What audio system do you have, or plan on getting?

[Mandryka]

@Spotted Horses When you design a solid stare amp, do you know what it will sound like before you have built it and heard it?

By the way my favourite amp is an antique tube amp, a Radford. People tell me that the thing which matters most is the quality of the output transformer.

[AnotherSpin]

The tube audio I used before was by far the best audio I ever had.

[Spotted Horses]

@Spotted Horses When you design a solid stare amp, do you know what it will sound like before you have built it and heard it?

By the way my favourite amp is an antique tube amp, a Radford. People tell me that the thing which matters most is the quality of the output transformer.

I did not build electronics for audio, but for specialized scientific projects. Most recently an amp with a bandpass filter which was part of a system to isolate a very small sinusoidal signal from a background of DC offset and broadband fluctuations. Before that a project which involved measuring extremely small currents, which involved matched, ultra low drift FETs.

Audio amplification is so basic, it is hard to imagine how to make a solid state that doesn't sound good.

[Mandryka]

Audio amplification is so basic, it is hard to imagine how to make a solid state that doesn't sound good.

But people do, though it may be partly a question of the poor quality of the components in commercial products. There's something in solid state amplifier design called "feedback" (you see how little I know about electronics!) Someone once said to me that that's where the all problems come from.

The other thing some people say - and my experience bears this out to some extent - is that in audio terms some amps are good with some speakers and not so good with others. The synergy between amp and speaker is what matters.

[AnotherSpin]

But people do, though it may be partly a question of the poor quality of the components in commercial products. There's something in solid state amplifier design called "feedback" (you see how little I know about electronics!) Someone once said to me that that's where the all problems come from.

The other thing some people say - and my experience bears this out to some extent - is that in audio terms some amps are good with some speakers and not so good with others. The synergy between amp and speaker is what matters.

It's easy enough to design a theoretically perfect-sounding amplifier, but making one in practice is a whole different story. That's why we're surrounded by so much crappy gear.

And for the record, this applies far beyond just audio equipment

[71 dB]

There's something in solid state amplifier design called "feedback" (you see how little I know about electronics!) Someone once said to me that that's where the all problems come from.

All amps have feedback, but in tube amps the amount of feedback is much smaller. Feedback is not any kind of problem when designed by a competent person. The general rule is the stronger the feedback is, the less there is distortion.

The other thing some people say - and my experience bears this out to some extent - is that in audio terms some amps are good with some speakers and not so good with others. The synergy between amp and speaker is what matters.

This is about how the output impedance plays with the speaker impedance. The higher the amp output impedance is the more the sound is coloured. Solid state amps have typically low output impedance while tube amps have high output impedance due to the output transformers. You can actually "simulate" tube amp output impedance with the solid state by putting resistors in series with the speaker cables that increase the impedance seen by the speakers and makes the speakers "think" they are driven by a tube amp. Very thin and long speaker cables have similar effect.

Speakers have their impedance peak at the resonance frequency of the (bass) driver. High amp output impedance makes it so that this resonance frequency gets attenuated less due to voltage division than other frequencies. This makes the sound more bassy. If the speakers have weak bass response, this can be even a plus, but if the speakers have good bass response, the sound becomes just bassier, rounder and less controlled (distorted).

So, for those who like the sound of tube amp + speaker "synergy", my recommendation is to use a good solid state amp and (power) resistors in series with the speaker cables. The "roundness" of the sound can be finetuned by testing different resistor values. Not round and bassy enough with 2.2 ohm resistors? Try 3.3 ohm. Resistors are MUCH cheaper than what tube rolling would cost and changing resistors in the speaker cable is easy/fast. This is what engineers do if they want "tube sound." Effective, easy and cheap, but no snake oil! 

[AnotherSpin]

[..]

So, for those who like the sound of tube amp + speaker "synergy", my recommendation is to use a good solid state amp and (power) resistors in series with the speaker cables. The "roundness" of the sound can be finetuned by testing different resistor values. Not round and bassy enough with 2.2 ohm resistors? Try 3.3 ohm. Resistors are MUCH cheaper than what tube rolling would cost and changing resistors in the speaker cable is easy/fast. This is what engineers do if they want "tube sound." Effective, easy and cheap, but no snake oil! 

Rather like when computer designers slap on a few noisy filters in a valiant attempt to make a digital snap look ever so 'analog', as if it had really been shot on film and lovingly printed on paper .

[71 dB]

Rather like when computer designers slap on a few noisy filters in a valiant attempt to make a digital snap look ever so 'analog', as if it had really been shot on film and lovingly printed on paper .

Not the same thing. The speakers can't know what causes the impedance they "see". If there's 3.451 ohms of resistance, the speakers don't know if it is caused by resistors or the output transformers. Try out my resistor hint before shooting it down. 

[AnotherSpin]

Not the same thing. The speakers can't know what causes the impedance they "see". If there's 3.451 ohms of resistance, the speakers don't know if it is caused by resistors or the output transformers. Try out my resistor hint before shooting it down. 

What I mean is that a solid-state amplifier can try to replicate the character of a tube amplifier, but achieving that perfectly is impossible, just as digital can only approximate analog without ever being identical. That doesn't mean solid-state amplifiers sound worse than tube ones. There are excellent tube designs and excellent solid-state, but they are different by nature and they do sound different.

[Valentino]

Ah! Tubes vs (AB) SS. I feel left out with my D stuff.
I sample my analog sources (vinyl and SACD) at 24/96. My speakers have digital crossovers and EQ you see. They approximate beautifully.

[Fëanor]

I did not build electronics for audio, but for specialized scientific projects. Most recently an amp with a bandpass filter which was part of a system to isolate a very small sinusoidal signal from a background of DC offset and broadband fluctuations. Before that a project which involved measuring extremely small currents, which involved matched, ultra low drift FETs.

Audio amplification is so basic, it is hard to imagine how to make a solid state that doesn't sound good.

So I've certainly heard that relative to other applications of the science, audio is indeed basic.

Nevertheless audiophiles of my acquaintance -- and my age -- will remember the solid state amps of the '80s and earlier did not sound very nice at all.  In particular they sounded harsh in the higher frequency range, this despite having very low, (for the time), total harmonic distortion (THD).  That phenomenon drove many audiophile back to tubes from solid state.

Objectively the tube amps had much higher THD, however the components of that THD were different.  Tube equipment typically is high in 2nd and/or 3rd order harmonics but relatively low in 4th and higher order harmonics.  By contrast solid-state of earlier vintage had low 2nd/3rd harmonics but relatively much higher high-order harmonics.

Consider that it's been know since antiquity, (as I recall, Pythagoras), that low order harmonics are pleasant whereas high order harmonics sound nasty.  Furthermore, IMO, (not proven science), 2nd/3rd order harmonics mask the nastiness of higher order harmonics and other distortions.  This is the underlying secret of why many audiophiles still prefer tube equipment.

The best contemporary solid-state amps, (including class D), have vanishingly low THD and what is detectable tends to be only a tiny amount of 2nd and/or 3rd order.  Hence these amps didn't suffer from the nastiness of solid-state a few decades ago. OTOH, they don't have the pleasantness of gobs of 2nd/3rd order harmonics tubes can provide.

[Spotted Horses]

So I've certainly heard that relative to other applications of the science, audio is indeed basic.

Nevertheless audiophiles of my acquaintance -- and my age -- will remember the solid state amps of the '80s and earlier did not sound very nice at all.  In particular they sounded harsh in the higher frequency range, this despite having very low, (for the time), total harmonic distortion (THD).  That phenomenon drove many audiophile back to tubes from solid state.

Objectively the tube amps had much higher THD, however the components of that THD were different.  Tube equipment typically is high in 2nd and/or 3rd order harmonics but relatively low in 4th and higher order harmonics.  By contrast solid-state of earlier vintage had low 2nd/3rd harmonics but relatively much higher high-order harmonics.

Consider that it's been know since antiquity, (as I recall, Pythagoras), that low order harmonics are pleasant whereas high order harmonics sound nasty.  Furthermore, IMO, (not proven science), 2nd/3rd order harmonics mask the nastiness of higher order harmonics and other distortions.  This is the underlying secret of why many audiophiles still prefer tube equipment.

The best contemporary solid-state amps, (including class D), have vanishingly low THD and what is detectable tends to be only a tiny amount of 2nd and/or 3rd order.  Hence these amps didn't suffer from the nastiness of solid-state a few decades ago. OTOH, they don't have the pleasantness of gobs of 2nd/3rd order harmonics tubes can provide.

That doesn't match my experience. My dad had a classic H.H. Scott 299B which was used to drive a pair of Wharfdale speakers (cabinets he made himself).



I remember it sounding pleasantly warm. The key is the Wharfdales were very efficient, with a hefty 12" woofer.

Then he got a pair of KLH Model 5 speakers.



The sound was bad, with ugly mid-range, probably because it was underpowered and the output impedance of the tube amp was a bad match to the modern speakers. He replaced with with an Akai integrated (solid state, about 60 Watts per channel) amp which was a smooth as silk. Later I got a pair of ADS L810 speakers and a NAD 2200 power amplifier.





I remember no harshness. Before I had the NAD I got a Denon integrated amp, highly regarded, but it couldn't drive the ADS speakers at all. Again, some sort of impedance incompatibility.

I doubt harmonic distortion measurements have much to do with sound quality. Harmonic distortion is typically so low as to be inaudible. The specs are relevant to the pre-amp stages. The problem is that speakers are a strange, non-resistive load (with inductance and capacitance contributing to the net impedance) and the output impedance and stability of the output stage is the most important factor. It's lovely to have a vacuum tube as an intermediate stage of amplification to give "warmth," if you're into that sort of thing. But give me a hefty power transistor at the output.

And, in the end it makes little sense to me that the custom is to have a power amplifier to drive a passive speaker. Speakers should have integrated power amplifiers that are engineered to complement the speaker design.

[Kalevala]

The contrast is interesting. Vacuum tubes are well before my time and I have never designed a vacuum tube amp. 

So, you're a youngin then? 

K

[71 dB]

What I mean is that a solid-state amplifier can try to replicate the character of a tube amplifier, but achieving that perfectly is impossible,

Why? What is the thing tubes can do but solid-states can't?

just as digital can only approximate analog without ever being identical.

What do you mean by that? What is the approximation here? Approximation because of quantisation error? Approximation because of limited bandwidth?

The truth is digital can easily be audibly distinguishable from analog, but analog struggles to sound digital (only the best analog studio tape machines approached CD quality digital sound). That's why the whole World has gone digital and analog devices are only used to add "analog flavour" to the sound. Only some audiophiles believing in he superiority of analog sound are stuck in analog, but if you ask them to explain theoretically why they think analog is superior, turns out they lack education into these things. They don't know what dither does. They don't know what Fourier Transformation or impulse responses are etc, but they KNOW engineers are wrong, because they have the super-human ears to hear things engineers can't measure. 

That doesn't mean solid-state amplifiers sound worse than tube ones. There are excellent tube designs and excellent solid-state, but they are different by nature and they do sound different.

Yes, they are different in nature, because solid-states offer cleaner sound with less distortion. One has to make solid-states to have dirtier sound with more distortion to make then similar to tube amps. 

[Fëanor]

That doesn't match my experience. My dad had a classic H.H. Scott 299B which was used to drive a pair of Wharfdale speakers (cabinets he made himself).

I remember it sounding pleasantly warm. The key is the Wharfdales were very efficient, with a hefty 12" woofer.

Then he got a pair of KLH Model 5 speakers.

The sound was bad, with ugly mid-range, probably because it was underpowered and the output impedance of the tube amp was a bad match to the modern speakers. He replaced with with an Akai integrated (solid state, about 60 Watts per channel) amp which was a smooth as silk. Later I got a pair of ADS L810 speakers and a NAD 2200 power amplifier.

I remember no harshness. Before I had the NAD I got a Denon integrated amp, highly regarded, but it couldn't drive the ADS speakers at all. Again, some sort of impedance incompatibility.

I doubt harmonic distortion measurements have much to do with sound quality. Harmonic distortion is typically so low as to be inaudible. The specs are relevant to the pre-amp stages. The problem is that speakers are a strange, non-resistive load (with inductance and capacitance contributing to the net impedance) and the output impedance and stability of the output stage is the most important factor. It's lovely to have a vacuum tube as an intermediate stage of amplification to give "warmth," if you're into that sort of thing. But give me a hefty power transistor at the output.

And, in the end it makes little sense to me that the custom is to have a power amplifier to drive a passive speaker. Speakers should have integrated power amplifiers that are engineered to complement the speaker design.

You make several points here that we may consider.

For a start, I confess that I have NO personal experience with tube power amplifiers.  However I suspect you're right about tube amps being under powered for less efficient speakers;  that category would include so-called 'acoustic suspension' speakers such as the KLH Model 5.  People using very low-powered tube amps today, such 'SET' amps of under about 12 watts invariably pair them with very high sensitivity speakers, say above 98 watts/1 volt @ 1 meter of input.

However something explains the trend in the '80s back to tube amps. Tube amps sounded different, folks said "better":  why was that?

Personally the worst sound power amp I've owned that I can compare with contemporary amps was my Phase Linear 400 power amp from the early '70s: piercingly shrill on the high-end but at the same time airless and opaque to detail.  My first amp replacing it was a solid-state NAD 220(?) from the '00s that was a huge improvement.  (In general NAD made above average sounding amps in '90s, '00, and more recently for that matter.)

I will certainly stick with my contention that 2nd/3rd order distortion is what folks low power tube amps are after, whether they admit it or not.  They often have distortion of more well above 1%, most of it been 2nd/3rd order. I have listened to many people expressing their preferences and also compared distortion spectra for hundreds of amp;  it's my conclusion that people like their tube amps because of, not despite, their high, low-order distortion.
~~~

Regarding whether power amplification ought to be internal or external to the speaker, I agree that in principle speakers with internal, purpose-designed, amplifiers ought to be the best approach.  However many speakers test well and sound great with external amps, and I don't seen external amps going away any time soon.

[StudioGuy]

But in a competently designed circuit an Op Amp would never "color" the sound.

And that ultimately is what it ALL comes down to and the fact it can be objectively demonstrated. Because this is NOT the case with so many of the audiophile marketing claims/myths, which have no reliable evidence, cannot be objectively demonstrated and typically don't even have a rational explanation.

Particularly annoying are the many false claims about digital, which are commonly predicated on analogue issues that do not apply to digital. Unlike analogue, digital has only two states (0 and 1), there cannot be a noisy or coloured 1 or 0 because there is no available state for that, for example: 1, 0, noisy 1, noisy 0 would obviously require at least 4 states and more for colourations. The only way to colour digital audio is to actually change the data.

In DACs and network streamers, the converter chip and clock circuitry largely determine resolution and coloration.

Resolution and colouration is dictated by the recording and the speakers/room acoustics that are reproducing the recording, not the DAC/Streamer or clocking. Back, long before digital audio was even released to consumers converter chip technology and digital clock circuitry was in its infancy and clocking circuitry could audibly affect resolution and colouration of the analogue conversion process. However, by the mid 1990's even $3 converter chips did not audibly colour the output and even the very cheapest converters ($50 OEM CD drives, DTV receivers, etc.) had clock circuitry over 1,000 times more accurate than any audible effect. Is the claim above really that modern DACs' clocking is thousands of times worse than cheapest DACs from 30 years ago?

On the contrary and as technology advancement dictates, today's DACs, even cheap ones, have clocking inaccuracies/artefacts so ridiculously tiny they cannot even exist as sound, let alone be audible.

As before, to the unwary/uninitiated; audiophile claims, especially those that fly in the face of historical facts are just audio equipment marketing BS unless supported by reliable, objective evidence and just repeating such false claims without reliable evidence is a serious disservice to those seeking information. Of course, as I'm not a hypocrite and I'm stating facts rather than marketing falsehoods, my assertions about chip/clocking performance can be supported with reliable, objective proof. I'll happily provide links if anyone wishes to verify.

G

[AnotherSpin]

And that ultimately is what it ALL comes down to and the fact it can be objectively demonstrated. Because this is NOT the case with so many of the audiophile marketing claims/myths, which have no reliable evidence, cannot be objectively demonstrated and typically don't even have a rational explanation.

Particularly annoying are the many false claims about digital, which are commonly predicated on analogue issues that do not apply to digital. Unlike analogue, digital has only two states (0 and 1), there cannot be a noisy or coloured 1 or 0 because there is no available state for that, for example: 1, 0, noisy 1, noisy 0 would obviously require at least 4 states and more for colourations. The only way to colour digital audio is to actually change the data.
Resolution and colouration is dictated by the recording and the speakers/room acoustics that are reproducing the recording, not the DAC/Streamer or clocking. Back, long before digital audio was even released to consumers converter chip technology and digital clock circuitry was in its infancy and clocking circuitry could audibly affect resolution and colouration of the analogue conversion process. However, by the mid 1990's even $3 converter chips did not audibly colour the output and even the very cheapest converters ($50 OEM CD drives, DTV receivers, etc.) had clock circuitry over 1,000 times more accurate than any audible effect. Is the claim above really that modern DACs' clocking is thousands of times worse than cheapest DACs from 30 years ago?

On the contrary and as technology advancement dictates, today's DACs, even cheap ones, have clocking inaccuracies/artefacts so ridiculously tiny they cannot even exist as sound, let alone be audible.

As before, to the unwary/uninitiated; audiophile claims, especially those that fly in the face of historical facts are just audio equipment marketing BS unless supported by reliable, objective evidence and just repeating such false claims without reliable evidence is a serious disservice to those seeking information. Of course, as I'm not a hypocrite and I'm stating facts rather than marketing falsehoods, my assertions about chip/clocking performance can be supported with reliable, objective proof. I'll happily provide links if anyone wishes to verify.

G

If I may, I'd appreciate it if you refrained from commenting on my posts. I have no interest in your personal grievances and won't be engaging with them.

[AnotherSpin]

I'm not entirely sure where this post belongs, but it seemed to fit under audio equipment. Let me explain why. People interested in sound are often accused of chasing expensive solutions, and I find that rather silly. Solutions can be very inexpensive and no less fascinating because of that.

A perfect example is the remarkable little device known as the Buddha Machine. It is one of those rare objects that sits somewhere between a musical instrument, a piece of art, and a pocket-sized shrine to sound.

Created in the early 2000s by the Beijing-based duo FM3, it is a small, battery-powered plastic box with a built-in speaker that plays looped ambient soundscapes. Each loop is a short fragment, ranging from drones to gentle melodic phrases, and it repeats endlessly until you switch to another. The idea is less about listening to "music" in the usual sense and more about letting sound inhabit your space, like incense for the ears.

The controls are wonderfully simple: a switch to change loops and a volume knob. The loops are fixed, which encourages you to surrender to repetition. Its inherently lo-fi speaker and tangible physical presence are part of the charm, since moving it from one corner of a room to another changes the way the sound blends with the space.

From the beginning, the Buddha Machine was embraced by experimental musicians, ambient composers, and sound artists worldwide. Brian Eno famously admired it, and reviewers described it as an anti-consumerist music player that resists the skip-happy habits of the digital age.

Some ten or more years ago, I bought several of these little boxes. They were identical except for color, made of plain, cheap-looking plastic. To be honest, I never fully managed to immerse myself in what this simple yet strangely magnetic device offered. I gave a few away to friends, while two or three remain in my home. Every so often, I come across one in a drawer or on a shelf. I pick it up, switch it on, but no sound emerges, since the batteries have long been dead.

Ultimate silence, at last in perfect accord with the teaching of the Buddha.

[drogulus]

     I looked up the Scott 299B and it uses 7189 power tubes in fixed bias mode producing 25 watts. This is supposed to be a higher power version of the EL84. A regular EL84 would top out at ~20 watts.

     

[Fëanor]

All amps have feedback, but in tube amps the amount of feedback is much smaller. Feedback is not any kind of problem when designed by a competent person. The general rule is the stronger the feedback is, the less there is distortion.
...

Nelson Pass famously argued the negative feedback, (at least more than a little), was bad because it caused high-order harmonics that sound bad.  In my observation higher order harmonics do, indeed, sound bad but using less feedback doesn't seem to follow.

Here is his famous article on the subject ... Audio distortion and feedback

I'm not an engineer but in his article Pass showed the following diagram, (attributed to John Linsley-Hood) ...



What I notice from that was that while more than say 10 dB of feedback produce higher order harmonics, but much more feedback reduce all distortion including higher order variety. That is, 4th order snuck in around 15-20 dB of feedback but ALL orders were reduced by 40 dB of feedback.

Bruno Putzeys argues "there is no such thing as too much feedback" and current class D's use a good deal more the 40 dB of feedback.  Or so to speak, while higher order harmonics may sound bad, they may be reduce by feedback to such a level as to unmeasurable and inaudible.