So, let's sum this up a little bit:
You want to make a single ended OTL amp without a cap at the output and without parallel tubes.
So you're going to put 60dB of feedback around a 300B and accept 2/10 of a watt of power, or you're going to make a hybrid power amp?
The magic of SET isn't just the output transformer, or the lack of feedback, or the linearity of directly heated triodes like a 2A3 or 300B, but rather the combination of all of these elements.
You want to make a single ended OTL amp without a cap at the output and without parallel tubes.
So you're going to put 60dB of feedback around a 300B and accept 2/10 of a watt of power, or you're going to make a hybrid power amp?
The magic of SET isn't just the output transformer, or the lack of feedback, or the linearity of directly heated triodes like a 2A3 or 300B, but rather the combination of all of these elements.
Eew god no. Going hybrid. Single stage tube hybrid SET. Thinking of using 01a, I hear good things.
The only difference between my amps and a typical SET is I don't use OT and I use high performance linearization and distortion cancellation/error correction schemes, and I guess my power supplies perform better than the typical LC filter.
A transformer is the most imperfect medium of signal transfer, thus this is where my focus currently leads.
My amps are crazy hi-fi but not "magical", at least not the ones I designed to be ultra high performance.
Of course a combination of factors effects the overall sound quality but I have a difficult time believing the SET midrange "magic" is due to a combination of factors due to the reasons I've already discussed. Something like that must come from an isolated source mechanism.
Last edited:
It's funny you say that. If we need 5x voltage amplification, I'd take a signal transformer over a single transistor.
Different horses for different courses I guess.
Member
Joined 2009
Paid Member
There is a whole priesthood behind the fear of passives, but what can you really control ? The speaker is a coil too, it may have a crossover made with coils caps and resistors. Where did the sound travel on its way to the input of your amp? Did the musician use a coil pickup? How many opamps were used? The mixing ? At the end I think you have most influence with the speaker and its interaction with the amp. This is where a common cathode into an OPT has a particular flavour.
Surely, if You want to develop amps that sound like a SET then You'd want a reference unit to compare them with not rely on words from strangers on the 'net
Surely, if You want to develop amps that sound like a SET then You'd want a reference unit to compare them with not rely on words from strangers on the 'net
Last edited:
Well the whole reason I got interested in audio electronic design is because of the changes I heard when I began cap rolling. Caps do have a sound, in certain circumstances. I've designed myself a pure class A pure current source amplifier for ribbon use so I soon won't need to bother with coil issues at all.
I also have some really cool ideas for graphene thermoacoustic drivers but that stuff is like 3 million dollars per gram, or a few hundred bucks for a barely usable sample. Unfortunately that means it's on the bottom of my priority list but the lure of potential is too great, I plan on building a prototype before the end of the year.
Transformers on the other hand have a slew of measurable issues. From a pure engineering standpoint I don't consider using them. But I'm not completely against the idea. If I were to use them I'd probably use a 1:1 ratio to get the highest quality out of it. Does anyone sell 1:1 signal transformers?
I've had limited experience listening to SET, I did notice the midrange was very pleasant to listen to. The words "sweet" "soft" come to mind.Surely, if You want to develop amps that sound like a SET then You'd want a reference unit to compare them with not rely on words from strangers on the 'net
The rest of the audio band was not up to par but the mids were really nice.
Last edited:
Member
Joined 2009
Paid Member
I've heard a couple of SETs. The first one was a test set up really, where I got hear a friend cycle through choice of output tubes including a 6C4C, 6P31S and something else. I really liked the smooth nature of the 6P31S for that short test but none were magical. Later I built a SE 6AS7 which being my first tube amp was greeting with the enthusiasm of any newborn, but it wasn't good enough to be my daily driver. I then heard a friends SV811-3 based SET and it sounded magical - hard to describe but sweet and soft it wasn't. It was simply 'music', with bass and treble and a superb midrange. I compared it directly with a Bryston SS amplifier, known for it's low distortion - it sounded soulless, unsatisfying, flat and bereft of any musical pleasure in comparison. It was an ear-opener for me and let me understand why people want this kind of amp.
Now, my best SS amp that I designed and build a few years ago sounds different. It has gobs of feedback and simulates with negligible distortion. It has much tighter bass but of course it has 40x the power too. It has an overall cleaner sound from top to bottom and is perhaps a little more extended at both ends of the frequency range. To me the sound is simply faultless, the nicest sounding amplifier I've ever heard, DIY or commercial. But it does not have the SET magic that I remember from the SV811-3. The two amps represent almost apposite ends of the spectrum of technologies and I like both approaches just as much as I like a good English beer (Tetley's is a good one) one day and a cold clean Canadian lager (Mill Street Organic is a good one) another day.
Now, my best SS amp that I designed and build a few years ago sounds different. It has gobs of feedback and simulates with negligible distortion. It has much tighter bass but of course it has 40x the power too. It has an overall cleaner sound from top to bottom and is perhaps a little more extended at both ends of the frequency range. To me the sound is simply faultless, the nicest sounding amplifier I've ever heard, DIY or commercial. But it does not have the SET magic that I remember from the SV811-3. The two amps represent almost apposite ends of the spectrum of technologies and I like both approaches just as much as I like a good English beer (Tetley's is a good one) one day and a cold clean Canadian lager (Mill Street Organic is a good one) another day.
Last edited:
I've never heard a commercial SS amp that I liked. I too have found my personal design is better than the commercial stuff. But I don't use Global feedback, it's kind of one of my rules. Even local feedback I try to restrict as much as is pratical. Of course tubes and things always have their own internal feedback which is necessary. I don't think local feedback is "bad" but I hear too many horror stories about NFB to not want to use it. I get really good results without it anyway.
Even though, without the use of purposeful feedback the theoretical distortion of my SS amp is 0.000000000225%. This is impossible to confirm but the harmonics are below the threshhold of my distortion analyzer so I can only assume it holds close to theory. What else is interesting is it should, in theory, maintain this distortion figure regardless of load or signal size. This holds up pretty well in the sims.
In terms of raw performance it's by far my best achievement. From a practical standpoint it sounds like garbage or like heaven depending on what I feed it, so I have no reason to doubt that it is completely transparent.
My journey towards extreme performance is almost finished, there's a few more designs I need to build before I'm done there. Then I move on to R&D of distortion generation. I don't even want to think about all the time and money I've dumped into this. I've got all of my eggs in the "start an audio business" basket. From a design and specs perspective I should dominate, but I can't even afford a car let alone a company. It will be quite a challenge.
Even though, without the use of purposeful feedback the theoretical distortion of my SS amp is 0.000000000225%. This is impossible to confirm but the harmonics are below the threshhold of my distortion analyzer so I can only assume it holds close to theory. What else is interesting is it should, in theory, maintain this distortion figure regardless of load or signal size. This holds up pretty well in the sims.
In terms of raw performance it's by far my best achievement. From a practical standpoint it sounds like garbage or like heaven depending on what I feed it, so I have no reason to doubt that it is completely transparent.
My journey towards extreme performance is almost finished, there's a few more designs I need to build before I'm done there. Then I move on to R&D of distortion generation. I don't even want to think about all the time and money I've dumped into this. I've got all of my eggs in the "start an audio business" basket. From a design and specs perspective I should dominate, but I can't even afford a car let alone a company. It will be quite a challenge.
Last edited:
No, definitely not. Due to magnetic anisotropy, transformer nonlinearity is mainly because of the core, the windings also have capacitances, and again, due to dielectric anisotropy some nonlinearities are due to windings but orders of magnitude lower.
Member
Joined 2009
Paid Member
For my hobby at least, my approach is different. I often begin by studying the journeys of others but I don't limit myself to their path. I observe things for myself, I always ignore their horror stories and I remind myself not to make any rules. I never try for a result, or any 'extreme' performance except to use simulations and pcb error-checking for reducing the number of errors.
Last edited:
Well there's an infinite number of ways to make an amplifier, and only a finite number of ways to make them based on a performance objective. I can't practically test all possible methodologies to find the truth of everything so I need to set some rules to maximize the odds of making a good design. Being methodical reveals patterns and trends more easily as well which leads to better future design.
topologically speaking I usually ignore the "journeys" of others and find my own path. I spend a lot of time in spice trying out quirky ideas, this is how I've gotten all my best designs.
From a commercial perspective, based on personal experience I think my stuff basically dominates everything out there, at least everything that I've heard, but my goal is not hi-fi as much as it is euphony so I won't be satisfied until I've perfected an amp that is ultra euphonic and ultra realistic at the same time.
topologically speaking I usually ignore the "journeys" of others and find my own path. I spend a lot of time in spice trying out quirky ideas, this is how I've gotten all my best designs.
From a commercial perspective, based on personal experience I think my stuff basically dominates everything out there, at least everything that I've heard, but my goal is not hi-fi as much as it is euphony so I won't be satisfied until I've perfected an amp that is ultra euphonic and ultra realistic at the same time.
Last edited:
No, magnetic anisotropy is a property of ferromagnetic materials, I do not know what you mean by removing the DC component, as for Maxwell equations and Fourier analysis it has different meanings.
If you mean a gapped core to deal with DC current in SET amps, it does help to linearity, not eliminating magnetic anisotropy, but reducing effective magnetic permeability.
Member
Joined 2009
Paid Member
Absolutely, this is often how progress is made in many areas of human endeavour. I often gravitate in this direction but eventually discovered that it turns my hobby into a 'job' (you see some of that around here). I now prefer to set no performance objectives or future expectations and give priority to enjoying myself each time I pick up the soldering iron or the wood glue....
Last edited:
Member
Joined 2009
Paid Member
No dc current to bias the magnetic core - I've read some folk say that a little bias is a good thing and don't like things fully symmetrical, especially when they are talking about their concerns over magnetic domain reversals in push pull output transformers.
Unfortunately I'm too much in the hole for this not to be a job. It's make or break for me. Which sucks because I really really hate having to think about amp design 24/7.
With any luck I should have a finished product line within a year. I would kill someone to split the load.
So, not only does it not eliminate the magnetic anisotropy, it does not improve the linearity either.
Yes, a DC current with a gapped core has two benefits
i) Increasing linearity due to reduction of effective magnetic permeability
ii) Avoiding zero crossing on the magnetic hysteresis curve B=μH, it also reduces Barkhausen noise.
Member
Joined 2009
Paid Member
There is a man who claimed he can make his amplifier sound like any other. That would be Bob Carver and his famous challenge, which in the one case he won. Perhaps that story holds some hope that a SET amplifiers sound can be achieved without the OPT?
The Carver Challenge | Stereophile.com
The Carver Challenge | Stereophile.com
Yeah I know about the carver challenge. I think he's using some sort of shade network or something based upon the descriptions of how he used his circuit to adjust the harmonics.
Since he was using 12 trimmers the variables are immense and it's probably only practical to copy an existing design using it.
His design imitates the transfer curve of the DUT but as far as I know he never released his design publicly? Of course such a design would completely break the market.
Since he was using 12 trimmers the variables are immense and it's probably only practical to copy an existing design using it.
His design imitates the transfer curve of the DUT but as far as I know he never released his design publicly? Of course such a design would completely break the market.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.