Hi Graeme
Conceptually your above idea is an embodiement of Fig 2 of Dr. Hawksford's paper
"Transconductance Power Amplifier Systems for Current-Driven Loudspeakers".
Which is in itself the topology of a current mode feedback op-amp, take a look
at the topology on the Burr-Brown OPA603 datasheet, it may not be obvious but
this is more or less the same thing as Dr. Hawksford's Fig 2, except they have
taken the current output and used it to drive a voltage follower. BTW, your
circuit is different in that it is differential.
Coming back to your circuit, if we do a quick analysis of the transconductance
we find that the voltage feedback network fixes the single ended gm = -1/R22,
this is low ! What is missing is the current feedback resistor Rf on the
voltage output of Fig 2 Dr. Hawksford. If we now add two resistors Rf to
ground from the voltage outputs the single ended transconductance is now
-R36/(Rf.R22), which is more reasonable ! (All this assumes we have enough
open loop gain in the first place). All this is theoretical, if we use a
feedback factor of 10 and Rf=4, we get gm = 2.5 S which is probably higher
than original gm of the transistor !
With Rf we retain all the linearity benefits of the global and local -ve feedback,
which I think was your intention + we should have all the benefits of a
current mode feedback op-amp, in theory
I still think you will need common mode feedback to set the dc on the current
output.
herisson
Hi Herisson,
Again thank you for your thoughtful comments and the data sheet references. I have looked at the data sheets you referenced but I haven't had time to study them thoroughly.
I am assuming a more modest gain of 4 or 5 and an output power of 5 to 10 watts. From the Zen articles it appears that there would be enough open loop gain available using IRFP240's.
My assumption was that the voltage output nodes would need to be loaded for stability purposes. It had not occurred to me that Rf would need to be as low as the value given in the Hawksford paper. The proposed circuit does not contain current mirrors, or use sensing of the current in the output node. I don't believe it needs to precisely "model" the load. My comparison to Hawksford was only in a general sense not a detail one.
Regards,
Graeme
Again thank you for your thoughtful comments and the data sheet references. I have looked at the data sheets you referenced but I haven't had time to study them thoroughly.
I am assuming a more modest gain of 4 or 5 and an output power of 5 to 10 watts. From the Zen articles it appears that there would be enough open loop gain available using IRFP240's.
My assumption was that the voltage output nodes would need to be loaded for stability purposes. It had not occurred to me that Rf would need to be as low as the value given in the Hawksford paper. The proposed circuit does not contain current mirrors, or use sensing of the current in the output node. I don't believe it needs to precisely "model" the load. My comparison to Hawksford was only in a general sense not a detail one.
Regards,
Graeme
Hi, Mr.Pass,
First Watt is a non-feedback commercial power amp.
You've decided to make it non-feedback. Must be a very strong reason why. From the review, I think the sound of non-feedback excel the feedback type (even 10W VS hundreds of watts). And another interesting fact is that the bass quality has nothing to do with Damping Factor (I always tought that's the reason for good bass).
Is it difficult to make a non-feedback power amp (with PassLab simplicity style, not with theory based complex cct) with bigger watt, like 100Watt? Or non feedback is only stable for small watt?
First Watt is a non-feedback commercial power amp.
You've decided to make it non-feedback. Must be a very strong reason why. From the review, I think the sound of non-feedback excel the feedback type (even 10W VS hundreds of watts). And another interesting fact is that the bass quality has nothing to do with Damping Factor (I always tought that's the reason for good bass).
Is it difficult to make a non-feedback power amp (with PassLab simplicity style, not with theory based complex cct) with bigger watt, like 100Watt? Or non feedback is only stable for small watt?
Designing a non-feedback amplifier becomes more difficult as the wattage increases. The voltage swing increases, which demands more of the parts that we have with available technology. The closer you push a part to its limits, the more feedback will be necessary to correct non-linearities.
That said, feedback games are much easier to play on tube equipment. My main tube amps are about 130W or so, and I have run them anywhere from zero feedback up to about 15 or 20dB, finally settling about 10dB. I have seen people argue that the low feedback used in tube amps is the key to their sound quality. Solid state pieces tend to have much, much higher feedback ratios.
What sound differences do you hear? Higher feedback tends to "sharpen" the highs, eventually to the point that they become overly etched. Some people like the effect, but it doesn't sound like real music. The bass will get somewhat tighter, but that's a diminishing returns sort of thing. On the downside, you start losing detail, such as imaging and ambience. People who listen to rock tend to want their highs and lows, and will frequently go for high feedback designs. Those who listen to classical, jazz, bluegrass, etc. often find low feedback designs persuasive due to their greater subtlety.
As a philosophical aside, I wonder how the stereo industry would have fared if rock had taken more of a folk-oriented turn, with acoustic guitars and so forth, rather than electric instruments. As it is, we have an entire generation of listeners who have grown up hearing music that's electronic from one end to the other. Since there's no reference to compare against, the sound has drifted away from the sound of "real" music. It's interesting to contemplate where the industry would have ended up if the music had made different demands on the playback equipment.
Grey
That said, feedback games are much easier to play on tube equipment. My main tube amps are about 130W or so, and I have run them anywhere from zero feedback up to about 15 or 20dB, finally settling about 10dB. I have seen people argue that the low feedback used in tube amps is the key to their sound quality. Solid state pieces tend to have much, much higher feedback ratios.
What sound differences do you hear? Higher feedback tends to "sharpen" the highs, eventually to the point that they become overly etched. Some people like the effect, but it doesn't sound like real music. The bass will get somewhat tighter, but that's a diminishing returns sort of thing. On the downside, you start losing detail, such as imaging and ambience. People who listen to rock tend to want their highs and lows, and will frequently go for high feedback designs. Those who listen to classical, jazz, bluegrass, etc. often find low feedback designs persuasive due to their greater subtlety.
As a philosophical aside, I wonder how the stereo industry would have fared if rock had taken more of a folk-oriented turn, with acoustic guitars and so forth, rather than electric instruments. As it is, we have an entire generation of listeners who have grown up hearing music that's electronic from one end to the other. Since there's no reference to compare against, the sound has drifted away from the sound of "real" music. It's interesting to contemplate where the industry would have ended up if the music had made different demands on the playback equipment.
Grey
GRollins said:
Today i read the review of the F1 through the link to 6Moons and i am pretty pleased to see that it does its job so well for the AKG K1000.
When AKG introduced it(1990i think) i was totally smitten with its looks, its red frame, even the silk layered mahogany box was a must-have. It was cheaper than the Stax towers and so much more comfortable.
As a bonus its analytical character made it a fetishist's dream.
A long time has passed since the days that i read Atti Mottala's findings, the works of Jean Hiraga and Akihiko Kaneda in l'Audiophile. A couple of times a year i was anxious going to Paris for getting my hands on another copy of L'Audiophile, later for La Nouvelle Revue du Son.
Nowadays my hair is grey, my kid thinks i am an old fart.
Then, air favored TIM less than HD, uneven harmonics less than even, a gradual decrease in harmonics more.
Today, still many manufacturers talk a lot about high power levels and low harmonic distortion figures.
15/20 years ago the Apogee Diva gave me the creeps, a Spectral amp made me nervous, Yamaha amplifiers were shear horror.
The solid state history, to me, is like the Hamburger.
MacDonalds conquered the world, even the Japanese resigned to the burger.
Now it is hard to get rid of the flying saucer, lately i suggested my kid making him a Mr Crab burger.
I am a Miles Davis, Chet baker, Luciano Pavarotti fan.
On the other hand ; ACDC, Japanese Kodo drummers, George Thorogood and Metallica still ring my bell too.
The punch Mr Davis can give to a sound from his instrument asks for an amplifier that can be dynamic, as does rock and heavy metal.
The voice of Mr Baker or Mr Pavarotti needs the ability for detail.
Dynamics need speed, detail needs accuracy.
I think Mr Pass describes it far better in his design philosophy of the X-2 than i.
To me the arguement is not quite fair. Solid state meant cheap power, push pull made it even cheaper.
What is a shame, is that the industry saw feedback as the overal cure, and did not bother to look at different approaches to obtain an amplifier that does the job best.
But some do.
BTW the dinamics produced by an F1(diy) and the Fostex FE206E are p r e t t y impressive . And for sure it is not an everyday system.It take some time for the mind to realize what is speed and what is detail.
Well this kind of amps and speaker do not push you to hurry.
In the past , crap systems made me nervous. So much crap are there in the market.
I must admit SOZ and the diyF1 (at least I believe it is) are my favourites amps . I like very much the absence of feedback.
But also a good x is good : the x feedback sounds far different from the almost totally amps out there.
Probably my preference goes to the differential pair in general but again the absence of feedback is a special planet.
Well this kind of amps and speaker do not push you to hurry.
In the past , crap systems made me nervous. So much crap are there in the market.
I must admit SOZ and the diyF1 (at least I believe it is) are my favourites amps . I like very much the absence of feedback.
But also a good x is good : the x feedback sounds far different from the almost totally amps out there.
Probably my preference goes to the differential pair in general but again the absence of feedback is a special planet.
It's not easy to isolate out all the elements that make up the
sound. Current Source is a big one, No feedback is another,
Pure Class A and Single-stage Mosfet are yet others. An
overlooked element is the difference between the 2nd vs 3rd
harmonic character of a single-ended (like the original Zen) and
differential (SOZ) topology.
Having built so many of the two varieties (SE vs Diff) and
listened and seen the reactions, I am beginning to conclude that
the world sees the 2nd and 3rd harmonic types as different
creatures. Audiophiles seem to have been conditioned to think
that they want 2nd harmonic, but they often pick 3rd as sounding
better in a given situation.
Much like red vs white wine, at one point or another you find
youself enjoying one instead of the other.
Of course I intend to explore this further, augmented by
the proper wine selection.
sound. Current Source is a big one, No feedback is another,
Pure Class A and Single-stage Mosfet are yet others. An
overlooked element is the difference between the 2nd vs 3rd
harmonic character of a single-ended (like the original Zen) and
differential (SOZ) topology.
Having built so many of the two varieties (SE vs Diff) and
listened and seen the reactions, I am beginning to conclude that
the world sees the 2nd and 3rd harmonic types as different
creatures. Audiophiles seem to have been conditioned to think
that they want 2nd harmonic, but they often pick 3rd as sounding
better in a given situation.
Much like red vs white wine, at one point or another you find
youself enjoying one instead of the other.
Of course I intend to explore this further, augmented by
the proper wine selection.
The question of second vs. third harmonics has been tainted by people saying that odd harmonics sound worse. As far as it goes, this is true. But higher order harmonics sound bad, also. Given a choice between the two, I choose low order (2nd or 3rd) harmonics over even order (in this case to include higher order even harmonics) any day of the week.
To be able to choose the relative harmonic structure is part of the "art of electronics" (ahem). Given that distortion is inevitable, you might as well do your best to shape it to be less disagreeable.
Those who are obsessed with very, very low distortion numbers get blindsided by the spectral distribution of the distortion that remains. Their equipment ends up sounding brittle; they in turn say "well, of course that's the way the music sounds...see, we've only got .0001% distortion, so it can't be our equipment." Only it doesn't sound like real music. That's not so difficult to figure out. All you have to do is go listen to listen to live, unamplified music. You don't even have to like the music, mind you, just think of it as one of those charts of letters they use to test your vision; boring, but a valid test, nonetheless.
Grey
EDIT:
I will nearly always choose a red wine over white, though I make allowances for Sauternes and the like. Can't stand Chardonnay. Other whites I take on a case by case basis. The question of headaches is interesting, in that there are regional myths that reds or more prone to inducing hangovers, or that whites are more prone to inducing hangovers. To my knowledge, no one has ever demonstrated a consistent tendency for either red or white wine to cause more hangovers.
To be able to choose the relative harmonic structure is part of the "art of electronics" (ahem). Given that distortion is inevitable, you might as well do your best to shape it to be less disagreeable.
Those who are obsessed with very, very low distortion numbers get blindsided by the spectral distribution of the distortion that remains. Their equipment ends up sounding brittle; they in turn say "well, of course that's the way the music sounds...see, we've only got .0001% distortion, so it can't be our equipment." Only it doesn't sound like real music. That's not so difficult to figure out. All you have to do is go listen to listen to live, unamplified music. You don't even have to like the music, mind you, just think of it as one of those charts of letters they use to test your vision; boring, but a valid test, nonetheless.
Grey
EDIT:
I will nearly always choose a red wine over white, though I make allowances for Sauternes and the like. Can't stand Chardonnay. Other whites I take on a case by case basis. The question of headaches is interesting, in that there are regional myths that reds or more prone to inducing hangovers, or that whites are more prone to inducing hangovers. To my knowledge, no one has ever demonstrated a consistent tendency for either red or white wine to cause more hangovers.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.