The jury's out still on the best term for it. Ripple is normally used to indicate stuff from rectification of mains (100/120Hz and harmonics) however I'm not meaning that, I'm meaning the hash put on the rails by virtue of the current going into the speaker. Agreed that CLC (or CLCLC) is about the best way to reduce mains ripple to insignificance. This hash is particularly bad news for SQ when it gets into the audio because its got very strong correlation with the audio itself so the perceptual processing of the brain does not recognize it as benign noise (and hence disregard it) rather it gets integrated into what we hear.
Stereophile | January 2017 Issue
- Thread starter NorthStar
- Start date
The jury's out still on the best term for it. Ripple is normally used to indicate stuff from rectification of mains (100/120Hz and harmonics) however I'm not meaning that, I'm meaning the hash put on the rails by virtue of the current going into the speaker. Agreed that CLC (or CLCLC) is about the best way to reduce mains ripple to insignificance. This hash is particularly bad news for SQ when it gets into the audio because its got very strong correlation with the audio itself so the perceptual processing of the brain does not recognize it as benign noise (and hence disregard it) rather it gets integrated into what we hear.
Well when we were talking about puny power supply caps, I assumed ripple. HF hash, now that is bad stuff.
Brad- thought I would get back to you on this in light of recent visit to Disney Hall to see Helene Grimaud play Brahms Concert No. 2 (the one on DG as well, which made for a nice comparison later). I used to be a more frequent visitor to the symphony, but only once or twice a year these days.
When I listen to SET amps, my mind feels absorbed into spooky midrange (coined by Kara Chaffee once over the phone) especially with my ears closed and its an unsettling, not relaxing experience. It feels like instruments are eerily coming into my space (if that makes any kind of sense). I tend to listen around 75-80dbs, so not particularly quiet.
Frankly, live performance wasn't similar. My seat was 10 rows up Orchestra East, with a very nice view of Grimaud as a bonus. I do think tubes are more like the tonality of live performance (some SS is as well, but much more rare). There is a cello solo at the opening of the 3rd movement that was spectacular- warm and rich, but not so rich like some SETs. Decays that most SET amps have are too long compared to real string and piano movements.
I also think most Hifi is bright and going to Disney Hall always reconfirms that conclusion - although there are many 'philes afraid of any highs (read: Audio Note guys). But never did I feel like I was swimming in distortion on this evening. I did pull out my phono spl meter and peaks were 100-105dbs.
I'm sorry I can't be more scientific here, but just relaying what I heard.
ps. Helene was phenomenal - well deserved, triple standing O
First, I would like to say thanks for the thoughtful response.
The next question I have though is which SET amps do you have long term experience with and how much of your experience is short show experience? Just like there are bad and not so bad SS amps and bad and not so bad PP tube amp there are bad as well as good sounding SETs. I would not make a blanket statement that all SETs are good...I have heard a fair number that are mediocre at best. However, their sins tend to be less obviously aggressive...borin perhaps but not so aggressive as other topologies done wrong can easily produce. I have not heard the designs from Kara Chaffee but based on the measurements I have seen from them she is using too small core on her output transformers and there is far too much bass distortion. Bass distortion has impact right up through the whole midrange and gives that overly saturated sound. Interestingly, OTLs do not have this overly saturated sound despite using triodes (or pentodes wired as such)...in some cases they sound even somewhat lean.
http://www.soundstagenetwork.com/measurements/dehavilland_gm70/
The better SETs have sufficiently large transformer cores that the increase in distortion is very small in the bass and this keeps the whole range much cleaner. The best I have heard so far in this regard is from Aries Cerat, KR Audio, Lamm and Ayon. All use large cores (thus very heavy amps) and there is audibly far less oversaturation of tone but still maintaining very natural tone. NAT is also quite good in this regard, although a bit more saturated than the ones mentioned above but nowehere near a DeHavilland model.
http://www.stereophile.com/content/...er-amplifier-measurements#hP6o5DoU0htiljmq.97
The other thing to note is that the core saturation of the output transformer is dependent on the power output, so when you use a speaker with high sensitivity and use typically < 1 watt then you are not entering into saturation nearly as often and the sound will be less "saturated" up through the range. Only the superior designs I mentioned will also work well with normal speakers and not turn to this kind of character of being too colorful.
Now, to the performance itself. You were not sitting very close and this will make tone colors a bit less vibrant. You have to remember that recordings are not made with the microphones in row 10...therefore, many recordings will be more vibrant and dynamic than the actual live event unless you sit close. A cello solo that is warm and rich from row 10 will be very vibrant and saturated when you sit up close (as I discussed in my thread a few months ago). I have a lot of experience with violin up close...and not just any violins but Stradivarius and Guarneri del Gesu among other top instruments. I can tell you a Strad (at least the one my ex had in her possession for a few months) up close is mind blowing...it is so dynamic and colorful...and loud! The other great violins we had in house were a good 3-6db quieter I would say and generally darker in character (particularly the Guarneri). Most chamber music is recorded very close and during playback it should sound like that as well...very close and projected. If you have sat close to a string quatet you will realize that the sound comes out to you.
Most hifi is bright and so are the recordings (due to the closeness of the microphones capturing significantly more highs than the middle to back rows in a hall). Why is most hifi bright? It has nothing to do with FR response but has to do with HF distortions that alter the balance of perception regarding loudness and also perception of distance.
You will find that there are designers out there that are tailoring their gear to create a more "distant" perspective, perhaps to counterbalance the overly close perspective of most recordings. You see this with a lot of speakers that have a dip in in the presence region that moves images back in the soundfield. A rolled off high frequency response (or restricted dynamics in the highs) will also enhance a sense of space and depth and move images back from the listener.
Other gear that has been well designed but is very clean in the highs will mimic the recordings presence and put it right in your lap if it is recorded this way BUT it will not have the aggressive brightness and flattened perspectives that become mentally fatiguing rather quickly.
If every recordings sounds a bit distant then this is just as wrong as if every recording sounds right in your face. Really good live recordings "breathe", meaning when the sounds are soft they stay close to the instrument and make the image smaller and a bit more distant as it gets louder the bubble expands and comes towards the listener and gets larger as well. Very little gear out there can do this effect although you can hear it live rather easily. Imaging should be 3d and palpable and soundstage should be both in front and behind the plane of the speakers. An intimate recording should give a "they are here" effect in your room and a more distantly recorded piece should give you the "I am there" effect. Most systems will give one or the other full time.
"Decays that most SET amps have are too long compared to real string and piano movements. "
This is contrary to my experience with sitting close to a lot of live practice and concerts. Most amps using negative feedback will have far to abrupt stops in the decay where real instruments keep vibrating for far longer. A SET stops decaying when the signal itself stops. An amp with negative feedback (a few SETs use a couple db as well) will truncate this decay because of the signal modulated noise floor obscurring the end of that decay. Based on what I have heard I absolutely cannot agree with you on this statement.
Hi Morricab,
A bit late to the party but I fully agree with your last paragraph.
I've used a few SE amps in the past, first were Unison Research Smart845 on Impulse Lali which gave a very pleasant sound with a large soundstage. I believe these use 6dB negative feedback so not the most dynamic of amps but it did convince me to pursue the SET route.
I changed my speakers to Avantgarde Duos some 14 Yrs ago and found the Smart845 a little too noisy for my liking. This lead to purchasing my first SET using the 45 tube, a Welborne Labs DRD45. A stunnng, at the time, little amp that could drive the Duos to pretty loud levels and remain controlled on all but the most bombastic music. I sold them to fund the latest amp from Ron Welborne, (don't ask about his dishonesty, it's well known), the Starchief. This amp was demoed against a Yamamoto A08 which uses a similar topology and 45 tube. It was a dead heat. I improved on the basic Starchief, changing caps, sockets and hookup wiring and after trying many of the respected NOS 45 tubes decided the EML45 solid plate globe tubes gave me the sound I was looking for.
Fast forward a few years and I built a direct coupled SET amp once again using the EML45 and following work done by Jeff Medwin and Dennis Fraker. A lot of design decisions go against normal tube design such as; non critical chokes, low value caps, multiple bypass caps, and low awg internal wiring. Also PSSS outputs from Electraprint wound for low IMD, oversized power transformers for low dcr and caps and resistors chosen for sound quality. Direct coupling is the key, only 150mm of, in this case, Mundorf Ag/Au wire, between the driver and output tube.
This first amp was so far ahead of the modded Starchief it was laughable and now my second amp uses similar topology though now using a EML45 V4 tube for even less noise.
So I agree that a SET will reproduce even the smallest sounds made by an instrument or voice and not stop until there is silence.
It does help if your speakers are 107dB (omega drivers in the Duos) and 12ohm.
The music is very present with huge soundstage and fills the room with a tactile quality.
Sorry, not sure that was all on topic but you did ask about the amps.
Cheers
Blue58
A bit late to the party but I fully agree with your last paragraph.
I've used a few SE amps in the past, first were Unison Research Smart845 on Impulse Lali which gave a very pleasant sound with a large soundstage. I believe these use 6dB negative feedback so not the most dynamic of amps but it did convince me to pursue the SET route.
I changed my speakers to Avantgarde Duos some 14 Yrs ago and found the Smart845 a little too noisy for my liking. This lead to purchasing my first SET using the 45 tube, a Welborne Labs DRD45. A stunnng, at the time, little amp that could drive the Duos to pretty loud levels and remain controlled on all but the most bombastic music. I sold them to fund the latest amp from Ron Welborne, (don't ask about his dishonesty, it's well known), the Starchief. This amp was demoed against a Yamamoto A08 which uses a similar topology and 45 tube. It was a dead heat. I improved on the basic Starchief, changing caps, sockets and hookup wiring and after trying many of the respected NOS 45 tubes decided the EML45 solid plate globe tubes gave me the sound I was looking for.
Fast forward a few years and I built a direct coupled SET amp once again using the EML45 and following work done by Jeff Medwin and Dennis Fraker. A lot of design decisions go against normal tube design such as; non critical chokes, low value caps, multiple bypass caps, and low awg internal wiring. Also PSSS outputs from Electraprint wound for low IMD, oversized power transformers for low dcr and caps and resistors chosen for sound quality. Direct coupling is the key, only 150mm of, in this case, Mundorf Ag/Au wire, between the driver and output tube.
This first amp was so far ahead of the modded Starchief it was laughable and now my second amp uses similar topology though now using a EML45 V4 tube for even less noise.
So I agree that a SET will reproduce even the smallest sounds made by an instrument or voice and not stop until there is silence.
It does help if your speakers are 107dB (omega drivers in the Duos) and 12ohm.
The music is very present with huge soundstage and fills the room with a tactile quality.
Sorry, not sure that was all on topic but you did ask about the amps.
Cheers
Blue58
Interesting. Can you tell me more about these designs from Jeff Medwin and Denis Fraker? So, they basically advocate a SMALL power supply and direct coupling? What about output transformers?
Wonderful stuff, morricab. This is more or less my experience with live music, location of the seating, and various recordings. I highlighted two sentences with which I have a slightly different opinion or question for you. I have found that the "they are here" effect versus the "I am there" effect, at least in my system, is also partly the result of the effect of my room on the overall sound of the system. As I have improved my room acoustics, I can also get an "I am there" impression from intimate recordings. I guess I should ask, do you think the difference between these two effects has anything to do with how well one's room disappears, or do you think it has everything to do with the recording technique and scale of the music?
False generalization. Case in point: just last weekend I had the pleasure to listen to Ack's system with Spectral amps, where decays were spectacularly realistic. Spectral amps use quite high levels of feedback, as far as I know, but with thoughtful and sophisticated engineering of superfast circuits that appear to avoid the usual timing errors and non-linearities introduced by extensive feedback.
I agree with you though that feedback which is badly or suboptimally implemented, as is often the case, will give you truncated decay.
Yes, small power supply so as a closer connection to the main power, ie. no huge caps to charge up and rob life from the music.
Output transformers I use are Electraprint PSSS, they have partial silver on the secondary plus Jack Elliano has a way of winding them that may be unique, don't ask me what it is I don't know. I think Magnequest also have a relatively special OT too.
Denis Fraker is the man behind Serious Stereo amps and Jeff Medwin 'notorious' on Audio Asylum for his unconventional thinking. Best not to enquire. Some measure, some listen. You can throw Morgan Jones book in the bin.
I believe Audiopax use a similar power supply. Caps are chosen for sound quality, ie best 4uf is Dynamicap, also used in Aesthetix products. Charcroft, Caddock and Shinkoh resistors plus caps from Vcap, Audyn, Mundorf, Clarity, Jupiter and Dynamicaps. These amps cost a fortune to build but imo worth every penny.
Direct coupling means nothing gets between the driver (triode) and final (triode). Remember the final has to be run at its sweet spot and not run to maximise power.
I run my 12ohm Duos from 8ohm tap to reduce distortion at the expense of a bit of power.
Some pics on my sleepy blog. http://smart845.blogspot.co.uk
Brad, I agree with Peter, these thoughts about performance and reproduction are excellent.
Thanks- you could have a point on seating position vs. tonality. That said, I'm not sure which is more correct - I mean, how often do we sit 5' from a player of any instrument and is that the correct soundstage to judge a home audio system?
I was recently at jeffrey_t home this weekend as well - he runs Lamm M2s on a full Lamm front end, but has a nice bit of room/ceiling treatment that gets the instrument to leap out very much closer to the listener. Tonality was very nice. So its not just a SET thing. Agree, there are speakers that tend to have a much more distant soundstage. But by bright, I'm talking B&Ws, Monitor Audio, old Thiels, Revels, etc...and anything with a rising upper midrange (Magico Q).
As far as recent SETs - I had the DeHavilland 845s on demo years ago. I wasn't enamored with the bass in my large room, but they weren't the only victims back then. If fact my 15 watt Quad Jubilees had much better bass control. Most recently I've heard the Melody AN845, Line Magnetic 219i, and the Audion Black Shadows in SET-land in my smaller room. I would honestly be curious how the Kronzilla sounds in my system as I believe with SS input stage, they likely have less 2nd harmonic which might work much better for my ears. That said, its aesthetics with 2' tall tubes don't work for my house.
The one thing still remarkable about live concert piano is the keystroke dynamics and how I could hear Helene in lightning fast motion up and down the instrument with very distinct motions - it wasn't really about leading edge either. This really isn't possible in a home setting and shows how large the difference still is between live and hifi. That said, SET decay tends to smooth over these transitions and is distracting to me. Very well could see string players feel differently (my long-time SET friend has played guitar for 40 years) - I grew up playing clarinet with a horn player sibling.
fwiw - I have basically only owned without negative feedback for over a decade (BAT, Quads, Valvets, Dartzeel, Shindo, SETs, etc).
Well, the room can certainly damage imaging precision as well as soundstage depth but it has to be pretty bad to really ruin it from my experience. However, I prefer to use speakers with relatively high directivity (planars and horns for example) rather than the typical wide dispersion direct radiator types. That being said, soundstage information is definitely on the recordings (whether real or artificially added) and I have heard far too many different pieces of gear over the years either have wonderful soundstage or destroy soundstage when replaced within a system that has demonstrated it can "breathe". I personally find electronics to the overall most damaging to sonic realism and therefore the limiting factor in most systems. I also think for analog that the phonostage is the most important part of an analog chain. The transition between transducer to electrical and vice versa is where a lot of bad things can happen.
I have not heard a full Spectral system but the bits and pieces I have heard were not as convincing as you are saying. Timing errors are not really the main problem with feedback but injection of back EMF into the amplification signal is potentially a big one. The formation of harmonics of harmonics of harmonics is another (artificial "noise" floor). Maybe Spectral has solved it, I cannot really comment on their designs until I hear them in more controlled conditions.
I am not sure that they are on to the primary reason for what they are hearing. I think it might have more to do with their choices giving a lower impedance to the power supply and this allows faster delivery of current and faster recovery. However, if one were to have a big capacity power supply with the same "speed" then it would probably work better for large transients and dynamic contrasts. Still, an interesting thought experiment...until someone can really test it.
Yes, IMO Spectral has solved the problem, I hope you get to hear the amps sometime in a system that allows you to hear this. I found the sound of the current Spectral top gear to be breathtakingly natural (and I think I know what to listen for, given that I have quite a bit of experience with live music, and triode amps with no feedback at all).
Thanks Al and Peter, just gettin' the thoughts out and trying to make sense of my observations.
I know you and Ack chat how Spectral is tubes without the hassle, but I have never shared that feeling - in fact, Spectral used to sound more Halcro-ish to my ears. Soulution is a modern-day contemporary to how I used to view them. Then again, there is no Spectral dealer in LA and I haven't heard a newer setup (been 10 years). I would definitely like to hear the new Spectral gear in a familiar setting based on yours and Marty's comments.
My other issue with Spectral is they are never at shows, never reviewed, and require cables that are very expensive to run. I'm not opposed to the system approach they use, but its not for me.
Yes, not too often we get the "front row seat". I was doing a lot of home recording of solo violin, violin and piano, violin and cello and string quartet so a lot of exposure to "up close". I was trying to make recordings though from a somewhat more distant perspective (2-4 meters from the performers rather than inside the violin
) for a more realistic live sound...sometimes it worked sometimes not.Funny enough, B&W speakers are often engineered with a "Gundry dip" in the presence region, so even though they sound a bit bright they also sound a bit recessed at the same time (I think they just sound wrong but that is my opinion).
What were your thoughts on the Melody, LM and Audion amps? You might be surprised to learn that KR will likely have MORE 2nd harmonic than these others relative to higher harmonics. KR distortion is monotonic and at a normal power usage, say 100mW-5 watts, the harmonics drop off very rapidly from the measurements I have seen. However, this behavior is consistent in the bass as well, something that might not be true for the two Chinese amps and the Audion, which can color the whole spectrum (you will easily get harmonics to 20th and beyond with core saturation). KR also has low power supply noise and very low intermodulation with the power supply. 2nd harmonic is really not the problem as it will be below audibility for 99% of your listening...it is the other harmonics that are the problem and their relationship to power delivery...and overall SPL. If you don't like a Kronzilla aesthetically then try a VA350,which has an 845 sized tube and a pretty similar character.
I get the power from keystrokes that are pretty lifelike in my system and they are not blurred over with my Ayon Crossfire III or KRs I have had at home. They were a bit smoothed with the Wall Audio Opus M50s but not enough to distract. You still hear the decay from the note before, as you would with a real live piano, but the new keystroke is still very clear. It doesn't sound sharp it has umph and weight as well as attack. Most systems either get the attack or the umph but not both. I don't hear smoothing of transitions...I hear a complex string vibrating and a hammer strike in two distinctly different soundspaces (the string vibration that sustains the longest without the pianist actively damping them are the low strings and they have a complex harmonic behavior that sustains right into the noise floor and extends between the speaker whereas the hammer falls are more directional and coming from one side, where the hammers are physically located). Do you expect the vibration from the previous note to stop when the hammers strike the strings again? Only if it is the same string or if the pianist dampens the vibration actively (then you should hear them working the pedal on a detailed recording).
darTZeel is not completely feedback free. They use nested local feedback. Quads use feedback. I think it was Bruno Putzeys who demonstrated mathematically that nested local feedback is in effect no different than global feedback. Shindo's SETs are probably without feedback but their push/pull amps probably have some. BAT does not use feedback but if you look at the measurements, their transformers are clearly undersized and the distortion shoots up in the bass...this will color the sound of what is otherwise and interesting design.