GaN amps: Peachtree or LSA Voyager?

question - in your view, what accounts for some amps having noticeably better soundstaging and imaging capabilities than others? be it solid state vs solid state, solid state vs tube, class a or ab versus class d types

most would say tube amps have the most expansive imaging, better at achieving ’holography’ and perceived image depth so to speak... while class a or ab amps can be all over the map in their spatial capabilites (some better one are quite good in this respect, while others are poor), and often, amps using class d modules really tend to flatten the soundstage and locate instruments/performers relatively poorly

curious to get your take on this...

@jjss49

My theory is the the way various amps make distortion has a lot to do with it. I used to think it was how well the amplifier was able to get phase relationships right and I still think there’s something to that, but by no means is it the only thing! One thing that seems to have a bit of correlation is the decay in harmonics as the order of the harmonic is increased. It needs to be on an exponential curve. Another thing that I feel needs more research is the role that the 2nd and 3rd harmonics play in allowing the ear to make out the detail required to reproduce the soundstage correctly. This relates to my first sentence in this paragraph. If the amplifier isn’t devoid of distortion but the 2nd or 3rd is too low, apparently that can prevent the amp sounding 3D.

IME the class of operation has little to do with how well the amp is able to winnow out soundstage depth and detail.

Frank gives a lot of credit to both the differential balanced design and the Exicon power mosfet output transistors...so hopefully Ralph can shed some light on why this might be some kind of secret sauce??

@snapsc

"Balanced Differential Design" Is a registered trademark of Atma-Sphere FWIW.

What happens in a differential circuit is even ordered harmonics are cancelled in the load of the output of that circuit. If differential throughout, cancellation occurs throughout too. This means that distortion is compounded less from stage to stage throughout the design.

This results in the 3rd harmonic being the primary distortion component which is very effective at masking higher orders in the harmonic spectra if its strong enough. Since there is less compounding of distortion in a fully differential circuit, the higher ordered harmonics decrease at a faster rate as the order is increased compared to a single-ended circuit. This means it will be smoother than the single-ended circuit, since the ear assigns tonality to all forms of distortion and the higher orders are assigned ’harsh and bright’. SET fans will comment about how smooth their amps sound, but it can get even smoother and with more detail bringing you closer to the music.

Mathematically speaking the non-linearity associated with this type of decay is called a ’cubic non-linearity’ as opposed to the quadratic non-linearity seen in a single-ended circuit. Probably more than you really needed to know; we’ve been relying on this method of controlling distortion in our amps for decades on now. Our class D does not use this method so much; it relies instead on having enough Gain Bandwidth Product that is feedback is properly supported at higher frequencies and so it able to control higher ordered harmonic generation that way instead.

Whatever the distortion signature actually is, that is the 'sonic signature' in amps that audiophiles talk about.

Cary's 805 had a variable feedback control.  I don't know how many other amps were made with this feature.

Lots. Starting in the 1950s.

Is it not possible with both Class d and Class A/B amps to engineer the amp to the lowest possible distortion and then add back enough 2nd and 3rd harmonics to give the amp the degree of musicality desired? And if so, could a "distortion" dial be added to allow the user to fine tune the sound to their liking?

Theoretically, to the former, possibly ’yes’. To the latter, likely ’no’ since such a control would have to modify some parameter of the amplifier that might be critical and not take kindly to such an adjustment.

I wouldn’t hold your breathe for such a thing. Usually designers of solid state are very conscious of the ’numbers guys’ that only look at specs and poo-poo the listening experience, even if the latter is what its all about. So they often try to get the distortion as low as they can and anything that raises it is anathema. There are a few that are careful about how their designs make distortion; Nelson Pass is an example.

I see an often quoted limit of 0.1% as being the absolute lowest than can be detected, and I think normally much higher.

@deludedaudiophile If that were primarily lower ordered harmonics that could well be true.

But otherwise that statement is false. The problem, frequently ignored, is that the ear uses the higher ordered harmonics (5th and above) to sense sound pressure. Keeping in mind of course that the ear has about a 130dB range!! -so its keenly sensitive to these harmonics! It also assigns tonality to all forms of distortion.

The higher orders can be only 0.01% or even 0.005%; if not masked by the lower orders, they will cause harshness and brightness. We've been hearing this in solid state amps for the last 60 years, so there are literally millions of references on the web!

We've known since the 1930s how sensitive the ear is to the higher orders (see Radiotron Designer's Handbook, 3rd edition). Its also very easy to demonstrate with very simple test equipment that the ear indeed uses the higher orders to sense sound pressure.

For the most part class AB solid state designs have ignored these facts in pursuit of the almighty dollar over the last 60 years. Of course, back in the 1960-1980s the semiconductors needed to overcome the problem simply didn't exist (and it seems the will to design something properly didn't exist either...). Instead, propaganda was generated to make you think you were hearing something 'neutral'. But its the Emperor's New Clothes; the simple fact is that brightness and harshness is a coloration- its not neutral.

The A/Bs I'm talking about are the Van Alstine DVA M225 monos. Surely you know Frank and his products. Care to comment on his matching GBP to feedback?

There's no way to tell from the information on the web. But I can tell you this: in conventional class AB amps the distortion numbers you need to see have to be quite low (0.000x), because higher ordered harmonics are a common thing in AB circuits. About the only way you can mask their presence since the lower orders are not available for this is for the distortion to be below -105dB! That means a lot of feedback, something quite challenging to achieve in an AB design. If the amp does not meet this criteria, its likely to have brightness.

reading up on Gain Bandwidth Product, it seems to be present with designs using the bipolar junction transistor. 

@mrdecibel to be clear, all circuits with gain have Gain Bandwidth Product. The question is does it have enough to support the feedback its designed with? If not, there will be a corner frequency where feedback will begin to decrease. This problem has been present in all amplifier designs until only very recently!

If you want the amp to sound smooth and relaxed at all frequencies, its crucial that the feedback be consistent at all frequencies, otherwise its not going to happen.

I’m very curious what good A/B brings to the party.

'Harsh and bright', IME, unless the designer has really done their homework and has enough Gain Bandwidth Product in the design to support the feedback it employs. If not, the feedback will decrease as frequency is increased and will be one of the reasons for increased distortion at higher frequencies. This in turn results in harshness and brightness (especially at higher volume levels) since the ear assigns that tonality to distortions in the upper end of the spectrum.

That is one advantage that some class D amps have over conventional AB amps.

Tweak1's point is also a good one. As a consumer with little to no background in electronics, most of the displayed specs for any piece of gear are simply beyond my understanding.

While I admire all you've done in this industry, I never needed to see measurements (nor would I understand them anyway) to validate, or not, my ability to hear whether something improves my listening experience, be they power cords/interconnects, going from rca kit to all dif balanced systems, dedicated lines, isolation/vibration devices, connection treatments, and in each of these examples  some/many are snake oil and others revelations

I'm fine with the listening experience. But when you are in it as a business having repeatable proof that your wares do what you say is a bit different!

My audio experience is quiet extensive, going back some 30 years before I was a partner in Audio Tweakers in the early 2000s.

@tweak1 I don’t have a dog in this fight but if I can offer some advice? I’ve found personally that testifying to however many years of experience I have really amounts to nothing. That isn’t how you develop cred. Also, in this particular case @kuribo has a point that isn’t being addressed by the attacks on him.

Its a simple fact that if you are a designer, you have to have measurements to confirm the authenticity of your work. I imagine a lot of people expect that our OTLs are a pretty tweaky amp, after all they have only one gain stage, are fully balanced and differential, are all-triode, class A(2) with a direct-coupled output and no feedback. That’s a mouthful and I expect that many people imagine us sculpting the sound (voicing) with certain exotic components, special wire and so on.

But we don’t do that- we rely on the measurements; if the amp fails that there is no reason to listen to it. We don’t tweak it for sound, we tweak it (if that is even the right expression) for measurable performance. More often than not, that likely involves changing a resistor value to fix an operating point of a tube...

Producing our class D amp would have been impossible without the measurements and of course the underlying math!

So for @ricevs to really address kuribo’s arguments, producing the measurable improvement in performance is how to do it. Failing that, using the argument that the ear hears more than we can measure (which is a false argument now, though probably true 30 years ago; like those that strive in so many different fields we audiophiles tend to live our lives based on tradition and what we learned long ago rather than using up-to-date knowledge) really isn’t going to wash! Why? If it sounds ’better’, why does it do so? Is it lower in higher ordered harmonic distortion? Perhaps less phase shift? Not knowing the answer is really a terrible thing!

Put another way, if you cause your hand to move and actually make the measurements that will answer that question (and one valid excuse might be that the test equipment needed to do so is a bit expensive) you should be able to sort out why it sounds ’better’, and now you have a concrete argument for the skeptics.

As an example some decades ago I found that the power cord could have a dramatic effect on one of our amplifiers. That really bothered me since I didn’t know why. Since it affect the sound of the amp that provided an access- so I measured the power and distortion based on the one variable of two alternate power cords, one ordinary one and one that was more ’exotic’ (edit: rather than measure the cords themselves...). A difference in power, distortion and also output impedance were all measurable. The difference in power was not slight either! It turns out that like all other electronic parts, power cords obey Ohm’s Law and the voltage drop across them was the culprit. Yet how often have we seen dissent over power cords on this forum alone??

I hope you see where this goes- there are a lot of ’objectivists’ that claim a power cord can’t possibly make a difference- but I can now challenge them because I’ve made the measurements and they haven’t (which is a bit ironic...)!