What are the reasons to choose Class A/AB/H over D in this day and age?

[H-713]

In absolutely no way do I have any loyalty to Topping, nor do I own any of their products (and don't intend to), but even still I will admit that the PA5 delivers extremely good performance, beating out most of the linear amps from the last 40 years in terms of distortion.

Also, in response to the SPL discussions above, who cares? Amplifiers haven't been a limitation in this regard for over 30 years. IIRC, Crest released the 8001 in the mid to late 1980s, and by the 1990s they had things like the 9001 and 10001. Now it's even less of an issue. MC2's E45 will do 8400W bridged into 4 ohms, and Lab Gruppen has Class D amplifiers that claim to do 14 kW into 4 ohms. Oh, and the E45 is a linear topology. All of that in a 2U enclosure.

And if you need MORE power? Then AE Techron has you covered. Really, amplifier power is not the limitation and it hasn't been for decades.

If you really want to recreate 160 dB SPL in your living room (I can't imagine why you'd want to), you'll need to solve the transducer problem first.

 

[DanielT]

In absolutely no way do I have any loyalty to Topping, nor do I own any of their products (and don't intend to), but even still I will admit that the PA5 delivers extremely good performance, beating out most of the linear amps from the last 40 years in terms of distortion.

Also, in response to the SPL discussions above, who cares? Amplifiers haven't been a limitation in this regard for over 30 years. IIRC, Crest released the 8001 in the mid to late 1980s, and by the 1990s they had things like the 9001 and 10001. Now it's even less of an issue. MC2's E45 will do 8400W bridged into 4 ohms, and Lab Gruppen has Class D amplifiers that claim to do 14 kW into 4 ohms. Oh, and the E45 is a linear topology. All of that in a 2U enclosure.

And if you need MORE power? Then AE Techron has you covered. Really, amplifier power is not the limitation and it hasn't been for decades.

If you really want to recreate 160 dB SPL in your living room (I can't imagine why you'd want to), you'll need to solve the transducer problem first.

It looks quite ok, PA5. As a desktop amplifier of good price and/vs performance. US $349.
Distortion 65 dB at 15 kHz, 50 W, 8 Ohm. If you think it's ok. A matter of taste.
PA5, too little power for me to have in the main system.However, here and now, but especially in the future, it will belong to class D. I think so.

Speaking of power, levels of distortion and price. Plus the low damping factor:

Air Tight ATM-300R
$ 14 650

Specifications
Power Output: 9W ＋ 9W (THD <10%)
THD below: 1% (1kHz / 1W / 8Ω)
Input Sensitivity: 290mV (9W)
Damping Factor: 7 (1kHz / 1W / 8Ω)
Frequency Response: 30Hz ～ 40kHz (−1dB / 1W)
Dimensions: 430 (W) × 275 (D) × 245 (H) mm
Weight: 24.5kg

https://www.akkelisaudio.com/forstarkare/air-tight/airtight-atm-300r.html

A little more power in this tube amp
Nice tubes it was with. A little more expensive than the above though.

Erik Signature 70​

2x30 W and 8 ohms.​

$ 24 750

Produkter | Knispel Audio | Sverige

Erik Signature 70 - Integrated, transformer coupled, single ended amplifier with the russian GM70 directly heated triode. Built in Sweden by Knispel Audio.

www.knispelaudio.com

But of course, it was nice with its tubes so people have different motives when they make nice choice.

 

[Lambda]

I don't know if that's correct. The loudest rock concert in history (or at least in the Guinness Book of World Records) is a 1976 Who show that was clocked at 120dB.

It is not.
Even if this measured values are always integrated over some time rms Averages and not true Peak!
They get also:

Weighting curve - Wikipedia

en.wikipedia.org

But whatever the SPL of a cannon, the audience is not listening to it at anywhere near the 1 meter standard distance for SPL ratings. Instead, the cannon is fired at a safe distance from the audience so that it will be very loud and impactful, but not painful and potentially deafening. So the 1-meter SPL equivalent of that cannon fire would almost certainly be less than 120dB.

For any standard way of measuring SPL it might be under 120dBA (i don think so)
But the true peak is would be way over 120dB SPL
Nothing rates true peak in SPL measurements because this is not how human ears work

The ear operates as an energy detector that samples the amount of energy present within a certain time frame. A certain amount of energy is needed within a time frame to reach the threshold. This can be done by using a higher intensity for less time or by using a lower intensity for more time.

We measure and hear sound "Energy" and the energy is what causes damage. not the pressure
(well of cause pressure can also cause damage but at way higher levels)

To record and reproduce a 120dBA SPL peak on a SPL meter you would need way more then a 120dB peak system

 

[skyfly]

I do not think Class D amps are unrepairable. I guess that, due to low price, replacing the whole board or whole unit costs less than having a service technician replace a couple of parts on the board.

 

[DanielT]

Then there is the question. Just because it's new, it does not automatically have to be good. Something that was most quickly mentioned in this thread:

Rega IO:

Rega IO Review (stereo amplifier)

This is a review and detailed measurements of the Rega IO stereo integrated amplifier with phono input. It is on kind loan from a member and costs US $649 on Amazon including Prime shipping. It is designed and built in UK and price there seems much more reasonable (379 pounds). The front...

audiosciencereview.com

Vs

Yamaha CR-600

Then add features and which one you think looks best. Best fit in the listening room, living room. What can the choice be then?

Yamaha CR-600 AM/FM Stereo Receiver Manual | HiFi Engine

www.hifiengine.com

 

[pjug]

I do not think Class D amps are unrepairable. I guess that, due to low price, replacing the whole board or whole unit costs less than having a service technician replace a couple of parts on the board.

Interesting that Sonos has announced they will be making an effort to improve repairability. Probably repairs by Sonos would be swapping whole circuit boards rather than component level repair, but at least it is something. Maybe 3rd party repair services can happen too?

Sonos Pledges to Make More Energy-Efficient, Repairable Products

Sonos reveals design changes that may make it easier to repair and maintain its products.

www.pcmag.com

 

[tvrgeek]

One can look at simple specs ( useful for weeding out the trash) and then one must use their ears. I have yet to hear a class D that I like compared to the better class A/B, including my own MOSFET design. Now, there are some really poor A/B amps out there too so do not make generalities of this.

That said, I have not had time to give a listen to March, Lyngdorf, or even the new NAD and compare it to the better, but reasonable priced A/B amps. There is a new player I just heard about: Stark. Can they beat a Benchmark or even a little Vidar? Maybe.

Remember, there are still those out there who can only be happy with tubes. You can't say they are wrong as it is their ear, not yours. Personally, I find I can do better with A/B and am not a tube fan. Been there, tried it. But again, a Lux or Mac in 1960 could stomp on most early SS amps.

Now, for all but the front three channels in a HT, I too see no reason not to use class D.

 

[tvrgeek]

Interesting that Sonos has announced they will be making an effort to improve repairability. Probably repairs by Sonos would be swapping whole circuit boards rather than component level repair, but at least it is something. Maybe 3rd party repair services can happen too?

Sonos Pledges to Make More Energy-Efficient, Repairable Products

Sonos reveals design changes that may make it easier to repair and maintain its products.

www.pcmag.com

Component repair is unreliable and highly likely to give birth to a future failure. Wisdom from my years in manufacturing quality/failure analysis. It is also usually labor cost prohibitive.

 

[beren777]

I'm not worried about distortion at 15kHz having a practical impact on my enjoyment when I can't hear anything above 10kHz. While my hearing range might be below average for gender and age, most people over 40 are going to be in the same boat. If the distortion profile at 15kHz chases teenagers out of the room, that could be considered a feature, not a bug. It really all comes back to practical versus absolute performance. Here at ASR, we chase absolute performance because we believe that it will be reflected in practical performance. (At least, that's my interpretation of what ASR chases and why, along with showing which products are most cost effective in achieving neutral performance.)

I also believe that tubes tend to appeal to a couple of different crowds: those who grew up with them, those with significant hearing loss, and those who are easily suggestible. For the hearing loss crowd, I have a theory that I'm unable to test, that the harmonics help to increase the perceived signal by replacing what the impaired listener can't hear anymore.

There's nothing wrong with wanting to tailor sound to one's liking. I personally just think that's more easily done in DSP these days versus mixing and matching gear with different distortion profiles. I'd like the gear to be neutral and if I want to mangle the reproduction, I'll do that in software.

 

[tvrgeek]

Even though I am an old curmudgeon, I can still hear a little above 15K. But do be aware, upper harmonics and worse, driver breakup, causes IM in the driver which manifests itself as distortion down in the audible range, below 8K. So plain old harmonic distortion in the lower frequencies can excite breakup modes in the tweeter which in turn generates even more distortion in the critical range. Bummer, but I can't change the laws of physics.

Now, there are several ways to deal with that. The first is obviously better behaved tweeters with lower distortion. Then proper crossover design which rolls off above 18 or so. A Zobel will do this as a side benefit. So, you are reducing the tendency of the driver ( tweeter in this case) from miss-behaving. If your tweeters and crossover are not up to par, then an amplifier topology using dominant pole compensation may be an advantage over ones using Miller or Transient Miller compensation. I do not know how tweeters behave with class D yet. One should also consider distortion of the mid-bass as it shows up in the critical range. A high pass filter does wonders to keep the woofer within it's happy range. ( I found about 1/2 X-max as a rule of thumb). If you have a 2 way sub crossover, great as you get that function for free. Crossing a tweeter too low is also problematic.

I harp on speakers as we are dealing with distortions in the .1 to several %. We often forget we are dealing with a system, not a string of components. Will I hear the difference between an Atom Dac+ and my Asgard? I kind of doubt it, but I am going to find out. I could hear the difference between my old Muse and the Schiit. I still have a slight edge on some music I do not think should be there. I may also swap amps again. The headphone amp in the Schiit is way better than my old DIY or Nak on my Grados, but not so much on my old "friendly" Yamaha cans. Yea, yea, if I listened to cans much, I should look into some designed in this century.

The absolute measured parameters at ASR are a great benefit to me. I do not buy solely on numbers as having been around, I know they do not tell the full story, let alone the biases you mention. But they sure can identify the garbage! Besides the information the numbers give, they also tell you something about the company when the advertised capabilities are far short of measured. I might mention, as far as power amps go, the ones I have heard with vanishingly low distortion numbers have sounded lifeless and just plain dull. Class D I have heard, ( old generations and limited) all sounded edgy to me even the basic numbers suggest otherwise. Ones with a bit worse numbers can, but not always, sound far more energetic. I have not had the pleasure to own a Benchmark with both superb numbers as well as a musical reputation or the newer well executed class D. Could be, as the cost goes up, you can get both low numbers and life.

So, love ASR numbers based reviews. Back in the day, I understood most of the glossy rag reviewers taste and bias so I could use their hyperbola reviews as a secondary filter before I tracked down something to hear, but not any more. Only my own ears as secondary now. Bummer we have so few high end shops any more so the chance to hear without buying is getting harder and harder. One hint is the best of the lot should not sound. I heard a Rotel integrated compared to an Arcam in the same room and there was a dramatic difference. I heard a Moon, Atoll, and Hegel in another room and could not tell them apart. The Arcam may have been good too as I would characterize the Rotel as typical no-power supply sound even though it was far larger on paper than the Arcam.

 

[DanielT]

I'm not worried about distortion at 15kHz having a practical impact on my enjoyment when I can't hear anything above 10kHz. While my hearing range might be below average for gender and age, most people over 40 are going to be in the same boat. If the distortion profile at 15kHz chases teenagers out of the room, that could be considered a feature, not a bug. It really all comes back to practical versus absolute performance. Here at ASR, we chase absolute performance because we believe that it will be reflected in practical performance. (At least, that's my interpretation of what ASR chases and why, along with showing which products are most cost effective in achieving neutral performance.)

I also believe that tubes tend to appeal to a couple of different crowds: those who grew up with them, those with significant hearing loss, and those who are easily suggestible. For the hearing loss crowd, I have a theory that I'm unable to test, that the harmonics help to increase the perceived signal by replacing what the impaired listener can't hear anymore.

There's nothing wrong with wanting to tailor sound to one's liking. I personally just think that's more easily done in DSP these days versus mixing and matching gear with different distortion profiles. I'd like the gear to be neutral and if I want to mangle the reproduction, I'll do that in software.

LOL.

So it is with aging and our hearing. We'll all go that way.

Maybe it's our creator who's trying to say something to us. That is, now that you are getting older, you will return to being a bass head as you were in your youth. The finesse with the highest notes is over.

In with subwoffers ...

.... then the neighbors come and knock on the door because then they have something to say ...

 

[DanielT]

I harp on speakers as we are dealing with distortions in the .1 to several %. We often forget we are dealing with a system, not a string of components. Will I hear the difference between an Atom Dac+ and my Asgard? I kind of doubt it, but I am going to find out. I could hear the difference between my old Muse and the Schiit. I still have a slight edge on some music I do not think should be there. I may also swap amps again. The headphone amp in the Schiit is way better than my old DIY or Nak on my Grados, but not so much on my old "friendly" Yamaha cans. Yea, yea, if I listened to cans much, I should look into some designed in this century.

The absolute measured parameters at ASR are a great benefit to me. I do not buy solely on numbers as having been around, I know they do not tell the full story, let alone the biases you mention. But they sure can identify the garbage! Besides the information the numbers give, they also tell you something about the company when the advertised capabilities are far short of measured. I might mention, as far as power amps go, the ones I have heard with vanishingly low distortion numbers have sounded lifeless and just plain dull. Class D I have heard, ( old generations and limited) all sounded edgy to me even the basic numbers suggest otherwise. Ones with a bit worse numbers can, but not always, sound far more energetic. I have not had the pleasure to own a Benchmark with both superb numbers as well as a musical reputation or the newer well executed class D. Could be, as the cost goes up, you can get both low numbers and life.

So, love ASR numbers based reviews. Back in the day, I understood most of the glossy rag reviewers taste and bias so I could use their hyperbola reviews as a secondary filter before I tracked down something to hear, but not any more. Only my own ears as secondary now. Bummer we have so few high end shops any more so the chance to hear without buying is getting harder and harder. One hint is the best of the lot should not sound. I heard a Rotel integrated compared to an Arcam in the same room and there was a dramatic difference. I heard a Moon, Atoll, and Hegel in another room and could not tell them apart. The Arcam may have been good too as I would characterize the Rotel as typical no-power supply sound even though it was far larger on paper than the Arcam.

Here is a compilation of the different distortions of the speaker element./drivers. A whole lot of them are used in various commercial speakers, if I remember correctly. Not me who made that compilation. Based on data from here:

HiFiCompass |

hificompass.com

faktiskt.io • Visa tråd - 3: Distorsion, jämförelser mellan element

Edit:
To reduce distortion, relieve individual speaker element / driver, there are those who advocate line speakers. See picture (although preferably they should go from floor to ceiling, plus adding tweets is probably not so stupid, even though there is a broadband element on those in the picture).

 

[MakeMineVinyl]

Component repair is unreliable and highly likely to give birth to a future failure. Wisdom from my years in manufacturing quality/failure analysis. It is also usually labor cost prohibitive.

Not to mention that every nCore module I've seen uses large amounts of adhesive (RTV?) between almost all the components, capacitors especially, which makes extracting a single component for repair a very difficult and very risky task. I know, I've tried, and I would put myself at the expert level of doing these kinds of repairs. I wouldn't doubt that component level repairs on these class D modules would be limited to die-hard hobbyists with considerable spare time and good soldering skills (and who also know how to make PCB trace and plated through via repairs!), but for commercial operations, the time required for component level repairs is not worth the money vs a simple board swap. I would seriously doubt that even Hypex repairs defective boards which have been returned to them.

 

[tvrgeek]

Here is a compilation of the different distortions of the speaker element./drivers. A whole lot of them are used in various commercial speakers, if I remember correctly. Not me who made that compilation. Based on data from here:

HiFiCompass |

hificompass.com

faktiskt.io • Visa tråd - 3: Distorsion, jämförelser mellan element

Edit:
To reduce distortion, relieve individual speaker element / driver, there are those who advocate line speakers. See picture (although preferably they should go from floor to ceiling, plus adding tweets is probably not so stupid, even though there is a broadband element on those in the picture).

Only part true. Multiple drivers help in that the excursion of each is lower keeping distortion lower but only to an extent. But if you have measured drivers, you know thy actually have a pretty broad range of excursion where the distortion remains quite low. More is not necessary better. One must also evaluate independently the breakup modes and the levels necessary to excite them from the primary excursion bases, suspension and motor based distortion. You have to measure each one and evaluate the use.

Line arrays have their place, especially in PA use, but they also have issues. I base some of my conclusions on 45 years of building, measuring and listening. It is quite clear many commercial speaker designers do not, or even read the literature, AES etc. or we would not have MTM center speakers, sharp edges and totally inappropriate crossovers. I have which cut the number of mistakes I have made down considerably.

Looking over that list of drivers, I would not agree with the author. For instance I have never found a Morel tweeter that did not sound like total crap. But the lowly XT-25 if used correctly is superb. ( most are not) I would like to hear the Purify mid-bass to see if what they fixed makes it overall better than the Seas and SB drivers I typically use or did they fix one set of problems ang get/leave another set? Always looking for a good mid. Not found one. There are good tweeters, but most are very difficult to use. I wanted to try the Transducer Labs but looks like they are not actually restarting production.

The biggest issue with board repair is ESD. I was able to change our corporate policy to scrap all repaired boards, new off line or field returns. Now, we did failure analysis on almost every one and often it was repaired for confirmation or test, but they never left the factory.

PS: Never been a bass freak. I prefer natural music. Not that it does not take some serious effort for Copland. As far as hearing loss, our evolution (be it natural or designed I will leave out) does not expect us to live much past reproduction age. Not very efficient. We hear to find food and to not be found as food. FWIW, when I was in my 20's, I could hear 22K. Tested.

 

[DanielT]

Only part true. Multiple drivers help in that the excursion of each is lower keeping distortion lower but only to an extent. But if you have measured drivers, you know thy actually have a pretty broad range of excursion where the distortion remains quite low. More is not necessary better. One must also evaluate independently the breakup modes and the levels necessary to excite them from the primary excursion bases, suspension and motor based distortion. You have to measure each one and evaluate the use.

Line arrays have their place, especially in PA use, but they also have issues. I base some of my conclusions on 45 years of building, measuring and listening. It is quite clear many commercial speaker designers do not, or even read the literature, AES etc. or we would not have MTM center speakers, sharp edges and totally inappropriate crossovers. I have which cut the number of mistakes I have made down considerably.

Looking over that list of drivers, I would not agree with the author. For instance I have never found a Morel tweeter that did not sound like total crap. But the lowly XT-25 if used correctly is superb. ( most are not) I would like to hear the Purify mid-bass to see if what they fixed makes it overall better than the Seas and SB drivers I typically use or did they fix one set of problems ang get/leave another set? Always looking for a good mid. Not found one. There are good tweeters, but most are very difficult to use. I wanted to try the Transducer Labs but looks like they are not actually restarting production.

The biggest issue with board repair is ESD. I was able to change our corporate policy to scrap all repaired boards, new off line or field returns. Now, we did failure analysis on almost every one and often it was repaired for confirmation or test, but they never left the factory.

PS: Never been a bass freak. I prefer natural music. Not that it does not take some serious effort for Copland. As far as hearing loss, our evolution (be it natural or designed I will leave out) does not expect us to live much past reproduction age. Not very efficient. We hear to find food and to not be found as food. FWIW, when I was in my 20's, I could hear 22K. Tested.

I thought, perhaps a little naively, that on the one hand it seems smart with line speakers. If they go from floor to ceiling. This with floor / ceiling reflectors. Plus good, high sensitivity, on the speakers. But then after studying it a little more closely, it turned out that it was not so easy to achieve anyway. What you are talking about and then with how sound is spread at different frequencies.

Then they would fit in a combined listening living room, so I put those thoughts down.

This speaker element is in the line speakers in the picture in my previous post:

Peerless TC9FD18-08 | HiFiCompass

The THD measurements at voltage level 2.83 Volt were carried out with High Pass Filter "on" (2nd order Butterworth type, 80 Hz cutoff, HPF2-80) The THD measurements at voltage level 4 - 11.2 Volt were carried out with High Pass Filter "on" (2nd order Butterworth type, 150 Hz cutoff, HPF2-150)...

hificompass.com

Here DIY, Array:

Murphy's Corner Line Array project

After having read many forum threads here and elsewhere I have decided to move forward with a Murphy's Corner Line Array project in my living room I thought that I would share my progress here in this thread as regularly as possible The design idea: The original MCLA design has a 45 degree...

audiosciencereview.com

 

[DanielT]

I do not think Class D amps are unrepairable. I guess that, due to low price, replacing the whole board or whole unit costs less than having a service technician replace a couple of parts on the board.

Does not only apply to Class D amps and Hifi. There is a lot of new where it has become more tricky. I was going to change the headlight bulbs in my sister's car a while ago. Had to unscrew the bumper, ... huh ?! Riduculous. Just to replace two lamps. Useless.
She has a fairly new car. And Opel Corsa.

 

[sgent]

Does not only apply to Class D amps and Hifi. There is a lot of new where it has become more tricky. I was going to change the headlight bulbs in my sister's car a while ago. Had to unscrew the bumper, ... huh ?! Riduculous. Just to replace two lamps. Useless.
She has a fairly new car. And Opel Corsa.

Sometimes items can be manufactured in a way to make them more repairable -- but optimizing for this increases engineering time / parts and raises costs. That means either a margin squeeze or an increase in price which the consumer electronics industry has said NO to for decades.

In addition, doing anything more than board-level replacement requires skilled workers which mostly don't exist, and would charge too much if they did.

 

[DanielT]

Sometimes items can be manufactured in a way to make them more repairable -- but optimizing for this increases engineering time / parts and raises costs. That means either a margin squeeze or an increase in price which the consumer electronics industry has said NO to for decades.

In addition, doing anything more than board-level replacement requires skilled workers which mostly don't exist, and would charge too much if they did.

Sure, you may be absolutely right, but even with that knowledge in the back of my mind, it would not have helped me there standing with two light bulbs staring like a fool at the car and scratching my head.

Ok it was OT. Back to the topic of the thread.

 

[H-713]

As much as I am a proponent of Right to Repair, the argument that class D amplifiers are "less repairable" than class AB amplifiers is rather weak in my mind.

Given a schematic, I do not expect that I would have any more difficulty troubleshooting a class D amplifier than a class AB amplifier. The concepts aren't difficult to understand. Without a schematic, it's about the same level of crapshoot as a class AB amp.

As for the physical rework itself... clearly some of you haven't worked on very many class AB amps. BSS EPC780. Crest CA series. Carver M400 cube amplifiers. Countless others. All of these multi-board amps that you can't service without complete and total disassembly? The single-board with surface mount construction is a walk in the park. I rebuilt a CA9 earlier this year. In order to properly troubleshoot it I wound up having to take everything out of the chassis and assemble it on the bench. PITA.

 

[Galliardist]

Not to mention that every nCore module I've seen uses large amounts of adhesive (RTV?) between almost all the components, capacitors especially, which makes extracting a single component for repair a very difficult and very risky task. I know, I've tried, and I would put myself at the expert level of doing these kinds of repairs. I wouldn't doubt that component level repairs on these class D modules would be limited to die-hard hobbyists with considerable spare time and good soldering skills (and who also know how to make PCB trace and plated through via repairs!), but for commercial operations, the time required for component level repairs is not worth the money vs a simple board swap. I would seriously doubt that even Hypex repairs defective boards which have been returned to them.

With boards full of components that run hot, if you get a repair with a "refurbished" board you can end up paying full price for a repair that only lasts a short time, so this might not be so bad.
The next level is recycling, and I'm left wondering if boards covered with adhesive will even recycle well at the level of crushing and reclaiming metals and minerals.