Purifi Class D

[DuckToller]

12 minutes ago, PYP said:

From Mola Mola website's description of the Kaluga amps:  2 pairs of Furutech binding posts. Biwired directly to the amplifier PCB using Kubala-Sosna cable.

There were certainly some question about that design from sales/marketing to the designer:

like
- Wouldn't these binding post from Elecaudio sound the same like these margin shredding Furutech ?
- Who the hell is Kubala-Sosna, couldn't we take these premium Transparent cables?
... perhaps not ....
rather like
- Would  the use of some well accepted brands for name dropping, like Furutech and Kubala-Sosna have any detrimental effect on the sound quality and the measurements? Just askin' Bruno, we're thinking the product may achieve even better acceptance in the target audience we're aiming at.  Also, the break-in times look negligable.

 

[barrows]

11 minutes ago, DuckToller said:

There were certainly some question about that design from sales/marketing to the designer:

like
- Wouldn't these binding post from Elecaudio sound the same like these margin shredding Furutech ?
- Who the hell is Kubala-Sosna, couldn't we take these premium Transparent cables?
... perhaps not ....
rather like
- Would  the use of some well accepted brands for name dropping, like Furutech and Kubala-Sosna have any detrimental effect on the sound quality and the measurements? Just askin' Bruno, we're thinking the product may achieve even better acceptance in the target audience we're aiming at.  Also, the break-in times look negligable.

 

Nothing wrong with that as far as I am concerned.  Furutech binding posts are very nice, and actually made of copper.  Good quality output speaker wiring is also nice to have.  There is no doubt that the Mola Mola product line is high end, and features expensive, very well done casework as well.  Sure, without these details the products could be offered at a lower price, perhaps 50% less, but they decided they wanted to make something nice.  I do not see any harm in doing so.  After all, these are audio products, not live saving medications, no one really "needs" these!  It is up to the customer to decide if such things as a really nice chassis are worth paying for, or not.  Mola Mola does not present these products as some kind high value audio, they present them as luxury items, for a luxury market.  There are other options for those looking for the best bang for the buck.  It is not like Mola Mola set out to make the most affordable products which they could, they left that to someone else.  They also have the other end of the spectrum covered, with their classD Audio DIY kits.

[PYP]

13 minutes ago, DuckToller said:

- Would  the use of some well accepted brands for name dropping, like Furutech and Kubala-Sosna have any detrimental effect on the sound quality and the measurements? Just askin' Bruno, we're thinking the product may achieve even better acceptance in the target audience we're aiming at.  Also, the break-in times look negligable.

That marketing ploy worked for me.  My thought bubble:  I already have KS interconnects and speaker cables using the exact same wire, so one continuous run of wire from source to speakers.  

 

But to be completely honest, the work was already done when I first saw a picture of the Makua and Kalugas.  Audio lust at first sight!  These are audio jewelry to be sure (the finish cannot be adequately captured by a photo though - it is so subtle -  pant, pant 🤭 ).   

[mocenigo]

7 hours ago, DuckToller said:

@mocenigo
My friend, I referred specifically to Jud's comment about the audio engineers, that was the one you quoted.

With everything else  from your arguments I am mostly d'accord, but I suspect that some sensitive ears aren't happy with your tiredness influencing your tonality. Especially when opinions and facts are clashing, the complexity of foreign languages and different cultural backgrounds (by origin or profession) show more importance than we would usually wish for. And feeling provoked in their comfort zone.

Even most members here are career audiophiles (means they are not just starting their hobby), not all of them can deal well with the sometimes inherent acidity of objective facts. Acidity in tems of telling them that they are plain wrong because these facts are different to their experience. And then are some that can't stop putting opinion and anecdotal experience into generalisations against objective scenarios and fight the facts.

Personally, I think it is a severe put off when people starting threatening you (Yes I think the reference of MansR was exactly that!), because you serve well facts which doesn't fit their personal views, and they start fighting you about your tonality.

I personally enjoy your posts most of the times, and would love to read a kind of workshop on signal processing from you in the Blog section.

Thank you very much for your words, and I apologise for misunderstanding your post.

 

This said, everybody should stay assured that I am all but aggressive or confrontational, I am more… a very strongly argumentative type. But if one proves me wrong, well, I can change the now outdated arguments!
 

note that I used to be a subjective audiophile!

 

This said, your wish is flattering but while I understand the mathematics involved in audio design, I am not an expert, and definitely I am not qualified to teach it. 
 

[PYP]

4 minutes ago, mocenigo said:

note that I used to be a subjective audiophile!

 

what cured you?

[mocenigo]

4 hours ago, PYP said:

I've never heard two DACs sound identical.  Or two pieces of any kind of gear.  

I do not want to ask whether that was done properly blind. And even then, the test must be a proper ABX, because just yawning may change how you hear the frequencies…

[jaytor]

This hobby is extremely subjective - everyone hears things somewhat differently and has different preferences for what they like. Telling people that what they hear must be confirmation bias is extremely disrespectful and rude. Just because you'd prefer to rely on test measurements instead of your own ears doesn't mean everyone else feels that way. 

 

I would love it if I could figure out how something was going to sound just from simple measurements, but in my 50 years as an audiophile, I have not found that to be the case. Many of the products that I have tried that appear to have stellar measurements sound, at best, boring and uninvolving, and at worst, dull or fatiguing. And this includes equipment I have built myself.

 

Measurements are a very useful tool for identifying design or manufacturing problems, and I use test equipment a great deal when I am designing and building my own equipment, but I have not found them to be a substitute for listening. 

 

This thread was (is) supposed to be about the Purifi class D modules and amps built using them. If you want to start a thread to discuss your beliefs that most audiophiles are delusional and wasting their money by using their ears, by all means go ahead, but please let's get this thread back to its original topic.

[Jud]

9 hours ago, mocenigo said:

 

How do you generate frequencies that are not an integer multiple of some fundamental? Suppose you have a multi-tone IMD test, with frequencies 500*I Hz with I=1,2,3,...,32 – this is not a valid test? But even if you picked 320, 417, 911, and 17100 Hz, they are still multiples of 1 Hz. In fact you need algebraic numbers of the frequencies since with rational numbers you always have a common integral submultiple.  So at which threshold do you stop? 

 

8 hours ago, Jud said:

 

As we used to say in law, this argument "proves too much." The inharmonics of percussion instruments and the inharmonic attacks of nearly every musical source including the human voice contain frequencies that are each multiples of 1Hz, but are not integer multiples of *each other*.

 

Let's talk about why this particular discussion is really about distinctions without differences, and what this means for testing of audio equipment like a Class D amplifier.

 

First, Fourier told us long ago that any acoustic wave (any sound) can be constructed from a fundamental and harmonics.  Yet inharmonic attacks are vital to such essentials of music listening as identifying what instrument is playing (and thus, for example, following an instrumental line in a piece of music). For example, from Elliott, et al., 1965:

 

Quote

Identification of musical instruments according to their individual timbres was the focus of this study. It was theorized that the attack and release of a tone could be a factor in identifying specific instruments. For testing purposes, a two-part master tape recording was prepared--part A containing 18 randomized instrumental tones with attacks and releases spliced out, part B containing 18 unaltered, randomized tones sounded by the same instruments as in part A. A total of 57 graduate music students served as subjects. Results showed that in part A (attacks and releases removed), only three instruments--Bb clarinet, oboe, and trumpet--were correctly identified a significant number of times; in part B (unaltered tones), all instruments except the cello were correctly identified a significant number of times. For all participants, the mean score was significantly higher on part B than on part A. Thus, it was concluded that attacks and releases may well be influential factors in differentiating between and identifying specific instrumental tones.

 

So if any sound can be broken down into a fundamental and harmonics, how can we say that inharmonic attacks even exist, let alone that they are so important?  An example with easy mathematics will illustrate (even though more complex inharmonicity, like percussion or plucking a guitar string, is what we are usually concerned with):

 

Let's have a synthesizer generate a 20Hz and a 30Hz tone simultaneously.  Since 30 is not an integer multiple of 20, the two pitches will not harmonize; the result will be inharmonic.  Moreover, the two pitches will intermodulate, causing energy at their difference (10Hz) and sum (50Hz).  These new pitches will cause yet more interactions, and so forth.  But look: we can construct the very same sound by using the appropriate amount of energy at a fundamental of 10Hz, with overtones (harmonics) of 20, 30 and 50Hz, and so on.  Thus we can use a fundamental and harmonics to create the very same sound as an inharmonic series.

 

Thus @mocenigo is correct that any sound you like can be generated as a fundamental plus harmonics. And yet, @leManu must also be satisfied, because the pulses and square waves with which amplifiers are tested plainly contain inharmonics.

[mocenigo]

5 hours ago, PYP said:

From Mola Mola website's description of the Kaluga amps:  2 pairs of Furutech binding posts. Biwired directly to the amplifier PCB using Kubala-Sosna cable.

they are there for marketing purposes, the Kaluga is a luxury product and needs “luxury” component. This said, the Fururech binding posts are surely robust and reliable, and without ferromagnetic parts.

 

I may shock you, but for exactly the same reasons I use ETI binding posts in my amplifier!

[mocenigo]

14 minutes ago, Jud said:

 

 

Let's talk about why this particular discussion is really about distinctions without differences, and what this means for testing of audio equipment like a Class D amplifier.

 

First, Fourier told us long ago that any acoustic wave (any sound) can be constructed from a fundamental and harmonics.  Yet inharmonic attacks are vital to such essentials of music listening as identifying what instrument is playing (and thus, for example, following an instrumental line in a piece of music). For example, from Elliott, et al., 1965:

 

 

So if any sound can be broken down into a fundamental and harmonics, how can we say that inharmonic attacks even exist, let alone that they are so important?  An example with easy mathematics will illustrate (even though more complex inharmonicity, like percussion or plucking a guitar string, is what we are usually concerned with):

 

Let's have a synthesizer generate a 20Hz and a 30Hz tone simultaneously.  Since 30 is not an integer multiple of 20, the two pitches will not harmonize; the result will be inharmonic.  Moreover, the two pitches will intermodulate, causing energy at their difference (10Hz) and sum (50Hz).  These new pitches will cause yet more interactions, and so forth.  But look: we can construct the very same sound by using the appropriate amount of energy at a fundamental of 10Hz, with overtones (harmonics) of 20, 30 and 50Hz, and so on.  Thus we can use a fundamental and harmonics to create the very same sound as an inharmonic series.

 

Thus @mocenigo is correct that any sound you like can be generated as a fundamental plus harmonics. And yet, @leManu must also be satisfied, because the pulses and square waves with which amplifiers are tested plainly contain inharmonics.

What you say is all correct, and in fact a measurement to determine whether the system can reproduce well a 20hz and a 30hz tone simultaneously exist: multitone IMD measurement.

 

this said, it would be interesting to see how impulses and square waves behave with some amplifiers, even though the audibility of ringing is widely debated. Of course Class D amps with their more limited bandwidth are at a disadvantage here - but the ringing should not fool the savvy reader of these measurements, as it is to be expected to be more accentuated than in, say, a 1 MHz bandwidth amp like a Spectral. But it does not necessarily have a bearing on subjective sound quality.

[Jud]

1 minute ago, mocenigo said:

What you say is all correct, and in fact a measurement to determine whether the system can reproduce well a 20hz and a 30hz tone simultaneously exist: multitone IMD measurement.

 

this said, it would be interesting to see how impulses and square waves behave with some amplifiers, even though the audibility of ringing is widely debated. Of course Class D amps with their more limited bandwidth are at a disadvantage here - but the ringing should not fool the savvy reader of these measurements, as it is to be expected to be more accentuated than in, say, a 1 MHz bandwidth amp like a Spectral. But it does not necessarily have a bearing on subjective sound quality.

 

Even though I had heard and was knocked out at RMAF by @barrows' DIY amp (thought the system of which it was a part produced the best sound in show that year), when I considered purchase of my present amplifier (a Class D amp constructed along similar lines as the one barrows demo'd), I was concerned.  Here's why: The amp I owned at the time was a Spectral.  Also at RMAF I attended a seminar on distortion given by Audio Precision.  The distortion I turned out to be most sensitive to was slew rate limiting. The Spectral had an unbelievably high slew rate.  The slew rate for the Purifi-based amp I was considering wasn't as high.

 

When I took delivery of the new amp, I planned to have a long shoot-out between the Purifi-based Apollon and the Spectral, perhaps write an article about it for Audiophile Style. But that didn't happen. The Apollon trounced the Spectral, in my subjective opinion. (There was also one objective test in which it beat the Spectral: With no music playing, I turned up the volume to the speakers with each amp.  With the Spectral there was a slight but clearly audible hum from the bass and midrange drivers, and a slight but clearly audible hiss from the tweeter.  With the Apollon, dead silence.)  IIRC it's been over a year since I put the Apollon in the system, and I've never had a subjective reaction to the way the system portrayed music that was less than very pleased.

[mocenigo]

I used to be a subjectivist.
 

1 hour ago, PYP said:

what cured you?

 

 

Not 100%  sure but I have a theory. It happened roughly at the same time I abandoned religion. So getting more rational, but there is more to it.

 

Definitely building audio equipment myself helped a lot because it changed my priorities. It was for me about the fun of listening music, and I understood that in order to get more out of the music I needed to let the equipment get out of the way of my enjoyment. Aiming at something I know to be extremely neutral and true to the source, and let my ears “burn in” in the sound of the hardware (as opposed to relying on the burn in of the hardware itself, that is vastly exaggerated) was the trick. Also, I read how our hearing and its processing in the brain works / actually how it does not work, and I frankly cannot believe in claims like barrow’s (whom I do hold in great esteem).

 

So, to return on topic: I built my amp around Purifi modules (and since I bought them during a phase when Hypex power supplies were unavailable and with a several months lead time - I had to be creative, and Kim Madsen at Purifi helped me a bit)!

[Jud]

5 minutes ago, mocenigo said:

Also, I read how our hearing and its processing in the brain works / actually how it does not work, and I frankly cannot believe in claims like barrow’s (whom I do hold in great esteem).

 

I've done a bit of reading myself. A topic at which you may want to have a look is pattern recognition by humans. Acoustic patterns specifically of course, but also our other senses.  A very interesting example of blind A/B testing vs. pattern recognition can be found here: https://en.wikipedia.org/wiki/New_Coke

[fas42]

IME, class D modules not in the class of the Purifi units, in terms of the classic distortion measurements, will still do a good enough job - my current active speakers just use a chip amp, and Dutch & Dutch speakers which use the more serious Pascal units intended for pro use, but still are well behind those from Purifi as regards their performance specs, did extremely well, subjectively. By contrast, amplifiers using the Purifi modules and engineered by a designer with an objective viewpoint, heard on the same occasion, did nothing to recommend them ... the answer, as always, is that the specific implementation, and overall status of the playback system in which an amplifier is used, will dictate the qualities heard.

 

What I find remarkable is that there is continuing reluctance to consider how electrical noise factors interfere with the subjective performance, and that close to zero attempts are made to measure this - this is paramount, but in the 35 plus years I've followed this, very little progress has been made in the understanding of its significance ...

[mocenigo]

4 hours ago, fas42 said:

What I find remarkable is that there is continuing reluctance to consider how electrical noise factors interfere with the subjective performance, and that close to zero attempts are made to measure this - this is paramount, but in the 35 plus years I've followed this, very little progress has been made in the understanding of its significance ...

 

Uhm… What?

[DuckToller]

18 hours ago, PYP said:

That marketing ploy worked for me.  My thought bubble:  I already have KS interconnects and speaker cables using the exact same wire, so one continuous run of wire from source to speakers.  

 

But to be completely honest, the work was already done when I first saw a picture of the Makua and Kalugas.  Audio lust at first sight!  These are audio jewelry to be sure (the finish cannot be adequately captured by a photo though - it is so subtle -  pant, pant 🤭 ).   

Had a second look on their product matrix & found their product's naming tends a tiny bit to the subjective side
... but all good ...

[PYP]

1 hour ago, DuckToller said:

Had a second look on their product matrix & found their product's naming tends a tiny bit to the subjective side
... but all good ...

Well, they don't have a Nirvana cable yet.  Perhaps they were working on it and in the process realized the futility of physical cables.  Maybe the next step is they will give their cables away to whoever wants them.  

[mocenigo]

On 6/19/2023 at 4:03 PM, barrows said:

On the Topping, I had the same experience.  I got, I think it was the D-90, with the new AKM chip (at the time) because it was an affordable way to hear what the AKM chip could do with high rate DSD input.  The DAC was just OK.  It was quiet and detailed, but certainly nothing special.

 

What does it miss in your opinion to make it special?

 

On 6/19/2023 at 4:03 PM, barrows said:

 I tend to straddle the line between the importance of measurements and listening personally.  I do believe everything can be measured, for sure, after all, we can "look" at things in even in the quantum realm!  But the standard set of measurements, as shown say at ASR are clearly inadequate to describe all aspects of the performance of audio components.

 

I think this is possible, but esp. for DAC almost everything is there, even the measurements of the response of the filters will tell us about the behaviour of the phase, if one knows how to extrapolate that. So one could probably even reconstruct how a square wave would look like...

 

[fas42]

14 hours ago, mocenigo said:

 

Uhm… What?

 

Noise factors tend to be dumped into the grab bag called "jitter"; once it was recognised as a cause of less than stellar sound, and it became fashionable to "do something about it!", SQ did get a boost. But most of the efforts were focused on clocks, and digital links. Unfortunately, there's a lot more going on than just that waveform "jittering around" - electrical activity anywhere in the neighbourhood, and associated with the chain in some manner can wriggle its way into the vulnerable conversion and following circuitry - and degrade SQ in the same way as, "jitter". What does it sound like? Exactly the type of thing mentioned in recent posts - listless, boring, dreary, unappetising reproduction; a toned down version of what made digital take so long to get approval in higher end rigs.

 

If you eliminate all the noise factors, you get - wait for it - "analogue sound!" Which is why it's worth pursuing the goal of sorting these issues out. Currently, ultra expensive DACs use OTT engineering to get on top of this - but what's actually required is better understanding of the root causes, and decent measuring of the impacts  ...

[mocenigo]

On 6/19/2023 at 4:09 PM, audiobomber said:

It was not a threat, it was a warning. I don't have the power to ban, but the moderator does, and has several times, for exactly the same arrogant and dismissive tone that our friend here favors. I don't like being called ignorant or stupid, which is the subtext in the last couple of posts @mocenigo sent my way. 

 

Well the person I consider most ignorant is myself, since I know that I do not know most of there is, and, well, I cannot claim to know the extent of the ignorance of others. I did not make any assumption on your intelligence, and even intelligent people sometimes fall into the mistake of refuting scientific facts because they would force them to go outside of their comforting cocoon of acquired lore. I speak from experience, assuming I am intelligent. Judging audio equipment in sighted way biases the experience so much as to be almost always worthless, in the sense that it provides no actual information. We can at most determine the characters of a piece of equipment, but even attempting to determine whether "like" the sound of component A over component B is disrupted by sight and the biases, often inconscious, that seeing stuff causes in our brain.

[barrows]

17 minutes ago, mocenigo said:

What does it miss in your opinion to make it special?

Lots of DACs are quiet, and somewhat detailed.  I find very few DACs are detailed, have realistic tonality, and very good representation of space, especially both depth and ambiance.  Additionally, dynamics were lacking both on the macro and micro levels, everything was bit "grey" in terms of dynamic expression.

Now you mentioned output impedance, that, and output current headroom could be part of the issue here.  I was running amp direct, and I suspect the D-90 might have done better with a good preamp, featuring a more robust output stage, to drive the amplifier.  This is one of those areas where I typically do not agree with the technical description of what is necessary for good performance in terms of both impedance and output current capability.  For example, the D-90's output stage, by common understanding, should be entirely adequate to properly drive the input stage of the amplifier used.  But the common understanding does not allow for the sense of current headroom being something desirable.  Do not ask me why, I do not know-I am sure a smarter person than I could come up with an explanation for this, and the appropriate measurements as well.  But I have found there is often a correlation between getting good subjective performance, and some additional current headroom for the output stage driving the amplifier (whether we are talking a preamp or a DAC).  A theoretically overbuilt output stage often sounds better.  I tend to suspect that IC opamps do not perform at their best when pushing their envelope in terms of current-but I do not have the proof of this.

I often prefer DACs with a discrete output buffer, and I suspect that one of the reasons is that the discrete designs can easily be made more robust in terms of current capability.  Alternatively, there are some IC Opamps which have higher output current capability, and I have done some DIY work with these-my DIY DSC-2 became next level when I added a two stage output buffer, with a higher current IC driving the output and little output R (.1 or .2 ohms as I recall).

[fas42]

39 minutes ago, barrows said:

Lots of DACs are quiet, and somewhat detailed.  I find very few DACs are detailed, have realistic tonality, and very good representation of space, especially both depth and ambiance.  Additionally, dynamics were lacking both on the macro and micro levels, everything was bit "grey" in terms of dynamic expression.

 

Yes. In the 40 years or so of the digital era, how slowly have things progressed ...

 

39 minutes ago, barrows said:

I tend to suspect that IC opamps do not perform at their best when pushing their envelope in terms of current-but I do not have the proof of this.

 

 

The 'proof' is in the spec sheets, from the manufacturer - if one chooses to focus on key areas, like PSRR vs frequency, it becomes very clear; higher current means that the voltage rails to the amp become more spikey, because the decoupling of the supplies is not up to the job; and every opamp performs worst at higher frequencies, in just about every area, especially PSRR. Combine poor voltage stability with poorer ability to cope with noise at higher frequencies - it is quite obvious what's going on ...

[PYP]

On 6/19/2023 at 4:40 PM, Jud said:

 Also at RMAF I attended a seminar on distortion given by Audio Precision.  The distortion I turned out to be most sensitive to was slew rate limiting. The Spectral had an unbelievably high slew rate.  The slew rate for the Purifi-based amp I was considering wasn't as high.

 

Very cool @Jud  It would be so helpful if there were an online diagnostic tool that helped identify the distortion one is most sensitive to and provided a threshold measurement for that distortion.  Then the measurements provided by manufacturers would be more meaningful to more hobbyists and some of these generalized discussions on threads would be more centered on one's hearing as a determinant of what's good/best.   I'm not sure it would result in world peace, but a good start.  

[Jud]

2 hours ago, PYP said:

Very cool @Jud  It would be so helpful if there were an online diagnostic tool that helped identify the distortion one is most sensitive to and provided a threshold measurement for that distortion.  Then the measurements provided by manufacturers would be more meaningful to more hobbyists and some of these generalized discussions on threads would be more centered on one's hearing as a determinant of what's good/best.   I'm not sure it would result in world peace, but a good start.  

I think this works as far as what types of distortion one is most sensitive to, but I don't think threshold tests are useful, for two reasons:

- We're not trying to hear distortion, but music. So the threshold question is not how much distortion you hear, but how much music you don't. Think about having a conversation at a restaurant. Long before the volume level at which you can understand conversations at other tables, it becomes harder to follow conversations at your own table.

- There's a psychology experiment called the Iowa Gambling Task. Participants gamble with decks of cards. They are given "good" or "bad" decks. If they get a bad deck, they can request a good one. On average, there is increased galvanic skin response (that is, the participant literally breaks out in a cold sweat) 10 rounds of play before asking for a new deck. So we subconsciously endure unpleasant stimuli for a long time before becoming consciously aware of them. Thus we should want to know when distortion begins to reduce our enjoyment of the music on even a subconscious level, rather than when we become consciously aware of it.

[mocenigo]

9 hours ago, barrows said:

Lots of DACs are quiet, and somewhat detailed.  I find very few DACs are detailed, have realistic tonality, and very good representation of space, especially both depth and ambiance.  Additionally, dynamics were lacking both on the macro and micro levels, everything was bit "grey" in terms of dynamic expression.

Now you mentioned output impedance, that, and output current headroom could be part of the issue here.  I was running amp direct, and I suspect the D-90 might have done better with a good preamp, featuring a more robust output stage, to drive the amplifier. This is one of those areas where I typically do not agree with the technical description of what is necessary for good performance in terms of both impedance and output current capability.  For example, the D-90's output stage, by common understanding, should be entirely adequate to properly drive the input stage of the amplifier used.  But the common understanding does not allow for the sense of current headroom being something desirable.  

 

I agree 100% with you here. I noticed that there was a difference with a more suitable input buffer to the Purifis - moving from direct with the EVAL1 bypassed to the EVAL1’s buffer, and then to the Neurochrome buffer. And after that to the Pre90. While it is true that the Pre90’s input impedance is low, I have only 30cm of cable between D90 and Pre90, and they are designed to work together.

 

Each step improved significantly what you said. Again, this must be considered as anedoctal, since it is sighted comparison.

But, if you have any have those incredible measurements also at full output, then the DAC cannot have per se any loss of details, loss of micro or macro-dynamics, “greyness” so to speak. So the cause of that is how it is used.  In other words, interfaced.

 

it would be like considering a low powered amp crap because it cannot drive Scintillas. No, it is used in the wrong way, and that does not really count. However, almost nobody thinks at the issue of interfacing of components (and when the interfacing is suboptimal, they think that cables can solve it, whereas it is the bad interfacing that may make the system susceptible to the cables, but that’s another discussion).

 

What another designed told me is that those opa1612 output stages rarely work optimally if they have to drive more than a few tens of cm of cables. Composite amps and/or LME49720, OPA1656 are examples that work better.

 

 

9 hours ago, barrows said:

Do not ask me why, I do not know-I am sure a smarter person than I could come up with an explanation for this, and the appropriate measurements as well.  But I have found there is often a correlation between getting good subjective performance, and some additional current headroom for the output stage driving the amplifier (whether we are talking a preamp or a DAC).  

 

I agree with this. Essentially, for most opamps noise and distortion characteristics are measured with ideal loads, and well inside the working comfort zone. Then they ask show you the maximum operating conditions. But that’s a bit like giving you the Purifi measurements at 5W in 8ohm, and the max power at 20% distortion. And then you use the 1ET400A to drive, of all speakers, Apogee Scintillas :-)

 

I also prefer to heave headroom. 

 

 

9 hours ago, barrows said:

 

A theoretically overbuilt output stage often sounds better.  I tend to suspect that IC opamps do not perform at their best when pushing their envelope in terms of current-but I do not have the proof of this.

 

I agree on the first because it is overbuilt only theoretically. That means that a good Purifi amp costs more than the EVAL1, because it would use a different buffer, but it does not have to cost twice as much.

 

As for the current envelope, this is actually a documented fact.

 

Measurements do not lie, but they are often in ideal conditions. An AP is a different load than a speaker. These measurements will tell you the truth about the device if it is inside the comfort zone, (which is smaller than the declared operating conditions because … marketing). And that’s how I am using everything I have. 

 

 

9 hours ago, barrows said:

I often prefer DACs with a discrete output buffer, and I suspect that one of the reasons is that the discrete designs can easily be made more robust in terms of current capability.  Alternatively, there are some IC Opamps which have higher output current capability, and I have done some DIY work with these-my DIY DSC-2 became next level when I added a two stage output buffer, with a higher current IC driving the output and little output R (.1 or .2 ohms as I recall).

 

Can you tell me something about your DSC-2?