Windsor Hifi Show - simple systems ruled
- Thread starter bonzo75
- Start date
+1......we tend to bend numbers to fit our own narratives........best to actually listen.
I am not saying you will know exactly how the amp will sound but you can get some idea where the limitations might be based on psychoacoustics. And based on psychoacoustics it IS amplitude dependent both from the amp and the listener.
I disagree with you that the power of an amplifier is independent of power...this might apply to a few high feedback designs but low or no feedback designs change with amplitude and many high feedback ones do as well but with a decreaseing relative THD. Distortion vs. frequency is a big issue and you can hear these effects and many amps screw up piano. BTW, your saying to look at the impedance is essentially like saying look at the power. For a given signal level the current draw, and therefore the power will not be the same as with a higher impedance. So power and frequency are important factors in the distortion presentation of an amp.
I do agree with you that power supply is important and one way that it is has to do with intermodulation distortion, where the harmonics from the power supply frequency intermodulate with the harmonics of the signal. Also, if the power supply sags or is delivering variable voltage/current then the operating points of the tubes or transistors will also be dancing around and affecting distortion.
The Jadis Defy 7 has quite good bass from the time I heard it on some Focal Maestro Utopia Bes at a friend near London. There was something literally wrong with the 300B monos in Stereophile...bad design or constructed wrong its hard to tell. I haven't heard their other big push/pull amps so I cannot comment on their bass but likely the measurments would show something unpleasant if it sounds as tubey as you say. Do you think Jadis amps have poor power supplies? They look hugely overbuilt but they are likely unregulated.
First, I don't think the measurements are as simplistic as you are claiming...a lot can be gleaned from them. Second I don't think there are many truly winning combinations. There are many amps that sound pretty good or are even musical but only a very few that can do realistic even to some degree and go beyond "sounds nice".
So you don't think correlating measurments with listening is a worthwhile thing to do? At the very least for designers it would be a way to zero in on what design choices impact listeners preference is. Of course one has to be careful and select listeners who are sensitive and know what what is natural vs. unnatural sound...especially when it is subtle and not obviously unpleasant. Vladimir Lamm claims to be already doing something like this, with his own models he supposedly developed, where he claims listening trials are not really part of his design process. He has been quite successful in this approach and his products sound overall very natural.
Read Cheever's thesis on this topic, there is also quite a history of engineers who tried correlating sound quality with measured data. Most agreed that as harmonic order increases the damage it does to the sound quality increases very rapidly. This was echoed by Jean Hiraga in the audio press. Cheever goes further and shows how it is also SPL dependent because of how the ear's tolerance threshold for these things changes with SPL. Higher SPL will mask some things better but then it depends on how strained your amp is and what it is delivering at that SPL.
And so on...
Distortion isn't what screws up piano itself. It might be one in the same as an indicator of why paino doesn't sound right, but is not the problem itself. You could have 10% distortion and not know it, with perfect piano.
However, noise, that can cause any amount of destructive behaviour you can think of... amplifiers are just as susceptible as anything else. Well, technically more because the ways you can reduce noise become less effective as power climbs, but due to gain it may not be as in-your-face.
However, noise, that can cause any amount of destructive behaviour you can think of... amplifiers are just as susceptible as anything else. Well, technically more because the ways you can reduce noise become less effective as power climbs, but due to gain it may not be as in-your-face.
Even if you had only 2nd harmonic distortion at 10% you would hear it. You might like it...or not but you would hear it. If that 10% was spread over harmonics up to, say, 20th harmonic you would most likely find it unlistenable. Noise is a big problem, when it interacts with the signal and modulates the signal...and distortion. This is where power supply filtration and power line cleanliness matter. You are inherently less sensitive to distortion at higher SPLs...it is how the ear/brain mechanism work.
Bob Hovland measured them when he did some upgrades for a friend of mine. Audion is voiced for speed, not amplitude however so it could be more the circuit, which is very simple even for SET. My guess is you would really like them. They wind their own silver transformers as well.
The Tri and Melody weigh 90lbs with huge transformers and neither of them had better bass either. The Sophias had the best bass/dynamics of any 845 amp I tried (and were my favs by a mile), but they blew up on me
Bass was my original SET hangup, but now the high 2nd order distortion really bugs me. I think I'm more of a 3rd order distortion guy, which is actually why KR Audio's VA-350i has interested me in SET-land as we've discussed.
I'm going to hear my Devore Xs with LM Audio tomorrow fyi.
Yes the type of distortion will play an affect. But really it can be much higher than you think. You may want to look up listening tests and thresholds for distortion are much higher than people think. Again 10% can go totally unnoticed. You have to keep in mind during musical playback even when hearing higher distortion it may only be for small snippets of the music because typically with power, distortion varies.
Noise and distortion are not the same thing, even if they do some forms of interaction. We have THD and THD+N measurements for a reason.
Power supplies are complicated in that just being low noise does not equal good sound. The behaviour of the current will matter just as much.
Noise and distortion are not the same thing, even if they do some forms of interaction. We have THD and THD+N measurements for a reason.
Power supplies are complicated in that just being low noise does not equal good sound. The behaviour of the current will matter just as much.
Technically, 2nd-order distortion is preferable over 3rd-order. While you're not explicitly looking for high values of each of them, 3rd order distortion is especially bad.
2nd order is a full integer multiplier of harmonic content, much as from one C4 to C5 is a doubling of frequency and sounds harmonious or musical.
As for SETs and PP, I know people have been combining them to their satisfaction, letting the PP (or sometimes SS) do the bass duties and the SET tackle things above.
I have also seen an open design where someone was bent on using SET for all frequency ranges, even for bass but tuned just for this.
Long-term, I'll do multi-amping keeping my SET for its sweet spot for sure.
Well we seem to agree everything matters. However the situation is so complex and the knowledge of applied psychoacoustics is so limited that no general correlations can be drawn. In these subject formal psychoacoustics mainly studies static thresholds, and amplifier designers protect their private studies and findings from curious eyes, leaking just enough to be used for marketing purposes.
Yes, the Jadis Defy 7 has better bass that the JA80 monoblocks that has larger output transformers ... I have owned both, and still keep the DA7. But weight and watts are not a sufficient indicator of power supply quality. Just consider that the conrad johnson tube amplifiers having exclusively film capacitors in the power supply section have better bass quality than the similar amplifiers using much high higher electrolytic capacitance, with much higher stored energy.
The problem is that IMHO for each really useful formal rule of analysis we have at less as many cases as exceptions. But debating them under the umbrella of our known preferences is always interesting.
well, psychoacoutics says it differently- from Nelson Pass white paper on feedback:
http://www.firstwatt.com/pdf/art_dist_fdbk.pdf
"Many audiophiles believe that 2nd harmonic is to be preferred over 3rd harmonic. Certainly it is
simpler in character, and it is well agreed that orders higher than third are more audible and
less musical. However when given a choice between the sound of an amplifier whose
characteristic is dominantly 2nd harmonic versus 3rd harmonic, a good percentage of listeners
choose the 3rd.
I have built many examples of simple 2nd and 3rd harmonic “types” of amplifiers over the last
35 years. When I say “types” I mean that they used simple Class A circuits described as
“single-ended” versus “push-pull” and so tended to have a 2nd harmonic versus 3rd harmonic in
the character of their distortion, but were not made to deliberately distort.
Anecdotally, it appears that preferences break out roughly into a third of customers liking 2nd
harmonic types, a third liking 3rd harmonic, and the remainder liking neither or both.
Customers have also been known to change their mind over a period of time."
Bob Hovland measured them when he did some upgrades for a friend of mine. Audion is voiced for speed, not amplitude however so it could be more the circuit, which is very simple even for SET. My guess is you would really like them. They wind their own silver transformers as well.
The Tri and Melody weigh 90lbs with huge transformers and neither of them had better bass either. The Sophias had the best bass/dynamics of any 845 amp I tried (and were my favs by a mile), but they blew up on me
Bass was my original SET hangup, but now the high 2nd order distortion really bugs me. I think I'm more of a 3rd order distortion guy, which is actually why KR Audio's VA-350i has interested me in SET-land as we've discussed.
I'm going to hear my Devore Xs with LM Audio tomorrow fyi.
I don't think that second order distortion is what is bugging you about SETs. Up to about 2% it is inaudible and most SETs within their limits produce less than this. It is not just 2nd or 3rd order that is what is causing the issue with amps...that is far too simplistic. The fact of the matter is that an amplifier that has a dominant third harmonic WILL also have predominantly odd order higher harmonics. Why is that? Well, it is because the amps out there that are predominantly third/odd order harmonic amps are push/pull amps and that is because the nature of push/pull results in a cancellation of even order harmonics. So, now you have cut out half the harmonics, shouldn't that be good? Well, no it is not because you have A) cut out half of the harmonics that don't sound as bad as the odd ones (possibly the 3rd harmonic excepted) and B) You have changed the pattern to something that is totally alien to nature.
A good SET will have monotonic harmonic distortion pattern...this means 2nd is largest followed by a 3rd then 4th each decreasing in amplitude as the order increases. This pattern needs to be an exponential decay or it does not follow the pattern of the ear's own harmonic distortion pattern that aids in masking of external distortions (probably not on purpose but as a side effect).
You mentioned the KR Audio but it does not have a dominant 3rd harmonic...all KRs have a dominant 2nd harmonic with a nice exponential decay. The one that I saw from KR that had the best measurements was actually the KR VA340i, which has a 300B XLS output tube. That had a nearly ideal harmonic distribution and sounds rather wonderful. The VA350i should be quite similar because the amp design is very similar. They don't sound exactly alike because the output tubes have different characteristics and the output transformers are not going to be the same either but they are close enough to easily tell the family relationship.
Probably what you are hearing that you didn't like was the impact of having inadequate output iron. This will create harmonics right up through the midband and creates a false "warmth" that many SETs display. SETs are far more dependent on this one component than push/pull amps for good sound quality. KRs don't have this extra warmth and neither does Ayon or Aries Cerat because they all use really good output iron. If what you are saying about the Audion Black Shadow is true then it sounds like they have good iron.
So, I think you are sensitive to the distortion from poorly executed SET, that has far more going on than just 2nd order...actually all amps have far more going on than just 2nd and 3rd order...that is why when people tried to correlate with sound quality high orders have a very heavy weighting in the equations. Read Cheever and Geddes on this. Back in the 1950s D.E.L Shorter from the BBC tried come up with a correlation where he used N2/4 where N is the harmonic order. Later studies found that this was too shallow of a slope and the effect was even greater.
Given that at high volumes where 10% is likely to be generated you are less sensitive to distortion you are possibly correct but only if there is no high order harmonics at significant level. BTW, it is rising high order harmonics that give us the signal that things are getting too loud. If there is too much high order harmonic distortion in an amplifier it will make things with a lot of HF content sound louder and therefore closer since loudness of HF content is an indicator of distance to an object. This has the effect of A)highlighting transients that have high frequency content...it gets exaggerated and B) flattens images and soundstage as things don't sound as far away as they should.
However, you can add 0.1% of a 9th harmonic to say like 200Hz (so at 1800 Hz) at low to moderate levels and it would very probably be audible.
You are right that noise and distortion are not the same but they can be intermodulated...take the 50hz from a power supply for example. That is considered noise and not distortion but it can intermodulate with the fundamental and distortion harmonics. It looks like little ripples around each harmonic on an analyzer in 50Hz (or 100Hz or 150Hz) intervals. Seen this on a lot of measurements to one degree or another. Crowhurst made some interesting comments on this about 60 years ago.
The correlation between THD and sound quality is just about zero...read Geddes on this.
Geddes and CHeever both demonstrated promising correlations with measurements. Geddes with simulated amps and Cheever with real ones. Clearly static thresholds are not sufficient and Cheever addresses that. FWIW, Crowhurst talked about a lot of these problems 60 years ago. Designers who have hit on something good are wise to keep some secrets of good sound.
THe impedance of a power supply and its regulating capabilities are really important. You want very low impedance for current on demand and regulation to keep the operating points of the ative elements tied down. Interestingly, the Lamm ML2.2 uses one 6C33C to regulate the voltages on the other one. That is rarely seen in amplifiers because the regulator is just as power hungry as the Class A output stage but it gives a stability to the proceedings that make a lot of other amps sound nervous. My NAT Symbiosis, although it wasn't perfect, had this serenity of sound and it single transistor output stage was mirrored by an equivalent regulator transistor. It ran hotter than nearly every SET I have owned...Crazy stuff in that amp...
About SETs: the circuits are quite simple. If you think about the ideal theoretical amp, we usually say 'a wire with gain'. Circuit simplicity can get you very good SQ done properly compared to gear with much more complexity in the signal chain like a chain of opamps.
Regarding simple systems overall, I would agree as well: the more gear and the more complex each piece of gear, the more the EMC (Electromagnetic Compatibility) interactions and other issues there are, and the more potentially detrimental to SQ things can become. Here are just a few:
1. EMI/RFI (non audio gear into audio gear, audio gear into audio gear, especially with SMPS, digital circuits have their own inherent noise profile)
2. AC Mains pollution (non audio gear into audio gear, audio gear into audio gear, etc...), Leakage Currents (can be made even worse with some AC Filters)
3. Grounding issues (large subject, but includes chassis grounding/star-grounding, potentially 'signal grounding', Balanced vs Unbalanced, if Balanced, the Pin 1 issue)
4. Vibration issues (external into audio gear from the earth or from the air, internal to audio gear from itself, from audio gear to other audio gear through feet into the shared platform, cabinet vibrations)
There is no way a manufacturer can test EMC with all combinations of existing other audio gear or non-audio gear like your computer or external HDD for the best SQ.
They can, however, potentially reduce the issues listed for one piece of gear.
They don't all do it. The very fact that you can get better SQ with your existing gear by re-working the PSU or replacing it by a Linear Regulated one, adding a ball-and-cup arrangement for some seismic vibration isolation, replace RCA with Balanced and ensure no Pin 1 issue exists, disconnect the GND or Vbus or both connections on USB of some DACs, is evidence of that.
So, here again, simple can potentially be excellent.
Regarding simple systems overall, I would agree as well: the more gear and the more complex each piece of gear, the more the EMC (Electromagnetic Compatibility) interactions and other issues there are, and the more potentially detrimental to SQ things can become. Here are just a few:
1. EMI/RFI (non audio gear into audio gear, audio gear into audio gear, especially with SMPS, digital circuits have their own inherent noise profile)
2. AC Mains pollution (non audio gear into audio gear, audio gear into audio gear, etc...), Leakage Currents (can be made even worse with some AC Filters)
3. Grounding issues (large subject, but includes chassis grounding/star-grounding, potentially 'signal grounding', Balanced vs Unbalanced, if Balanced, the Pin 1 issue)
4. Vibration issues (external into audio gear from the earth or from the air, internal to audio gear from itself, from audio gear to other audio gear through feet into the shared platform, cabinet vibrations)
There is no way a manufacturer can test EMC with all combinations of existing other audio gear or non-audio gear like your computer or external HDD for the best SQ.
They can, however, potentially reduce the issues listed for one piece of gear.
They don't all do it. The very fact that you can get better SQ with your existing gear by re-working the PSU or replacing it by a Linear Regulated one, adding a ball-and-cup arrangement for some seismic vibration isolation, replace RCA with Balanced and ensure no Pin 1 issue exists, disconnect the GND or Vbus or both connections on USB of some DACs, is evidence of that.
So, here again, simple can potentially be excellent.
Geddes is just one good example of what some designers think about the measurements typically carried. I quote him:
"In the study that Lidia and I (Earl Geddes) did we confirmed that THD and IMD were basically useless indicators of sound quality because the measured values for either of these metrics did not correlate with the perception of music played through the system."
He suggested alternative techniques to carry measurements that he thinks correlate with the perceived sound quality and debated them in the DIY audio forum http://www.diyaudio.com/forums/solid-state/109147-geddes-distortion-measurements.html.
This 2005 Keith Howard article published in HFNRR also refers to this subject : https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=14&cad=rja&uact=8&ved=0ahUKEwinjNvjjaLQAhXHJsAKHb_6CZk4ChAWCC4wAw&url=http%3A%2F%2Fwww.gedlee.com%2Fdownloads%2FTHD_.pdf&usg=AFQjCNFGCM3Nkz5ALkMX-Z-06TxyyQFCUA&bvm=bv.138493631,d.d24
As far as I know no one pursued Cheever studies presented in his master thesis - that also presented the shortcomings of conventional techniques and his "figure of merit" parameter was never tested. Do you know of any followup of his work?
I have the Crowhurst book on my desk. He and many others have found the problems long ago, but IMHO until someone solves them measurements will be of limited value for high-end customers.
Yes, so you're sort of just going in a circle here. Most of which you've mentioned is just electronics 101, really. It's not really anything that helps someone design an amp since they know it and still have a lot of work to make something.
No I am not going in a circle. I have made it clear here and elsewhere that the content of the distortion is more important than the bulk amount.
When I talk about THD vs. frequency the increase at low or high frequency indicates more than a level change. It indicates a new mechanism for generating distortion in that range and this usually means a change in distortion content as well as level. High order content to the distortion should be avoided at all costs but you see it all the time in most amps.
