|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
82.35.157.35
In Reply to: Is there a good set of plans for a good tube OTL amp posted by djn on January 2, 2005 at 15:07:58:
Hi,Warning 1 - I prefer SE Valve Amp's to ANY OTL I have heard.
Warning 2 - I have not heard too many.Notes.
All (and I mean ALL) PL/EL509/519 based OTL's I heard where VERY MUCH NOT too my liking. I heard them in many different configurations (SEPP and Circlotron from a range of makers). Best avoid, if your tastes in sound run like mine.
The best OTL I have heard so far are the the small 6S33 based Graaf OTL's. Atmasphere (M60 and the big (g)uns) runs second. Croft's smallest OTL's with 6AS7 run third, the bigger ones loose the "magic".
All in all i have always felt that while OTL's have some virtues, they tend to generalise too much, probably a result of the large amounts of NFB that invariably are required.
I wonder if someone could make a 800 Ohm voicecoil Driver with a really good sound, so we can make sensible OTL's, instead of forcing the output valves to work overley hard and having to use tons of NFB.
Thus for DIY I think you might be best off doing something like an AS M60 to "hear how it sounds" if you want to use normal speakers. Try to get 16 Ohm Speakers. Otherwise you end up quite likely adding and autoformer, in which case you might as well have started with a properly designed transformer coupled output stage.
A somewhat carzy idea would be to take a bunch of the 2/3/4" tangband Drivers and make a high impedance line array by putting them in series. If you put 16pcs 8Ohm in series you get 128 Ohm nominal impedance and a sensitivity of 12db more than the original single driver, so even with 84db/W/m Drivers you get 96db/W/m, of course, 1Watt is now around 12V.
Still, driving around 35V into 128 Ohm for 10W with 106db/1m is not a big deal, you only need around 0.4A Peak current for that. Maybe a pair or 3pcs of 6AS7 as SRPP and you get that sort of current/voltage quite easily.
Anyway, my thoughts.
Follow Ups:
Thorsten, I also yield to your knowledge of circuitry (about which I am woefully ignorant). I would merely like to point our that Joule OTLs have a knob that can increase/decrease NFB, although I do not know to what degree. That said, I find this amp exceptional (the VZN-80) and for the record, I also use Welborne Laurel SET and Manley SET/pp amps (in SET mose, the Manley also offers controllable NFB). Proper mating with speakers is, as we all know, the key element.
Hi,> I would merely like to point our that Joule OTLs have a knob
> that can increase/decrease NFB, although I do not know to what
> degree. That said, I find this amp exceptional (the VZN-80) and
> for the record, I also use Welborne Laurel SET and Manley SET/pp
> amps (in SET mose, the Manley also offers controllable NFB).Hmmm. Not sure about Joule and Manley, the Laurel Amp's I heard where okay sounding, but far from great.
> Proper mating with speakers is, as we all know, the key element.
Not sure I'd agree. Most competently designed Speakers will operate with very little power (as the only practical ways to make speakers with low compression and distortion tends to increase efficiency and sadly also bulk) and only need to be correctly impedance matched by the amplifier, which may be a problem for (some?) OTL's but is not one for transformer output SE Amp's.
Negative feedback in "long loops" may make an Amplifier more tolerant of incompetently designed speakers, but at a cost in sonic quality.
I yield to your superior knowledge in this field, Thorsten, but you should know that the Atma-sphere amps operate well (in my system, optimally) with no NFB whatsoever. One can apply a few db of NFB if one likes, to reduce the output impedance when and where necessary. Moreover, there are fundamental differences between designs based on the circlotron (Atma-sphere et al) and those based on the original Futterman circuit, in terms of stability, use of triodes in a balanced output stage, etc. Also, I hesitantly would say that the use of an autoformer at the speaker-end of cables, to boost the impedance seen by OTL amps on the order of 2X does not present the problems inherent to the design of a transformer-coupled output stage. In other words, the autoformer/OTL strategy is potentially less "harmful" to the signal than the OTC strategy.
Hi,> you should know that the Atma-sphere amps operate well
> (in my system, optimally) with no NFB whatsoever.Good on you. Compatibility will always depend. I also note that I did not remember correctly the amount of NFB on the M60, or maybe it has been reduced in recent revisions.
> Moreover, there are fundamental differences between designs
> based on the circlotron (Atma-sphere et al) and those based
> on the original Futterman circuit,Absolutely. I much preferred the Circlotron type Amp's over the SEPP (aka Futterman) versions.
> Also, I hesitantly would say that the use of an autoformer
> at the speaker-end of cables, to boost the impedance seen
> by OTL amps on the order of 2X does not present the problems
> inherent to the design of a transformer-coupled output stage.Sorry, transformers are transformers. They have magnetic cores which do all the hard work. These cores are subject to certain issues, as are the actual windings. Autoformers do not change this in the least. It is still a magnetic coupling device. It may be easier to make a decent low ratio autoformer, but it is not that hard to make a good higher ratio transformer.
> In other words, the autoformer/OTL strategy is potentially
> less "harmful" to the signal than the OTC strategy.Moreso actually, when compared to a good SE Design. Simply because the bias current in an SE Amp and the Airgap make the transformer core very linear. The key problem in SE Transformers is at very low frequencies, where the hearing and perception is a lot less acute then in the midrange.
As a 5 year owner of a Transcendent Sound Stereo OTL, and having auditioned an early A-S M60 in my system, (and various tube and SS amplifiers) plus also having listened to a pair of M60's and MA-1 in a dealer's showroom, I have the following impressions, which might shed some light into Thorsten's comments:Both amplifiers sound more similar than different. Both present a very big, layered and natural soundstage which is also very transparent and refined. Anyone used to the typical tube or SS amplifier might find this type of presentation as recessed because it is presented behind the speaker's plane, it is not an upfront presentation. Another characterstic is that they are lightning fast in their dynamics. Let's put it this way, they sound different to what you are used to listen from the typical amplifier.
I had an early A-S M60, version I, I believe, for a few days in my system, which compared to the TS presented a little bit more refined or maybe ethereal quality in some instances. However, both were very natural, smooth, sweet and both exhibited that aforementioned soundstage presentation and the same dynamic quality. The T8 had a little bit more bass control and slam though, I guess due to it's NF in the circuit. I believe the T8's output impedance is below 1 ohm.
This experience has also been experienced in my dealer's showroom, which most of the time has been using the same Wadia front end and speakers and different tube and SS amplifiers and the aforementioned A-S models. Again, The A-S amplifiers difference in presentation compared to other tube and SS amplifiers follows the same type of presentation ascribed to the OTL description given above.
I am not saying that one type of presentation or amplifier type is better; there has to be differences in the sound of each type of design. That is something that we have to live with in this industry and hobby and is part of the fun. Each design has it's own engineering and design constraints and also the designer's preferences built in.
BTW, my speakers are fairly efficient, 92 db and present an 8 ohm impedance (Reference 3A MM deCapo's) and I also use Thorsten's favorite preamp, a pair of Stevens and Billington transformer volume controls.
Do you know djn and what his tastes are?"Best avoid, if your tastes in sound run like mine."
How would one know what your tastes are so as to apply any relevance to your statement.
What does your phrase "they tend to generalise too much" mean?
As for my tastes, Atma-Sphere M-60's, driving Köchel K-300 horns, using an Audio Synthesis DAX Discrete for a source, provide me with ecstasy beyond words. I have heard you mildly diss each of these products, or at least their progenitor, at one time. What does this mean? People have different tastes?
We all value our own opinions, but your outburst is more than a little over the top IME. Having said all that I still enjoy reading your thoughts and would love the opportunity to meet and listen to your system some day.
One last thing, I don't think Atma-Sphere amps require NFB.
Hi,> Do you know djn and what his tastes are?
No, therefor I stated my preferences AND qualified my statements.
> "Best avoid, if your tastes in sound run like mine."
> How would one know what your tastes are so as to apply
> any relevance to your statement.Given that I have published quite a few reviews (in TNT Audio and at ETM) as well as loads of other writings that is easy enough.
> What does your phrase "they tend to generalise too much" mean?
Exactly that. In real music there is (unless computer genrated) an unavoidable minute variation from note to note, phrase to phrase and so on. I find that in subjective terms Amplifiers using large amounts of negative feedback tend to obscure these variations more than amplifiers with no loop feedback.
> As for my tastes, Atma-Sphere M-60's, driving Köchel K-300 horns,
> using an Audio Synthesis DAX Discrete for a source, provide me
> with ecstasy beyond words. I have heard you mildly diss each of
> these products, or at least their progenitor, at one time.The Koechel Horns, you should be aware, are Speakers I quite like.
I listed the M60 as my second favourite OTL Amp of those I heard (which does not say I don't prefer something else though).
The various DAX Dac's tend to also come up occasionally in my writings as worthy of note, however I feel that a really well implemented Non-Oversampling DAC offers sound more to my liking.
> What does this mean? People have different tastes?
Exactly. Recording & Reporduction is invariably drastically flawed. In order to get the experience we want from listening to recordings we need to arrange our systems such that they avoid emphasising the areas where we as individuals tend to react to the difference between recorded and real music while emphasising or at least excelling at what are the core areas of music that us spell "reality".
> We all value our own opinions, but your outburst is more than
> a little over the top IME.OUTBURST?
Someone asks "what is a good OTL Design to build".
I list my experiences with OTL Designs I have heard and simply note that contrary to the Hype some people generate OTL Amplifiers are not perfect or the end to it all.
And that is an OUTBURST? Pulleeese.
> Having said all that I still enjoy reading your
> thoughts and would love the opportunity to meet
> and listen to your system some day.If you find yourself in London, you are welcome to drop in.
> One last thing, I don't think Atma-Sphere amps require NFB.
The commercially sold ones have it and quite a good dollop too. If you remove the NFB, the output impedance gets very high. High enough that you actually need to design a speaker especially to co-operate with that high output impedance (which is not per se a bad thing, on the contrary, I have been advocating "current drive" to speakers for a long time).
And of course the distortion goes up and as in OTL's in generally the distortion open loop must be fairly high and is generally composed of quite a few high order components, which may or may not be audible as sonic problems.
As with all engineering, everything is about tradeoffs. Going OTL means you trade off some things for the ability to say "there is no Output transformer coloring my sound". I believe for somone who wonders if he should invest into such a device it is important to be aware of the tradeoffs.
The M-60s have about 0.5db of negative feedback if I recall correctly (hardly a huge amount). The MA-1 and MA-2 have always had ZERO negative feedback. The newest ver.3 of the MA-1 and MA-2 now includes switchable negative feedback for 0, 1 or 2db.
Hi,> The M-60s have about 0.5db of negative feedback if I recall
> correctly (hardly a huge amount).The current M60 lists as follows:
Output Impedance: ~4.1 Ohms
Feedback: 1 dbIt seesm the feedback has been reduced from earlier versions (good idea), if at the cost of an output impedance a bit on the high side for the real world.
> The MA-1 and MA-2 have always had ZERO negative feedback.
> The newest ver.3 of the MA-1 and MA-2 now includes
> switchable negative feedback for 0, 1 or 2db.So NFB was added.
I agree however that these leves are much less than what memory told me, especially on the subject of the M60.
Still, it would seem that M60 has around 5 Ohm Output Impedance without NFB which makes using it with conventional speakers a little difficult. The 6db or so NFB I remembered would have brought this down to a more usefull 2.5 Ohm, now more less on parity with a typhical SE Amp on the 8 Ohm Tap.
However, these fairly low levels of NFB (as also found in the Graaf) underscore why I listed them among my favourites, strictly on sound, as I did not dive too deeply into their designs. And why the SEPP OTL's I have heards all where much worse.
It's probably worthwhile for people considering OTL's to read carefully the various specifications found on AtmaSphere's website.
So my recommendation to djn remains as written, try a M60 Clone, there is enough info out there to do that.
Ciao T
BTW, here a page with an Amp one might call AS M60 Inspired....The Guy who build it omited NFB and clocked 10 Ohm Output Impedance, suggesting that really some 12db of NFB would be needed to get the output impedance down to tolerable levels for conventional speakers, or using 4 X the number of output Valves.
Ciao T
Your posting, read in the wee hours especially, appeared quite condescending, almost as if you believed yourself the sole arbiter of good taste and sensibility. Re-reading your post in the light of day, along with your further responses in this thread, temper that impression considerably. Again, pardon my outburst.I am still not quite tracking your thoughts on 'generalisation'. Any additional insight is appreciated. You seem to be saying OTL's using NFB severely limit resolution.
Hi,> I am still not quite tracking your thoughts on 'generalisation'.
With that I mean a diminished perception of the differences between the individual notes playing style. In other words, each note or phrase sounds more like the others except for diferences in pitch and general loudness.
> You seem to be saying OTL's using NFB severely limit resolution.
No, I am saying that all amplifiers I have heard, designed and build which use large amounts of NFB around certain types of fairly non-linear structures (among which arguably fature OTL's) seem to lead to an effect that subjectively seems to rob music of some of it's life, of some of the things that for me demark differences between "real vs. memorex".
This even appears to apply to certain forms of digital feedback (Delta Sigma Noiseshaping for example).
You rightly emphasize two things that everyone probably knows, but which are often overlooked, I think. (1) Excellence in hi-fi does not yield a single linear ordering (a single ranking, best to worst), because what is desirable in sound reproduction is not given a priori - different people will value different things. BUT this does not mean that it is all subjective, that there are not objectively or intersubjectively valid rankings etc; it merely means that there are many orderings (which may criss-cross in various ways). (2) The bottom line for engineering in any domain is, does it work? Does it do what we want (at acceptable cost, reliably etc.)? Our hi-fi world is infested with component or design-fetishism - this type of amp, this type of wire is absolutely the best. Fetishism, by definition, is the belief that certain properties are contained in some object which are not really contained in the object, but only in certain contexts in which the object occurs. One falls prey to it precisely when one is ignorant of the way in which the context works - eg, of the ways in which circuits actually work. It is easier to pick out some element and imagine that the magic is all contained in that. So for an amateur like me, all I can do is to be very sceptical of what anyone who is not a real engineeer says about hi-fi, except, at most, what something sounds like.
Hi,> (1) Excellence in hi-fi does not yield a single
> linear ordering (a single ranking, best to worst),
> because what is desirable in sound reproduction is
> not given a priori - different people will value
> different things.Absolutely. Which is one of the key reasons why so far no formal measuremenst of quality have been devised that show any significant correlation with percieved "good sound".
> (2) The bottom line for engineering in any domain
> is, does it work?How do we define "it does work"?
Does it work if it fulfills the German DIN 45500 Standard?
> Does it do what we want (at acceptable cost, reliably etc.)?
How do you define this then, unless on an individual, subjective basis?
> Our hi-fi world is infested with component or
> design-fetishism - this type of amp, this type
> of wire is absolutely the best.Perhaps.
> Fetishism, by definition, is the belief that certain
> properties are contained in some object which are not
> really contained in the object, but only in certain
> contexts in which the object occurs.In other words, a given thing may be considered relatively, in a given context, but not absolutley.
An example may be SE Valve or OTL Amplifiers and Apogee Scintilla Speakers or old 15 Ohm Tannoy Coaxials in GRF Autograph Enclosures. The SE Amp or OTL will be very much "a bad thing" in the system with the Apogees, while in the system with the Tannoys they would be both likely exceptional, as overall system. Equally, the Monster Krell Solid State Amp that would be "a good thing" with the Apogees would not sound too good with the Tannoys.
> One falls prey to it precisely when one is
> ignorant of the way in which the context works -
> eg, of the ways in which circuits actually work.Absolutely.
> So for an amateur like me, all I can do is to be
> very sceptical of what anyone who is not a real
> engineeer says about hi-fi,Hmmm. How do you define "real engineer"? It it someone who constantly bangs on about how great an engineer he is or somone who actually understands things? And how do you judge if you yourself do not understand the subject well enough?
As a degreed EE myself with experience in both industrial/military electronics and audio and as someone with quite a few years as sound/recording engineer I find myself highly sceptical of what many engineers claim about things audio/electronics. I have had enough time to collect practical experience which suggests that traditional theoretical foundations are sufficiently incomplete to be relied upon to the exclusion of other (empirical) methodes.
Therefore those subscribing to the orthodox readings of electronics (and acoustics/electroacoustics) suggest to me usually the presence of academics, as opposed to real engineers.
> except, at most, what something sounds like.
I would argue that few Audiophiles are qualified to comment on how something sounds, lacking sufficiently points of reference and schooling in critical listening, at least when it comes to commenting on anything beyond "what it sounds like to me".
I am actually planning an article for the British Journal of Aesthetics on hi fi - never been done before and I'm persuaded that hi fi raises some genuinely distinctive philosophical issues - or at least, they intrude practically in a way that is unusual. I am planning to interview several hi fi people (I can tell you who is on the list by email), which will also make the article rather unusual in a philosophical journal.As I say, it is not true that things sound or look to one as one think they do (though it is natural and excusable to think they do). The key is that (as philosophers long suspected, and empirical results have now demonstrated beyond doubt) the perceptual field arises independently of, and anterior to, cognitive/evaluative processing. For example, the phenomenon of blindsight: subjects with certain brain injuries can perceive things that they claim not to (this is shown by their propensity to guess correctly - or rather they think they are only guessing). Many other phenonemena point the same direction. Judgements based on perception can be influenced in many ways, and in perfect sincerity. That is why, for example, people will insist on the dramatic difference their new power cord makes to their system, without being in the slightest bit insincere in claiming to hear the difference. They are wrong about what they hear, as paradoxical as that sounds. Anyway if you are interested in contributing to the project, let me know (it would not be onerous in the slightest).
Hi,You may forgive me for declining, but like so much in modern science modern, so-called Philosophy has gone well past any point where it still is useful or relevant to me personally at least. I could not make any points any better than they have already been made by Bishop Berkley, Hume and Kant, if they do not suffice, what can I possibly add.
That said, I agree that people have a much greater acess to absolute reality than it may seem at times.
However music reproduction is in fact the art to maximally remove and isolate the listener from absolute reality and engender in her or him the perception of listening music and not to air vibration created by mechanical devices reacting imperfectly to electrical signals which imperfectly encode the airwaves from an original musical event.
Thorsten, have you heard a Bruce Rozenblit OTL? (Transcendent Sound) - I may have asked you before, but have forgotten. Like CB I have them and thus the bias of an owner, but I would surprised if they did not sound different from other OTLs. He's also got a true single-ended one now.
Hi,> Thorsten, have you heard a Bruce Rozenblit OTL?
I heard a DIY one from his designs. It used EL509's.
> I would surprised if they did not sound different
> from other OTLs.I would be surprised if they did. They are a rather traditional implementation of an SEPP OTL output Stage with large amounts of overall feedback and many stages. The "novel" part is the way BR attemps to impbalance the drive signal to the output valves in order to compensate the different gain between the common cathode and common anode part of the output stage. An extensive discussion of the merits (or lack thereoff) of his solution can be found in the Tubecad Journal. Ignoring the different approaches to the Drive Signal Imbalance, the average Croft OTL and one designed by BR have more in common than they differ.
However, I reapeat, if you like what you got, good on you. Enjoy.
> He's also got a true single-ended one now.
That one I have not heard, if I find myself with too much time I'll probably build something similar (the basic design goes back to a 1950's Japanese design BTW).
Hi Thorsten,A few years ago you helped me by email with some issues in an SE 300 B amp using your C3M driver circuit. I have been interested for some time in the idea of an SET OTL, and am interested the Transcendent Sound produced one. Never heard it though. If you do get around to experimenting with such a circuit, I for one would be very interested in learning about it.
I built one of the first, if not the first, pair of M60 kits that Atma-Sphere sold quite a few years ago. (It used 12AT7's and 12AU7's in the driver circuit.) I modified them quite a bit with Ralph Karsten's help. Later I traded them for another component. While I personally like SET's better, I do think--from a subjective experience viewpoint--that each has their advantages. The OTL's can have amazing speed and transparency. The Atma-Spheres did sound good when I briefly tried them with some warm-sounding Tannoy Stirlings I used to have.
So I think an SET OTL with at least about 5 watts per channel is a very intriguing idea.
I just can't imagine how one would get the output impedance of an SET OTL amplifier down into the range where it could drive any real world (8-ohm) speaker, using a realistic number of any real world triodes of which I am aware. For 500-ohm headphones, maybe. (Obviously, since Transcendent has apparently done it, it can be done, but I want to see the schematic. If it requires tons of NFB, it's a Pyrrhic victory.)
I suppose you know that according to BR, the author of the Tubecad article completely misunderstood his circuit. btw, I don't go for the 'If you like it, then good on you' school (and I doubt you really do either!). That is is just a way of avoiding disagreements. It is not true that if a person thinks something sounds good, then it it really does sound good to them!.
Hi,> I suppose you know that according to BR, the author of
> the Tubecad article completely misunderstood his circuit.I suppose you know that the Tubecad Article author analysed the circuit according to how it ACTUALLY works. IF BR's understanding as to how it works differs, well that BR's problem. I cited the article simply because it saved me repeating the main points.
I would prefer to not get into exhaustive discussions of any of BR's Circuits. He has been off base so many times (including Discussions I had directly with him) and is way too constantly puffing himself up on him being an EE (That don't impress me much - so, BR got an EE degree, well so do I, but I don't keep banging on about it and tub-thumping about it all the time) for me to take him overly serious.
> btw, I don't go for the 'If you like it, then good on you'
> school (and I doubt you really do either!).Actually, I do. With the implied limitation that just it sounds good to you it by no means has to sound good to anyone else. Personal truth is just that and unarguable. However personal truth is rarely universal.
> It is not true that if a person thinks something
> sounds good, then it it really does sound good to them!.Why not? If I feel I enjoy a good Redwine because it tastes good and I don't enjoy good White Wines as they don't taste good, who is to argue that in reality I'm imagining redwine tastes good and in reality white wine is what tastes good TO ME?
This post is made possible by the generous support of people like you and our sponsors: