|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
108.5.225.95
In Reply to: RE: Is there any real benifit sonically between class A, AB etc... posted by morricab on June 28, 2016 at 07:10:11
Why is SET 'better'? It is certainly less linear than push/pull.
Follow Ups:
type of distortion (harmonic pattern) is suboptimal with push/pull due to even order cancellation. Additionally, most A/B and A PP tube amps still use negative feedback that further pushes the harmonic distortion pattern further away from an optimal pattern.
So, more distortion, usually, but the pattern is less optimal.
No amplifier is linear so the pattern becomes paramount with regards to audibility.
But SET amps compress the negative half of the signal relative to the positive half. They aren't even symmetrical in their amplification
My 'best' amp for me is a non feedback, triode, push/pull amp and preferably with high power(leaves out SET) to handle even massive peaks. The Berning BA150 comes close to this. If I accept feedback then the Melos amps come into the picture.
I think this is primarily problem when you are pushing close to clipping...
Nevertheless, the ear makes distortion that is monotonic in pattern, starting with 2nd order and decaying exponentially with increasing order. The relative levels of these harmonics fluctuates with SPL but the pattern is always the lowest order harmonic has the highest level (2nd higher than 3rd higher than 4th etc.). This pattern is important for masking and what the ear/brain expects to hear from natural sounds.
Cheever came up with a metric that involves both harmonic distortion level and overall SPL. That metric gives the lowest (i.e. best) score the more closely an amplifier's distortion pattern fits what he calls the "Aural harmonics", meaning the pattern that is self-generated within the ear.
This is similar to what the French journalist/amp designer Jean Hiraga was claiming in the 1980s (I think it was then) about an ideal amp has a monotonic distortion pattern.
This was also tested by audio Journalist, Keith Howard, who actually added distortion digitally to a recording with a piece of software he wrote. He found that the unaltered file always sounded the best but that the one with a monotonic even/odd pattern with exponential decay was the least offensive of the added distortions. An all odd pattern (the ideal for a push/pull amp) was the worst sounding.
That is probably why a lot of the better sounding push/pull amps don't use feedback and don't have such good even order cancellation... because the pattern is important!
Obviously, high order harmonics that are caused when an amp is in Class B and the use of negative feedback resulting in high orders is an issue as well and standout to the ear as unnatural.
The only type of circuit topology that gives a monotonic pattern is single ended. Push/pull by design results in at least some, if not total, even order harmonic cancellation...this is not a pattern found in nature and our ear brain to some degree rejects this. Probably push/pull triodes sound quite good (I used to have a nice sounding VAC 30/30) in Class A because they are linear enough to not need negative feedback and probably the even order cancellation is not absolute.
I guess most acoustic waveforms in nature are also not symmetrical...where compression and rarefaction happen with different intensities.
`This was also tested by audio Journalist, Keith Howard, who actually added distortion digitally to a recording with a piece of software he wrote. He found that the unaltered file always sounded the best`
- Yes of course !
`but that the one with a monotonic even/odd pattern with exponential decay was the least offensive of the added distortions. An all odd pattern (the ideal for a push/pull amp) was the worst sounding. `
-Realizing this few simple facts many known and unknown amp designers was starts practicing the new type of modern art ,
`The Art of the Amplifier Voicing` :)
Regards !
__
"Art which does not have the appearance of art is true art."
- Old Roman saying -
My point here was mainly that those who claim SETs and other amps with a high 2nd order harmonic are "euphonic" are not correct in this claim.
SET's are different from all others, including other tooby's , there are many good SS amps that will match or better PP Toobs, nothing matches a good SET on the right speakers, well for acoustic instruments.If one does not listen to classical music and or acoustic instruments much, then there is no need to look at SET or PP Toobs in general...
Edits: 06/30/16
As long as you keep the amp in a comfortable operating range then the music can be as complex as you like and it doesn't get confused, compressed or glassy. This means you need a sensitive speaker and/or a small room.
If you go with lower sensitivity speakers, say 90db, then a 30 watt SET will work fine for complex music at moderate levels but not realistic concert levels.
I was able to get very good sound from a 30 watt SET on 84db electrostats because the drop to the listening position was only about 1db in my room. So to get in the mid 90s was possible without driving the amp to death, which was loud enough for me to easily enjoy classcial orchestra. Part of this was due to how well the system sounded with the SOFT parts of a symphony. I could set the lower end around 60db and then peaks to 95db were no problem. Most speakers just can't handle soft parts well until you boost them to where the peaks are now driving the amp to clipping. If the speaker can handle quiet passages without losing the sound quality (very few can do this even moderately well) then you don't have peaks of 105db at home, where most systems and rooms tend to be overloaded.
Was it the NAT or Lamm you prefered on Ribbons ...?
I have never heard NAT and LAMM on the exact same set of Apogees so it is not really possible to give a direct comparison.
KR Audio is far better on STudio Grands though than the Lamm M1.1. How an all Lamm would fare I can only guess...maybe the ML1.1 is really good on ribbons. My friend now has a Kronzilla SXi on his Studio Grands and a VA350 on his Acoustats (he liked my sound so much he basically copied the system).
The Lamms drove his Scintillas while he had them...it sounded not bad but not amazing either...Scintillas with a Sphinx Project 16 was significantly better sounding, IMO.
I never heard my NAT Symbiosis SE hybrid on an Apogee but I did try it on a pair of Relco Sinus One ribbon hybrids. That sounded REALLY good! I didn't buy the speakers only because they were too tall and wide and the WAF was maybe negative :). The ribbon was 1.7 meters long and was in a kind of assymmetric waveguide. The bass was on the bottom and integrated suprisingly well and went deep enough. This speaker was everybit as good as most of the Apogees I have heard and the NAT loved it.
I have heard NAT all tube amps on a pair of Apogee Grands. Really good sound overall.
Good one Banat ...:)@Morricab , you should try that Rotel , disbeliever sounds very Serious, i mean , really now everyone knows there's no way antique toobs can deliver ....
Edits: 06/29/16
What makes you think I haven't ;-)??
I have heard some Rotel gear in the past...can't remember what it sounded like...it was that (non) memorable.
Rotel gear sound pretty good for what it is.I'll keep my DIY SET 300b amps. No holds barred LCLC B+ supply for the finals with Electra Print partial silver single secondary OPTs, separate LCLC feeding a CCS into VR tube voltage regulated power supply for the driver tubes, driver tubes CCS plate loaded......line stage is a AVC, the DIY phono pre is no holds barred LCLC B+ feeding a CCS feeding VR tube voltage regulated and the 6sl7 and 5687 are CCS plate loaded, passive RIAA EQ.
The Rotels are OK but please........
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 06/30/16
No pic ..? Sounds like a beast , will it drive a 1.5 ohm load ...
Edits: 07/01/16
Here's a picture of one output stage. The other is below it. The driver chassis is there to the left. The power supply for the driver stage is not shown.
No, they won't drive a 1.5 ohm load but then again I don't have any 1.5 ohm loads.
They drive my midranges and tweeters in a biamped system using SS to drive the JBL 2231 woofers crossed at 200Hz.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Very nice.
I have a very interesting amplifier at home right now, the Aries Cerat Diana. It is a 110Kg monster that uses the 813 DHT pentode wired in triode. The driver and input tube are both 814 DHT pentods wired in triode. All are liked up through interstage transformers that are amorphous double C cores. No coupling caps. Sound is incredibly smooth and liquid but with stunning inner resolution and texture.
The power supply is using 3 of these 3kg super caps, SS rectification and I think 6 chokes for filtering.
That looks like a very serious amp.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
I wouldn't be surprised if these theories of harmonic distribution have validity. But I believe that dynamic linearity, accurate changes in level from micro to macro may be the most significant factor in sounding 'real' and an amplifier that must compress the negative half of a waveform can hardly be called linear. Add that SET amps are very low power and peaks in the great classical music can easily top 30 dB and that adds more compression.
In deed the Berning BA150(really about 225 watts by the way) sounds best in its highest distortion(no feedback) format because in spit of added distortion that's where it's most linear.
We could go on and on. Obviously all distortion is bad. And with all the variants of amps there are various distortion combinations and to a certain degree we are now in a subjective region to a certain extent. I suspect that we differ in what we concentrate on when listening and the factors that most offend us are the ones we can't ignore and determine the amps(and other gear) we like the least. And that may be what makes this hobby fun, differences of opinion and that all gear is flawed in some way so we never reach audio nirvana which would be the end of our hobby.
"But I believe that dynamic linearity, accurate changes in level from micro to macro may be the most significant factor in sounding 'real' and an amplifier that must compress the negative half of a waveform can hardly be called linear."
While I would agree with you a zero distortion amplifier would be the ideal, no such thing exists. Therefore, you decrying this compression issue on one half of the cycle (and I think this is only an issue as you approach clipping) and claiming this is a worse problem than those generated by other topologies is not completely logical. One has to look at the outcome of what the topologies produce in terms of distortion products and how these affect human hearing.
For sure if you put a SET on a low impedance insensitive speaker you will get gross distortions that are easily audible. But if you let the amp run in its preferred power range then good ones that don't have transformer core saturation (a big problem with SET but also with many PP tube amps) will sound strikingly natural and uncolored. This is because then the distortion products are reasonably low in level but most importantly low in order and largely monotonic (most still have too many higher order components though). It certainly won't sound compressed despite what you think the scope is showing you.
I don't worry about low power as there are plenty of very good high sensitivity speakers on the market. Since there is no stereo system on the planet that I have heard that can truly do a full orchestra justice, I am not so concerned with reproudcing this kind of music at true concert level at home...it just can't be done by any system, anywhere in the world at this time...maybe someday, maybe not.
What I do care about is can it realistically do a jazz ensemble or chamber music. This is possible with a very good system and a SET with a good horn does it just about the best. Some planars with SETs, OTLs or a good push/pull triode amp can also get pretty realistic sound from this kind of music.
I am an analytical chemist, so making measurements is my living and I believe in analyzing data to reach conclusions. What the data is saying up to this point is that for what listeners generally prefer (it is always statistical...there will always be exceptions...such is the nature of humans) will be a pattern that "hides" in the ear/brain distortion and tone masking. This pattern is specific and the closer an amp gets to this pattern the more chance it has of sounding right to the largest number of listeners.
"
In deed the Berning BA150(really about 225 watts by the way) sounds best in its highest distortion(no feedback) format because in spit of added distortion that's where it's most linear."
I have yet to hear an amp actually improve when feedback was added on amps with variable feedback settings. Probably those that have a lot of feedback probably need the feedback just to work properly. In the Cheever thesis, he bypasses the feedback loop in a Hafler DH500 and gets something like 25% THD! With feedback I think it was 0.05%. So for an amp that is grossly non-linear without feedback then perhaps it at least makes the amp listenable. I would imagine a majority of Class AB SS and tube amps fall into this camp.
Funny enough, Nelson Pass shows how it is possible to make quite interesting, extremely simple designs with a few transistors that apparently sound really good. His SIT-1 and SIT-2 are the simplest possible amplifier...just 1 transistor. Very limited in power, but again high sensitivty speakers are now pletniful. I would love to hear this. Distortion is simple and acceptably low.
His white paper is interesting as he explores what happens when you add multiple active elements and then feedback to the distortion components.
Of course we cannot ignore execution of a design. This is where a lot of SETs fall short. Often there the output transformer is inadequate and saturates...this makes audible distortion of the "warm and wooly" kind. Often the driver stage is inadequate and distorts...and so on.
Same though for push/pull. That is why the VAC 30/30 is an intersting amp. It is Class A, makes around 30 watts, uses 300Bs and has variable feedback (0 to 6db). It is quite heavy thanks to the well proportioned output transformers and bass is deep, tight and controlled but well textured (no one note thud here). Mids are smooth and resolving with nice natural tone...same for the highs. However, in head to head competition with a number of SETs it is found wanting in realism...not easy to say why but the word coherence keeps coming up in conversations with friends about it ( I used to own it as an experiment, I sold it to a friend who sold it in favor of a KR Audio and then to another friend who has now relegated it to a second system in favor of 845 SET monos).
We also compared against the Lamm M1.1, which is a Class A push/pull hybrid and considered to be one of the better amps with a transistor output. That also falls short in believability.
I haven't heard the Berning amps but would love to one day.
If you believe Morricab you will spend the rest of your life throwing money away buying different tubes & amplifiers never being satisfied and not enjoying the music.
If you are satisfied with the level of fidelity you have achieved with your system then that is great for you.
I am actually satisfied right now listening to youtube and apple earbuds while at work but that doesn't mean I don't want maximum quality at home.
Morricab : I get maximum fidelity at home with my excellent system a well designed class AB amplifier with THD 0.005% @ 1 kHz , damping factor 200 at 40Hz can be considered blameless. I have been involved with many tests of speakers for comparing Live v recorded sound and the technicians involved would never consider using antique tube amps.
Edits: 06/29/16
I have no idea what you think is maximum fidelity but based on my experience to date all I can say is that I am doubtful of your claim. Very doubtful.
maximum fidelity means closest approach to live sound.
Now I am even more doubtful of your success.
"But SET amps compress the negative half of the signal relative to the positive half"
That's why using the most linear tubes in the most linear way is so important.
It's important in push pull amps also for, what amounts to, the same reason.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
but it's the topology of SET that causes the affect. Tube linearity can't fix it. Nothing can since the topology is set.
No it's the tube that causes the affect. Push pull masks it, but not entirely."Tube linearity can't fix it. "
More correctly stated "tube linearity can't be fixed". That's why it's so important to choose the most linear tube you can get.
If you load a DHT into a high impedance and don't drive it anywhere near the cutoff region or the saturation point you won't have much power but you also won't have much (if any) asymmetry.
All of the examples in the books of Class A operation show this type of operation, just using the "sweet, linear, center cut" of the dynamic curve.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 06/28/16
Sorry the asymmetry is there inherent in single ended topology. I checked with Nelson Pass who is a single ended fan and he confirmed what I knew. And in an imperfect world he did give the correct reason for using them. He likes the way they sound. I guess I just can't handle knowing that even a simple sine wave gets fouled up.
simple sine waves and all wave forms are always fouled up by amplification unless that amplification has no distortion.
I've never seen an amp without distortion.
and BTW this might just be semantics but the asymmetry is in the tube and SE has no way to hide that. Push pull hides it but not completely.
If you had a tube that could amplify in a truly symmetrical way then the output from a SE amplifier would be symmetrical. BTW a lot of preamp stages and input stages in power amps are SE, even input stages in push pull amps.
In this graph the input swing is symmetrical but the output voltage and current wave forms are not.
If a truly linear tube was used then the output wave forms would be linear (symmetrical).
A 45 or 2a3 or 300b operating only in the sweet spot in the middle (input signal limited) of the most linear operating condition (idle current, plate voltage, load impedance) are very close to being symmetrical.
So again, the tube used is what is causing the asymmetry in the output of a SE amp.
If you look at the output of each tube separately in a push pull amplifier you will see that same asymmetry.
A push pull amplifier combines the two out of phase outputs of the two tubes to form an output that looks symmetrical while at the same time creating other problems.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
This would suggest then that a single ended MOSFET or bipolar transistor would be even less symmetrical than a triode because their curves are actually far less linear than the triode.
If one wants a linear (low distortion) amplifier one has to start with linear parts.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: