LSA Voyager GAN Amplifier

in fact, an amp designed by anyone to have the distortion structure that Ralph says can sound blah, OK, good, very good or great depending on the implementation of a million different factors. Ralph thinks its mainly one factor.....to each his own.

Its clear that you either didn’t read my post or didn’t understand what is says. As a result, the first statement is false. The second is too- because there is way more than one factor at play!

In order to get consisent distortion with less than 1 degree of phase shift at all frequencies in the audio passband there are basically two means to do it- either have wide enough bandwidth with no feedback or to have enough gain bandwidth product that it can support the feedback, especially at 10KHz and above.

That’s way more than just ’mainly one factor’: Its more than just two...

Look at the two distortion charts two posts above. You can clearly see a difference- the Purifi module has the gain bandwidth product to support the 70dB of feedback that module employs. There’s a lot going on there- Bruno has patents regarding the process. Its not ’mainly one factor’- there is a host of factors- poor choice of words on your part??

 

Since most polystyrene comes from Germany from the same plant.....there was practically nothing different in each custom cap.

The Germans got out of the polystyrene film business about 10-12 years ago. Its now made in China, but the Chinese film is thicker (this according to the prior owner of REL, now retired), resulting in lower performance. Termination of the foil used in the cap makes a difference- this is well documented. There probably were differences in the caps- if they were made with German film or not for starters. I’m sure if you did some measurements a correlation to their sound quality could be found. Usually you don’t have to look very hard to find it.

Many people think that because we make an all-triode class A fully differential zero feedback OTL with only a single stage of gain that the amp is really tweaky. IME they are surprised to discover that real math went into it and measurements confirm the math, and as a result the amps are really consistent and stable. I’ve been challenged by SET fans that it can’t sound right because it cancels even ordered harmonics; but since it has a cubic nonlinearity the higher orders fall off at a higher rate as the order of the harmonic is increased, allowing the 3rd (which is treated by the ear nearly the same as the 2nd harmonic, meaning its innocuous as long as its at a reasonable level) to mask the higher orders.

They are smoother and more detailed than any SET made.

As you can see, there’s a lot more to it than just ’the right distortion signature’; to get that a lot of stuff is going on.

The same is true of any amplifier. If you get the right distortion signature the kind of amp it is becomes irrelevent. There’s nothing special about any of the technologies- DHTs, pentodes, Static Induction Transistors (VFETs), GaNFETs and the like. The trick it that distortion signature. You can arrive there by tweaking, but if you are only using your ears you’re really taking a shot in the dark.

The chief engineer at HH Scott once said "If it measure good and sounds bad, it is bad. If it sounds good and measures bad, you’ve measured the wrong thing."

That statement is very much in play in class D.

history shows and I also say , nothing touches big Krell amps for bass.

This statement is false. If the amp is behaving as a voltage source then its bass won't be any different from any other amplifier which is also able to act as a voltage source. Add to that the fact that almost any solid state amp has had enough feedback to play bass properly in the last 40 years, which is why solid state has a good reputation for making bass.

'Nothing touching it' suggests that the amp has a coloration- for example **too much** bass, which seems really unlikely. So this one seems a furphy.

Wow!! Krell bass quality detractors can really make themselves look

I've not seen anyone detracting from Krell on this thread.

Pays to look a little harder, or you only see what you want to see

" Larger quantity of bass is always associated with lower quality and accuracy, as shown by the Krell."

What exactly is being said here? Its possible I misinterpreted it and if that's the case my prior statement would be false. FWIW I've not heard excess bass from any system I've heard with Krell amps. As solid state amps go, if anything I'd say they are a bit dark on top compared to the brightness of many solid state amps. And that's not a bad thing- most solid state amps are bright due to distortion.


**If** you can change a fuse in a Krell and the bass is different, I think you'll also find that the mids and highs are different too.

Every one knows big Krell’s bass is their forte.

I didn't know that, which seems to belie this statement. Other amps are known for their bass too- does that put them in the same category??

You never heard anyone use the saying "Krell like bass"

Nope. I hear about natural bass a lot. Shouldn't bass be like the music in which it exists rather than a certain brand of amplifier??

I don't mind someone saying something is "better," IF he describes what "better" means.  For you and me, "better" means closer to neutrality with more clarity.  For a tubaholic, "better" means warmer, sweeter, fuller, etc.

@viber6 For a tubeaholic, "better" means 'closer to neutrality with more clarity'. Just so you know. That might depend on the 'tubeaholic' but all the ones I know are looking for neutrality and clarity that they can't get with solid state.  This quite literally is what has kept Atma-Sphere going the last 46 years.


I mention this only so that you know that 'tubeaholics' are not so crazy as you seem to make out (I apologize if I'm putting words in your mouth).


I can identify why this is so: between 1960 and 2005 or so **all** amps employing feedback didn't have enough feedback for it to do what it is supposed to- get rid of distortion. In all cases, while suppressing distortion it also added some of its own. The distortion it adds is higher ordered harmonics and the ear perceives that as 'harsh and bright'; the 'solid state' sound. You can get a tube amp to sound like that too if you add enough feedback. In the cases of both tube and solid state the problem is insufficient loop gain; insufficient gain bandwidth product and poor phase margins (which, if exceeded, results in oscillation).


It is literally the brightness of solid state which is why tubes still exist! You don't have to know anything technical to understand this; its economics and nothing else.


Self oscillating class D amps offer a way around all this. Essentially you put so much feedback on the amp that its phase margin is exceeded and the amp oscillates. That becomes the switching frequency, and now you have a lot of feedback (more than 35dB) which then allows the amp to clean up the higher ordered harmonics generated by the feedback itself. A single formula, called the 'oscillation criteria' defines the loop gain, amount of feedback and oscillation frequency.


One thing about human hearing is that it is relatively insensitive to the lower ordered harmonics (the 2nd and 3rd in particular, which are treated in the same way).  Another thing about the ear is the masking principle. The reason solid state has a reputation of harsh and bright with 'tubaholics' is the higher ordered harmonics are not masked by the lower orders as they are in a tube amp. Tube amps actually make more higher ordered content than solid state amps but they are smoother due to the masking principle; you can see how important this is if distortion is going to be present (and it always is)!

It turns out that the non-linearities of class D can cause lower ordered harmonic distortion to occur. These mostly are issues in the encoding scheme and distortion due to deadtime. With the high amount of feedback, the THD you get is at the same levels that you get with a very low distortion solid state amp but there **can** be an essential difference: the distortion *signature* can be more like a tube amp in that it favors the lower orders. Simply due to the lowered distortion, such an amp can be both very neutral and very relaxed- just like a tube amp (until it gets overloaded). Sort of the best of both worlds- an amp with excellent transparency, relaxed, musical presentation and acts like a true voltage source all at the same time.

Shouldn’t bass be like the music in which it exists rather than a certain brand of amplifier??

Yes, it’s obvious from that answer/question, you have not sat in a jazz venue when the unamplified drum kit is being severely belted.
It’s closest in audio as I said (Krell like bass) and similar amps to it, like Gryphon etc etc etc and defiantly not OTL and most tubes with the majority of hard to drive hi-end lower impedance speakers.

Your response does not address the question (which by definition makes it a Strawman logical fallacy). And not knowing me, you'd be hard pressed to know what I've done and what I've not done. But FWIW, I play string bass and have played in jazz bands. Sheesh. I suspect you've no idea that your response here was simply so much rubbish.

You are right about OTLs except for the larger ones, which have no troubles playing lower impedances. But the ability to play a low impedance speaker is not the same as saying the amp has excellent bass.


If you're into bass, you want unmeasurable square wave tilt at 20Hz (which is something our OTLs do). If you know what that means you also know that most tube amps can't do that. Getting good bass isn't about driving low impedance so much as it is making sure that the amp has the LF bandwidth, that its low frequency bandwidth does not exceed the timing constants in the power supplies and otherwise has low distortion in the bass region. Its not about output impedance, since an amp with higher output impedance will simply have to be used with a higher impedance speaker, and that speaker can then demonstrate that the amp does indeed have 'good bass'. I do this all the time at audio shows; for a long time we had the Atma-Sphere Bass of the Year list on our website, which was a list of recordings that shook the walls very nicely.


In solid state amps that are direct coupled, since the 1970s they have had enough feedback at 100Hz that they have no problems playing good bass. Once you have that feedback several things occur. The first is that the amp can reject power supply noise as the feedback will correct for that. The second is that the distortion at 100Hz will be low (we've been seeing that for decades). At this point the only real concern is making sure that the power supplies don't sag under heavy demand and there are plenty of high end audio solid state amps that fit that bill. The feedback thing is the key here; if the amp employs it and the power supply does not sag, it will have 'Krell like bass' since the Krell has sufficient feedback at 100Hz and otherwise has good supplies. So it then follows that if it has **more** bass then its the result of a coloration and in a nutshell, that's not happening with a Krell at that frequency.


The differences you hear in solid state amps, as long as the power supplies are not an issue (and this includes class D) are going to be in the midrange and highs. Admittedly, you'll hear some class D amps that really can't do the bass right, but you'll also find in those amps that they have power supply problems (which I excluded via the caveat above). Now you might be able to get a slight improvement in the bass by improving speaker connections and other minor tweaks but that has nothing to do with the brand if those tweaks are applied across the board.

I have finally (boy I am dense some times) learned he will never hear what I have to say....so, there is no sense responding. I will correct something he says that is way, way off (but only once)......but there is no sense going over the same thing with him over and over and over again

@ricevs I've been at the same conclusion for a long time. I don't make the responses for his benefit, I do it so others reading this stuff at least have a different viewpoint rather than obvious misinformation.

Our Cherry Amplifiers use both clocked and "continuous time" digital circuits with a variable switching frequency (up to 2 MHz), allowing audio bandwidth over 100kHz.

If you're avoiding feedback you'll need that sort of bandwidth to prevent significant phase shift.

I have huge respect for your technical expertise, but sorry, if you are not a violinist or any other trained classical instrumentalist you may not realize that close exposure to real instruments teaches that these natural sounds are BRIGHT (in the good natural sense, not for example in the artificial types of distortion that is obvious if you tune the radio slightly off its freq and get the static which brightens the sound).

@viber6  I agree! I've been playing instruments since I was three years old (piano); started harpsichord when I was in 6th grade. I picked up string bass in 7th grade and played in orchestras and ensembles well after college. These days I'm in a rock band playing keyboards again only now they are synths and a Mellotron. But I also play flute and have 2 albums of that; used to regularly play out until the pandemic.


The brightness I'm referring to isn't the correct natural brightness of instruments; its caused by higher ordered harmonic distortion generated by all electronics. Your radio tuning example is a great way to illustrate how this works. The ear interprets all harmonic information as a tonality (this is how we can tell the difference between wound and gut strings for example). Traditional solid state amps and tube amps with feedback sound bright because the distortion I'm referring to isn't masked, so the ear interprets it as brightness- quite independently of the instruments being portrayed in the recording being played back. This is why two amps can measure perfectly flat on the test bench but one will be bright while the other is not. Put another way, this kind of brightness is not a frequency response error.

Like Class-D have with switching frequencies at 600-800khz, like this Icepower has https://ibb.co/5Kj0rcP that has significant feedback, and still has rubbish phase shift measurements(red trace).

I agree. In order to allow the amp to correct phase shift, the feedback has to be more than about 35dB. You either run at least that much or none at all, IMO.

what specific sonic improvements does 150K input impedance show compared to 38K, using the same resistor brand/materials?

I can answer that. The benefit may be to your preamp, especially if you have a tube preamp. The higher input impedance will mean that the preamp is making less distortion which will translate to smoother sound with more detail.

This was the correct thing to say

"Changing the input impedance from 38k to 150k will do absolutely nothing if the "active" source has a "normal low’ish" (say <150ohm) output impedance that will drive both"

This is correct. But many tube preamps have a much higher output impedance; some might be 4,000 Ohms, some are even higher!

IME there's no point going over 150kOhms input impedance on an amp; it will cause it to be more sensitive to the interconnect cable capacitance and for that matter, induced noise in the cable.

Thanks for your input on the impedance question.  I use the Rane ME 60 equalizer as a small gain stage with EQ.  This SS source has a low output impedance, I forgot, maybe 100 ohms.  The ratio of input impedance to output impedance would be 380 for 38K input, or 1500 for 150K input.  Both ratios are very high, way over the minimum 10, so I wonder whether there would be any audible or technical difference between 38K or 150K.

@viber6 With something like the Rane (which is really meant more for PA and recording use) you *might* have slightly less noise if the input impedance is lower. Otherwise the Rane itself will not interact in any way with load impedances like that. IOW no worries.

Many audiophiles have closed minds and mock pro audio equipment and concepts.

The Rane is an opamp-based circuit and as long as they are not demanding too much gain out of any particular opamp it should perform really well. Rane stuff is designed by actual engineers :)  I use one in my keyboard rig, but when I've used the mic input its proven that its very transparent.

Will these monos be a good match for the LSA 350?

Have they been serviced out- new filter capacitors in the power supplies and the like? If yes, then its certainly worth a try. If no, I would not run them at all until that's been done- your power transformers are at risk.

The objective measurements just show numbers.....they have little to do with the sound of an amp.

This statement is false. People talk about the sonic signature of an amplifer- that is its distortion signature. The problem isn't that the 'objective measurements' have little to do with the sound- they have everything to do with it- the problem is that the important measurements are rarely published anywhere.

If you want to know what is needed to create an amp that sounds like music, here are the things you need to see:

less than 1% phase shift at 20Hz and 20KHz

the same harmonic distortion at all frequencies, not just 100Hz (as is typically measured).

Either one of two things- either you have the primary distortion components be the 2nd and 3rd such that they can mask the higher orders. or the distortion is well below 100dB down (keeping in mind that this is true even at 10-15KHz). Or both.

IM distortion must be low- 0.05% or less.

If you satisfy these specs, you'll have an amp that sounds very musical. And if the amp does not, you can forecast what it will sound like. But most of this information is never published.

 

 

I make amps for active speakers. Only music lovers like them because they don’t carry the biases of separates/passive speakers do

Anyone else spot the failure of logic in this statement?

@amplifierdude If I can offer a tip: If you want credibility you're going about it the wrong way. Apocryphal statements like 'I make the best measuring and sounding amplifier' won't get you anywhere- in fact they only serve to erode your credibility. If you really want street cred you need to do two things:

1) get out there and show us how its done

2) keep doing it for a really long time so people can see you really mean it and have dedication.

Put another way, a true shaman or medicine man never says he is any such thing- others say that instead. That is why there are universities and colleges that confer degrees- others have said that the recipient of the degree has demonstrated worthiness.

 

No it will just add a bunch of noise and distortion.

This statement is false on two counts.

External crossovers don't contribute noise to any system nor do they add distortion. If either were to happen, it would point to a severe flaw in the amplifier driving the crossover.

Passive crossovers of themselves cannot cause noise or distortion. 

Well we will see next week. Do you own an audio analyzer?

Of course, but Its irrelevant.

Passive components downstream from the final output of an audio system cannot contribute noise or distortion, unless the use of timing constants to create cutoff frequencies is somehow considered distortion.

You have a 'proof' issue in that no image you can produce will be considered 'proof'. For that, you need to have an uninvolved party do the actual testing- someone who does not have a conflict of interest.

This is why I gave you that tip earlier- to continue down this path you risk eroding your credibility even further.

Here's another tip: When confronting something you disagree with on a forum site like this one, generally speaking you'll do much better if you apply this simple rule:

Attack the post, not the poster. (FWIW this is actually in the rules you agreed to when you signed up)

It does not matter if someone is personally attacking you. If you respond in kind, you get caught by the internet tar baby. No-one on the outside looking in will be able to tell who is right and they also will not care. If you want to erode good will and credibility, to be known as a troll, that is a very quick way to do it.

+1

Which model of analyzer do you own?

Hp 8903B Old but it does the job.

Clearly some 'tube guys' do have analyzers on hand...

 

 

 

I never said the amp with the lowest MEASURED distortion is the best (Halcro or whatever). What I am saying is that the amp with the lowest SONIC distortion is the best.

The distortion signature of any amplifier is the 'sonic signature' to which people often refer. Since you can't get rid of it, it needs to be correct. FWIW this is why tubes are still being made.

What is so funny is that these "science is truth" people have no truth.   There are no double blind linstening tests that confirm one measurement is better than another.

These statements are false as are some of the apocryphal claims that followed them in the same post. Dr. Earl Geddes used double blind testing to show that off axis response and total room energy were extremely valuable for a speaker to be considered musical- and thus backing the measurements. Again, the important thing here is that if you cannot correlate the measurement with the apparent listening experience, you are measuring the wrong thing.

When someone claims numbers, science, engineering, etc., doesn’t matter, that it’s a scam, a religion, all faith, especially someone whose livelihood is dependent on denying anything that might give evidence of the veracity of his claims, it is an obvious and clear fraud.

+1 Exactly! Yet oddly, those same people will take a ride on an airplane, use electric lighting, the Internet and a host of other technologies all in a single day, and all that only exist entirely due to science and engineering.

 

You know very well we are talking about electronics here......of course, certain measurements of speakers are very important. 

That isn't what your statement said. You said:

What is so funny is that these "science is truth" people have no truth.   There are no double blind linstening tests that confirm one measurement is better than another.

Saying that this applies only to amps and preamps after saying something like that is changing the argument- the first statement remains false.

That old junker maxes out at -80dB THD+N. Useless for modern audio gear engineering. Especially for class D design work. No wonder you didn’t know Johnson noise existed.

To be clear, the conclusion isn't supported by the opening statement. You don't know anything about me; the conclusion is not only unsupported but also false. And I made my statement in the context of crossover components in a loudspeaker. I would be really interested if you could show that they make even a 10th of a dB difference to the noise floor. In practice that seems difficult at best since you'd need an anechoic chamber to prove it. Regarding the analyzer, its proven itself very useful during the class D project as have differential probes, smart tweezers and a decent thermal camera plus a fair amount of simulation.

With my analyzer cheap dummy loads have a -92dB noise floor.
 

Lol! This statement suggests that your analyzer isn't being run properly or its broken. 

Would there  be harmonic distortion from a 15khz tone that humans could hear or a band limited amp would amplify?

@djones51  No. The reason that tone is used is to look for intermodulations. Those can be quite a bit lower in frequency depending on what other tones are used in the test. How well it does with this test can say a lot about how relaxed/easy going the amp is to listen to since intermodulation distortion is very audible.

The point of the 15khz tone is to see if the amp struggles with those frequencies. As most amps have a much harder time as the frequency climbs above 1khz.

I seriously doubt any class D amp would 'struggle' with a 15KHz tone. If they are at all competent, the amp would just sit there and do it all day. 

I agree with the rest of this quote though- most amps have insufficient feedback owing to insufficient gain bandwidth product to support the feedback at higher frequencies- so the distortion often rises above 1KHz (contributing to harshness and brightness). I think this might be why distortion is often measured at only 100Hz, so as to cover up that issue. I agree that the spec sheets are mostly marketing tools because they almost never tell you any of the more meaningful information, such as you pointed out above ('...THD+N vs power vs frequency over a 20-20k bandwidth'). Such a graph would instantly show who has the gain bandwidth product and who doesn't. Heaven help us if the spec sheets actually showed something useful...

 

 

Hypex and Purifi offer quite a bit of performance data on their amp products, much more than most.

They do- and are the exception rather than the rule. Bruno uses a self-oscillating technique where the phase margin of the circuit is exceeded by all the feedback added- and so goes into oscillation. The oscillation is then used as the switching frequency- and the triangle wave derived from it. So that allows for 50dB or more of feedback even at 15KHz. You can see this in the distortion vs frequency curves.

the Bel Canto e.One ref501S I now own that is equipped with Bruno's  Hypex's NC500 OEM module.

@jerryg123 Just as a FWIW; that suggests that if BelCanto (fun trivia dept.: one of their staff used to work for me...) were to use the Purifi module it would sound very similar to the amp they make now.

If you use a toroid coil instead of an incased coil it would sound different. 

Just to be clear, both are toroids.  The reason it might sound different might be because the inductor can play an enormous role in how the output section switches and if there are problems with it, can cause spurious noise that can be many multiples of the switching frequency. Its not the sort of thing that you play with and not also subject the amp to extensive testing afterwards to make sure it doesn't block out channel 4 on the TV or the like... I might add that the output filter caps have to be chosen for low inductance characteristics and this relates very specifically to the way the leads are terminated in the device. If you replace those parts with something different that has greater inductance, RF noise is the likely result. The module has to meet EU Directives which are pretty strict in this regard- and those directives serve as a model for the UK and China and to a lessor extent, the US and Canada. If you don't meet them the consequences can be dire for a smaller company!

So here is my question. If these Purfi equipped amplifiers are so great why is it I returned my NAD M33 after 45 days because it sounded awful?

@jerryg123 To my understanding the Purifi modules are not offered with an input buffer/gain stage. To the best of my knowledge, the buyer is expected to install their own. This is one of the reasons you see such variable comments on Bruno's modules- some buyers know how to design a proper input circuit and apparently many do not. I don't know anything about the NAD in particular, never heard it. But I do know that the input circuit has to be really neutral and there are some pretty bad examples out there wired into Bruno's modules. So if I were you I would not take your experience with the NAD as particularly meaningful.