For single amplifier SET systems then, yes, the OPT is a big issue and can certainly flavour the sound by changing the alignment of the bass channels of the speaker and perhaps rolling off a little early up high in the treble. I often hear that when a SET amp is used on an inappropriate speaker.
Break the bass channel away from the SET amplifier, or add a separate channel/amplifier especially for the bass so that the back EMF from the woofers do not impact the mids and tweeters so much and you get away from what some call "SET sound", which can be heavily flawed, and you get into much more fullrange sound that is cleaner through the entire audio frequency spectrum. Those SET traits (when used with an appropriate speaker) of unbridled dynamics and wonderful tonal colours are still there and are improved, but they also extend an octave or more further up and down than they did before.
So yes, transformers I am sure are part of "SET sound" as many people would identify it.
Member
Joined 2009
Paid Member
I would only build a hybrid SET for commercial sale. Speaker matching would not be an issue. I'd much prefer to stick to an OTL design though, which is one of the reasons why I'm trying to unravel the mysteries of "magical" tube sound since it would effectively be a single triode with the plate directly coupled to the speaker.
100% agreeAs for "good engineering practice" I must express my praise for hellokitty at attempting just that. Good practice is not to keep trying the same tired and populist ideas in spite of failure to give improved sonic results. Good engineering is doing something that gets results.
In this community, some members get highly fixated on overly simplistic theories and naive applications of textbook models and all too often shoot down people trying to break out of the resulting mental prison. An example of this is the ludicrous stand-offs between THD and the human auditory system, with the human usually being dismissed as unreliable.
So I say to hellokitty, ignore anything that sounds familiar advice and do something different, with my blessings.
Last edited:
From what I can figure reading this thread you have had some sort of sound for a month or so and you have only listened through headphones. As you are aware speakers are different beasts to headphones...have you tried your circuit with speakers? Low sensitivity? High sensitivity? Impedance peaks/dips in the speaker response?
What I would call the "SET sound" only occurs on speakers under certain circumstances that I outlined earlier. If a SET amp is forced to work at all it loses dynamics and it harmonics can be screwed up making it sound ordinary or even rubbish. The OTL amps that I have heard either had good dynamics or good harmonics but never both. The lively ones seemed to screw with the sound, sort of like it was pulled apart by the amp but it could not quite figure out how to put it back together again (like Humpty Dumpty). You may have solved these problems and good for you if you have...but just saying you have does not make it so.
Hybrid. There's a zillion ways to go about it. Saying it won't have any form of feedback would be a lie of course. The type of feedback depends on the design. My choice of design would be effectively open loop as far as the tube is concerned if we ignore the obvious internal NFB of the triode.
My plan revolves around using a single stage hybrid design, but I don't really know if there's a practical difference between that and just buffering a tube with a follower. I'd rather not use a follower though, not as cool and I can think of some possibly incorrect or non-science based reasons why direct tube connection may be better.
No need to try it with speakers. I can set it to output any arbitrary power I want with an adjustable output impedance from near 0 to the tubes output impedance.
Yes, that's why you use a hybrid design. The the tube transfer curve is imposed upon the SS parts, or the SS parts simply copy the signal, or they apply some form of error correction, there's a lot of ways to do it. In all scenarios the tube is effectively functioning into a near infinite impedance in the entire audio range.
Member
Joined 2009
Paid Member
There has been, for many years, this idea that you can impose the tubes characteristic curves onto a SS power stage. It's not that difficult of course, whether a tube + follower or some variants (there's an interesting one with feedback to a screen grid wrapped around a FET). And many amps have been built with varying degrees of commercial and sonic success. I don't believe any of them were accused of being 'SET' like though. I would imagine that far more R&D $ and time has been expended on this search than that which you have invested yourself. It really will take something new if you want to achieve something new. What that is I don't know !!!
Last edited:
Oh my...
So these are not SET circuits you are building, they are hybrid SS. The triode is driver stage only and power stage is SS...ok clear now. Yeah, this has been done quite a bit before: I even had some Class D amps here with a tube input stage. Still sounded rubbish.
Assuming I understand correctly, I think that SET OTL is not the correct topology name for the circuits you have there...Hybrid SE SS is more accurate. So you have done all this and do not have the "SET sound" then maybe your best next step is to build a nice two or three stage SET the traditional way with lots of iron in it and see which bits you can leave out. Afterall, it is the speaker system and circuit that determines the SET sound, not the transfer function of the driver tube.
I salute the decision to use tubes in driver and pre-amp stages, they are superior to any transistors of course! take one EL84 for example.
However using transistor in the output in Class A to me is a transistor amplifier, it is not a SET... why use "set-sound" in the title?
And you need something to block the DC in your class A SE amplifier, how do you bias your transistors? How are they superior to tubes in your application? Do you seriously think you will get a SET sound with a SE transistor output stage?
More: on the SS copy of the signal. why the heck use any tubes if the amp is able to copy a signal?
However using transistor in the output in Class A to me is a transistor amplifier, it is not a SET... why use "set-sound" in the title?
And you need something to block the DC in your class A SE amplifier, how do you bias your transistors? How are they superior to tubes in your application? Do you seriously think you will get a SET sound with a SE transistor output stage?
More: on the SS copy of the signal. why the heck use any tubes if the amp is able to copy a signal?
Last edited:
No not a tube input stage + SS output stage. That is done to death.
The idea I like best would be just a single stage with a master slave 100% feedback loop where the tube transfer curves are completely imposed upon the SS part. A parallel configuration where the transfer curve of the SS part is eliminated and it acts as the tube does. The plate of the triode is still directly connected to the load, just a single triode amplifying in the truest sense while also being a hybrid.
In an input stage + SS output stage the tube just passes the signal to the output stage and hopes for the best. In my version the tube forces the SS part to act as itself.
It's really fundamentally different but at the same time I don't really know if the difference is practical.
Of course I can think of a million other ways to hybridize a tube but the elegance of the single stage one is too great to ignore.
In either case Bigun is right, I need more than just the design itself. My only idea at the moment is to try to find a non linear bias area of the tube. I think biasing 6sn7 at very low currents might be a decent start.
Last edited:
Yeah it's a bit of a PITA though, don't have a source to work with at the moment. I've got to wait another few weeks to be able to afford finishing building my new dac. Then another month or so optimizing and convertig my DAC output stage design to PCB.
Then I need to optimize and order PCBs for this super-super reg I accidentally discovered. (would anyone actually be interested in buying regs that out-perform the jung super reg? Curious to know)
Then I need to build the WCF 6c33c stage for my friend.
Then I need to order PCBs for my latest line stage design I just built.
Sure, I wasn't impressed with the sound because it has no sound, but the performance of the design is probably the best you can get out of a tube line stage, at least from a theory perspective that I can think of. Harmonic measurements show a 15-40db improvement from comparable designs, but there aren't too many 4P1L line stage THD graphs floating around so I can't say for sure.
From a commercial perspective it's more than qualifying, my subjective goals aside.
After that I may or may not want to convert my solid state amp to PCB, because it is a pure SS design the prototype is pretty small and mechanically stable, I may be able to let this one slide for a while.
Then I can start playing with my shelved ideas like OTL set amps and whatever else I have in the backlog. Probably in the late summer or fall. This is why I'm doing some theoretical R&D now to while I wait.
Then I need to optimize and order PCBs for this super-super reg I accidentally discovered. (would anyone actually be interested in buying regs that out-perform the jung super reg? Curious to know)
Then I need to build the WCF 6c33c stage for my friend.
Then I need to order PCBs for my latest line stage design I just built.
Sure, I wasn't impressed with the sound because it has no sound, but the performance of the design is probably the best you can get out of a tube line stage, at least from a theory perspective that I can think of. Harmonic measurements show a 15-40db improvement from comparable designs, but there aren't too many 4P1L line stage THD graphs floating around so I can't say for sure.
From a commercial perspective it's more than qualifying, my subjective goals aside.
After that I may or may not want to convert my solid state amp to PCB, because it is a pure SS design the prototype is pretty small and mechanically stable, I may be able to let this one slide for a while.
Then I can start playing with my shelved ideas like OTL set amps and whatever else I have in the backlog. Probably in the late summer or fall. This is why I'm doing some theoretical R&D now to while I wait.
Last edited:
Those two quotes are referring to two different designs, not that I couldn't increase the output impedance of the SS amp if I wanted I guess. I'd probably want to stick to its strengths and keep it low impedance.
The tube hybrid will have nowhere near such a distortion figure and it will obviously always have gain.
I'm almost done with my R&D into high performance, HI-FI designs. What I'm looking for now is to inject some euphony or "tube magic" into it. This requires some form of imperfection, just need to find out what that is.
Based on what people have said there seem to be different experiences of "tube magic". My experience with euphony OTL amps is more of a "something" in the sound that just makes me feel like I got a dopamine shot. Present in the entire audio range, not just the mids. Consistent across all forms of music. My very limited experience with a SET amp revealed something similar in the mid range but not quite the same kind of euphony as I'm used to. Statements I've read about SET sound seem to confirm this. In either case, finding the source of these "euphonies" is where my aim is beginning to shift.
The tube hybrid will have nowhere near such a distortion figure and it will obviously always have gain.
I'm almost done with my R&D into high performance, HI-FI designs. What I'm looking for now is to inject some euphony or "tube magic" into it. This requires some form of imperfection, just need to find out what that is.
Based on what people have said there seem to be different experiences of "tube magic". My experience with euphony OTL amps is more of a "something" in the sound that just makes me feel like I got a dopamine shot. Present in the entire audio range, not just the mids. Consistent across all forms of music. My very limited experience with a SET amp revealed something similar in the mid range but not quite the same kind of euphony as I'm used to. Statements I've read about SET sound seem to confirm this. In either case, finding the source of these "euphonies" is where my aim is beginning to shift.
Last edited:
Looks like you’ve got too many projects cooking at one go Hellokitty. Perhaps you could share your regs design here & do a GB for the boards or ready build & tested regs. If your looking for euphony, you could try small tube rectifiers for low current side of your design if your game that is.
Cheers
Cheers
You have no idea
I've got at least a year or two worth of testing before I've exhausted my to-do list. It must be done though, gotta be thorough.Yeah I've been thinking about it. Not like I gain much by keeping a pico/nano ohm theoretical output impedance reg a secret, I doubt the performance would ever be audible in a design and kit sales would be the most optimal place to sell.
In the end I may just sell all my designs as kits. Would remove the middle man of having to blow money on "business" costs. Plus I could post my designs on here without worry. Everyone benefits.
I don't want to jump the gun though, need to carefully consider my options.
The results of my R&D are my ticket out of this crap hole of a life situation I'm in, at least that's the hope.
I'm not against trying tube rectifiers. I just have no room for them any time soon in my slot for "things to test" unless I have significant reason to believe they are a benefit. More so because it would be expensive and annoying to build.
Once again you are using "NFB" and "GNFB" without distinguishing them. Can we assume that each time you say "NFB" here you actually mean "GNFB"?Hellokitty123 said:
I was just trying to clarify whether he uses feedback or not. It is common for some audio designers to claim 'no feedback' then use lots of feedback. When queried, the usual answer is that 'no feedback' actually means 'no global feedback' (in which case, why not say so) or that massive degeneration is not feedback (it is feedback). I think we have now established that Hellokity123 uses local feedback, and no global feedback - which is fine.traderbam said:
My point is this: if someone announces "I don't use feedback" then we should be able to assume that this is what he means, and that he thus obtains the required linearity by some other means such as balancing or feedforward. If he actually means "I don't use global feedback" then that implies he probably uses lots of local feedback instead - which is fine and a perfectly valid engineering decision. I am just seeking clarity in language. It matters because some people believe that feedback of any sort is bad so for them 'no feedback' is a positive thing to say.
Why drag THD in? Naive applications of textbook models may be better than ignorance of textbook models, as seems to be the case for some audio designers. Better still would be intelligent thoughtful applications of textbook models.traderbam said:
Member
Joined 2009
Paid Member
One of my earliest thoughts was along the lines of this thread
SET amp using Triode-FET complementary feedback pair ?
SET amp using Triode-FET complementary feedback pair ?
Not really the way I would do it. It looks to me like a consolidation of a source follower into a single stage. Cool idea but ultimately I could just make a really nice power follower and probably achieve better results. The Gm of the mosfet is overpowering the tube in that example, it's sort of the opposite of what I had in mind.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.