the ratio of peak to average in signals and the thermal coefficient of the resistivity of copper or aluminium, the thermal resistance and the thermal time constant of coils, thermal effects on them do not need to be a problem in standard domestic conditions.
I would be glad to know your interpretation of Howard's data.
More about the amp driver interface at RF is coming later, I hope.
Prior to Keith Howard's article, I made some simple calculations on thermal effects in coils of drivers and submitted them to Mike Renardson. He suggested to publish them. I am sorry to be currently unable to find them in my computer.
The conclusion was : considering the ratio of peak to average in signals and the thermal coefficient of the resistivity of copper or aluminium, the thermal resistance and the thermal time constant of coils, thermal effects on them do not need to be a problem in standard domestic conditions.
I would be glad to know your interpretation of Howard's data.
More about the amp driver interface at RF is coming later, I hope.
I find it interesting that people say they've never heard of output coils affecting sound quality. It's hardly news.
I also find it sad that after all the numerous times that "impossible" things have been proven (audibility of passive components, audibility of absolute phase, etc.) there are still some who scream "IMPOSSIBLE" every time something is mentioned that doesn't immediately make sense to them.
Is it intuitively obvious that absolute phase should be audible? No. Is it logical that there should necessarily be a mechanism in the human ear/brain pairing that should be sensitive to such a thing? No. But people were saying it was audible for years. And being criticized for it. And being ridiculed for it. Then someone went and proved it. Oh, how rude!
Of course, we all know that it was inaudible the day before those test results came out, and only became audible the day after...right?
Sigh...
I got great pleasure from reading the chapter in Douglas Self's book wherein he "debunks" audio "myths" and seeing him (rather harrumphingly) eat his words. It took courage to do that in print, and I respect him for it, though I disagree with him on other matters.
Yes, there are things out there that are snake oil. But there are also things that are audible. And stranger things than dreamt of in your philosophy...
The only rational response when confronted with something new is either:
--I have no opinion, as I have not tried it.
--I have tried it, but heard no difference. (There are any number of reasons this might be so, without having to conclude that the phenomenon doesn't exist.)
Both options are intellectually honest, ethical, and may potentially save you from having to eat your words later.
Grey
P.S.: These endless arguments get damned tedious, particularly when it's plain that those who cry, "It ain't so!" have never tried anything of the sort. Armchair theorizing has its place, but if it's never tested against reality it's just hot air.
I also find it sad that after all the numerous times that "impossible" things have been proven (audibility of passive components, audibility of absolute phase, etc.) there are still some who scream "IMPOSSIBLE" every time something is mentioned that doesn't immediately make sense to them.
Is it intuitively obvious that absolute phase should be audible? No. Is it logical that there should necessarily be a mechanism in the human ear/brain pairing that should be sensitive to such a thing? No. But people were saying it was audible for years. And being criticized for it. And being ridiculed for it. Then someone went and proved it. Oh, how rude!
Of course, we all know that it was inaudible the day before those test results came out, and only became audible the day after...right?
Sigh...
I got great pleasure from reading the chapter in Douglas Self's book wherein he "debunks" audio "myths" and seeing him (rather harrumphingly) eat his words. It took courage to do that in print, and I respect him for it, though I disagree with him on other matters.
Yes, there are things out there that are snake oil. But there are also things that are audible. And stranger things than dreamt of in your philosophy...
The only rational response when confronted with something new is either:
--I have no opinion, as I have not tried it.
--I have tried it, but heard no difference. (There are any number of reasons this might be so, without having to conclude that the phenomenon doesn't exist.)
Both options are intellectually honest, ethical, and may potentially save you from having to eat your words later.
Grey
P.S.: These endless arguments get damned tedious, particularly when it's plain that those who cry, "It ain't so!" have never tried anything of the sort. Armchair theorizing has its place, but if it's never tested against reality it's just hot air.
GRollins said:
That is irrational.
How can something be new if you have already tried it? And if it can, there is at least a third equally good response: "I have tried it and I heard a difference". So either your points should be three or only one.
However. let me add a response which I find just as acceptable, although theory hating pragmatics will probably run away and hide in a corner:
-- Interesting. If it is actually audiable, what then could the reason be and what do we fail to measure and what are we missing in the way we are applying the theory.
Here's an experiment that I tried a couple of years ago.
The LTSpice simulator has the capability to assign a WAV file to a voltage source and capture a voltage to a WAV file.
By using a simulated output inductor/resistor combination and a simulated speaker load, the error voltage for an actual musical input (which is just the voltage across the inductor) can be captured to a WAV file and you can listen to it.
In case anybody wants to try this, I've attached an LTSpice file of the simulated loudspeaker load that Stereophile uses.
I have not tried the experiment with the Stereophile speaker simulation yet, but I did try it with Graham's simulated Ariel speaker. The first music I tried it with, I got total silence for the error signal. Then I tried it with a Miles Davis track that had piercing trumpet riffs. In that case, I could just barely detect a sound with the volume control cranked all the way up. But at this level, listening to the input WAV file would have blown out my headphones, eardrums or both. Listening to the WAV file was done with a Squeezebox hooked up to a Benchmark DAC1 using the DAC1's headphone outputs and Sennheiser 580 'phones. YMMV.
This prior test was done with a 6uH inductor in parallel with 8.2 Ohms. That's a much larger inductor than you'd normally use for a domestic power amp.
The LTSpice simulator has the capability to assign a WAV file to a voltage source and capture a voltage to a WAV file.
By using a simulated output inductor/resistor combination and a simulated speaker load, the error voltage for an actual musical input (which is just the voltage across the inductor) can be captured to a WAV file and you can listen to it.
In case anybody wants to try this, I've attached an LTSpice file of the simulated loudspeaker load that Stereophile uses.
I have not tried the experiment with the Stereophile speaker simulation yet, but I did try it with Graham's simulated Ariel speaker. The first music I tried it with, I got total silence for the error signal. Then I tried it with a Miles Davis track that had piercing trumpet riffs. In that case, I could just barely detect a sound with the volume control cranked all the way up. But at this level, listening to the input WAV file would have blown out my headphones, eardrums or both. Listening to the WAV file was done with a Squeezebox hooked up to a Benchmark DAC1 using the DAC1's headphone outputs and Sennheiser 580 'phones. YMMV.
This prior test was done with a 6uH inductor in parallel with 8.2 Ohms. That's a much larger inductor than you'd normally use for a domestic power amp.
FWIW, a while back I had to design a custom power amp for a research application. The specs for noise and distortion were extremely low and they couldn't find a commercial manufacturer for it. I found that the design of the output inductor was critical in whether or not I could meet the specs. With an air wound coil of fairly heavy wire, there was no problem. Within *anything* in the coil, like winding it on a resistor, or attempting to reduce its size by using a powder core, there was a definite increase in THD. I'm not saying THD is the end-all criteria, or that the coil was audible (I didn't listen), but the thing definitely affects the signal in the audio band, in spite of its small value. If you can design a stable amp without one, you're that much further ahead.
[Remember the Dynaco Stereo-something, with the coils wound around the big output capacitors- talk about two strikes and you're out, and that's without mention of the 2N3055 output devices!]
[Remember the Dynaco Stereo-something, with the coils wound around the big output capacitors- talk about two strikes and you're out, and that's without mention of the 2N3055 output devices!]
hi Forr,
peak to average might looks ok , but how little is considered okay ?
and also because of the construction voice coil does not dissipate heat in linear manner, so dynamic effect is worse that one might assume. It's asking a speaker too much to expect coil to dissipate in a linear manner at all input level at all speed and offset.
the data in the stereophile article clearly shows the woofer impedance throughout the test, I think that's substantial
stereophile data: averaged over 5 second(way too long) the graph shows that when the music plays the woofer impedance vary by 0.3 ohm averaged(discounted ? ) over 5 second that is 7.5%
even the result for the tweeter 0.25% (0.01 ohm) modulated impedance change for 4 ohm tweeter is much worse than any capacitor, amp, etc
well at least I don't want to buy any capacitor, cable or amp with 0.25% modulation, or 7.5 %
what is the point of having 0.01% THD or IMD if the tweeter output is modulated by 0.25% ?
as for the output inductor , as I said before it depend on the design goal, if one is to strive for near 0 ohm output impedance up to 30khz
the inductor can be a set back.
cheers
peak to average might looks ok , but how little is considered okay ?
and also because of the construction voice coil does not dissipate heat in linear manner, so dynamic effect is worse that one might assume. It's asking a speaker too much to expect coil to dissipate in a linear manner at all input level at all speed and offset.
the data in the stereophile article clearly shows the woofer impedance throughout the test, I think that's substantial
stereophile data: averaged over 5 second(way too long) the graph shows that when the music plays the woofer impedance vary by 0.3 ohm averaged(discounted ?
) over 5 second that is 7.5%even the result for the tweeter 0.25% (0.01 ohm) modulated impedance change for 4 ohm tweeter is much worse than any capacitor, amp, etc
well at least I don't want to buy any capacitor, cable or amp with 0.25% modulation, or 7.5 %
what is the point of having 0.01% THD or IMD if the tweeter output is modulated by 0.25% ?
as for the output inductor , as I said before it depend on the design goal, if one is to strive for near 0 ohm output impedance up to 30khz
the inductor can be a set back.
cheersGRollins said:
The debate isn’t strictly over wether or not a sufficiently big inductor can induce audible differences – it is over wether or not the inductor necessarily incurs sonic degradation.
I don’t think that a small inductor necessary to ensure stability into capacitive loads in a well-designed amplifier has a significant effect. This comes from building and listening to amplifiers. I can’t hear the difference between a uH or two in series with my speaker leads and my hearing isn’t too bad. The last time it was clinically measured, it was pretty well spot on.
Now I’ll concede that if the inductor was big enough, an audible difference could very well be induced. But wether or not the audio quality was degraded or improved would be something for the listener to subjectively decide. No body here can therefore make the blanket assertion that all output inductors “sound bad”, so that we should build amplifiers without them by necessity.
Graham Maynard said:
Distortion? The major contribution would be a small amplitude/phase variation. You can call that distortion if you want. What significance do you think a 1-2uH inductor has in comparison to the voice coil inductance of the speaker, or the tweeters crossover components such as the big 180uH inductor that would be necessary to shunt an 8 ohm tweeter for a crossover of 10kHz at 12db/oct??
Cheers,
Glen
Hartono said:dimitri posted :
"anything NEAR the coil gave me rise in distortion"
sometimes it's not possible to remove the output coil but it's becoming a hard component to live with.
Halcro uses an output inductor in their amps to ensure stability when driving capacitive loads. Their amps have the lowest distortion performance of any on the market. Proof enough that a properly designed output inductor need not limit an amplifiers distortion performance.
Cheers,
Glen
Interesting discussion.
I for one will continue to use an output inductor of about 2uH to ensure stability and isolation from capacitive loads, to say nothing of its usefullness in stopping all sorts of EMI gunk (e.g. from those CFL's we've got all over our houses now days) getting into the amp.
What I do not understand is how people fail to get a grip on the fact that the output inductor (2uH) is a small fraction of the total series inductance in a speaker system (look at it from a systems point of view).
I for one will continue to use an output inductor of about 2uH to ensure stability and isolation from capacitive loads, to say nothing of its usefullness in stopping all sorts of EMI gunk (e.g. from those CFL's we've got all over our houses now days) getting into the amp.
What I do not understand is how people fail to get a grip on the fact that the output inductor (2uH) is a small fraction of the total series inductance in a speaker system (look at it from a systems point of view).
hi Glen
we're talking about audibility here, which is not only distortion. frequency response variation can also cause audible change.
I only agree with the dimitri point on distortion because this is what exactly happen in many case.
output coil will cause frequency response variation especially at the higher end of spectrum.
as I said earlier, output coil is becoming more hard to accept as people trying to optimize specific design goal like flat output impedance, and not only distortion.
Cheers
we're talking about audibility here, which is not only distortion. frequency response variation can also cause audible change.
I only agree with the dimitri point on distortion because this is what exactly happen in many case.
output coil will cause frequency response variation especially at the higher end of spectrum.
as I said earlier, output coil is becoming more hard to accept as people trying to optimize specific design goal like flat output impedance, and not only distortion.
CheersThe inductance or the inductor?
Now it is starting to get interesting. One obvious question which has been hoovering in the air without anybody asking it is whether the (assumed) audible difference is due to the inductance or the inductor, ie. the electrical properties of an ideal inductor or the actual component, or even the components magnetic fields interacting with its environment. Interestingly we seem to get answer supporting both these.
Andy reports an audible difference from simulations!!! That clearly suggests it is the inductance that matters, or at least the inductance in combination with the complex RLC network of a (simulated) speaker. In this case it seems that audible means a hard fact considering the type of experiment Andy made.
But we also get experiences from Conrad and Dimitri that a measureble increase in distorsion seem to arise from the inductors interaction with its environment and/or the nature of the inductor (core type). I presume these measurements to be either without load or with a resistive load, so in this case it is not an interaction with the reactivity of the speaker.
So maybe we actually do have several verifiable reasons for the coil affecting the signal (whether to an audible degree is a different question).
Disclaimer. I deliberately write "(assumed) audible" in an attempt to passably please both those who say they hear a difference and those who doubt such a difference can be heard, in an attempt to avoid the usual discussion of hearing, believe hearing etc. etc.
Now it is starting to get interesting. One obvious question which has been hoovering in the air without anybody asking it is whether the (assumed) audible difference is due to the inductance or the inductor, ie. the electrical properties of an ideal inductor or the actual component, or even the components magnetic fields interacting with its environment. Interestingly we seem to get answer supporting both these.
Andy reports an audible difference from simulations!!! That clearly suggests it is the inductance that matters, or at least the inductance in combination with the complex RLC network of a (simulated) speaker. In this case it seems that audible means a hard fact considering the type of experiment Andy made.
But we also get experiences from Conrad and Dimitri that a measureble increase in distorsion seem to arise from the inductors interaction with its environment and/or the nature of the inductor (core type). I presume these measurements to be either without load or with a resistive load, so in this case it is not an interaction with the reactivity of the speaker.
So maybe we actually do have several verifiable reasons for the coil affecting the signal (whether to an audible degree is a different question).
Disclaimer. I deliberately write "(assumed) audible" in an attempt to passably please both those who say they hear a difference and those who doubt such a difference can be heard, in an attempt to avoid the usual discussion of hearing, believe hearing etc. etc.
Bonsai :
What I do not understand is how people fail to get a grip on the fact that the output inductor (2uH) is a small fraction of the total series inductance in a speaker system (look at it from a systems point of view).
now that's true but people just try to get better
now let's design amp with -1 ohm output impedance
then add a series 1 ohm resistor ( or less if considering speaker cable resistance ) , voila!!!! perfect damping ~~~~~~~ i'm just joking
What I do not understand is how people fail to get a grip on the fact that the output inductor (2uH) is a small fraction of the total series inductance in a speaker system (look at it from a systems point of view).
now that's true
but people just try to get better now let's design amp with -1 ohm output impedance
then add a series 1 ohm resistor ( or less if considering speaker cable resistance ) , voila!!!! perfect damping ~~~~~~~ i'm just joking
Hartono said:hi Glen
we're talking about audibility here, which is not only distortion. frequency response variation can also cause audible change.
Hi Hartono.
I agree that an inductor will cause small amplitude fluctuations with a reactive load at high frequencies. Though with a small inductor, these amplitude fluctuations will be very small and their audibility is questionable.
as I said earlier, output coil is becoming more hard to accept as people trying to optimize specific design goal like flat output impedance, and not only distortion.
It’s not quite as clear-cut as that, I think. As the old saying goes, there is no free lunch.
If we remove the output inductor from our amplifiers, we then have to make other design compromises to maintain an equal degree of stability when driving capacitive loads.
Without the inductor, the amplifier will need to be frequency compensated more heavily to give a greater phase margin so that the extra phase shift incurred by a capacitive load does not drive the amplifier into high frequency peaking or oscillation.
That is simple enough and practically any amplifier can be made to run stably into highly capacitive loads without an output inductor if the open loop frequency response is rolled off by the compensation components sufficiently brutally.
However, over compensating an amplifier can drastically reduce the negative feedback factor above the dominant pole frequency, and that causes the THD to rise proportionally. Less negative feedback also means higher output impedance, so some of gain made in that department by getting rid of the inductor is lost.
It is up to the designer to juggle with the design compromises that he is happy with. Personally, I feel that a small output inductor has more pro’s than con’s. I don't buy silly arguments like those put forward by some here that inductors necessarily sound bad or that amplifiers using them are only good for driving woofers.
Cheers,
Glen
- Status
- Not open for further replies.