geeez ....
i'm sorry i even mentioned the article now. it appears to have had the opposite effect to what i was hoping for
silly me, after years of hanging out here, you think i would have learned this by now ...
i'm sorry i sent us off on a tangent. can we get back to the tech discussion now?
i'm still looking forward to understanding more about how the experts build good sounding, non-overall feedback preamps so i can build something and see what i think for myself.
thanks.
mlloyd1
who promises he won't "fart in the theater" anymore
i'm sorry i even mentioned the article now. it appears to have had the opposite effect to what i was hoping for
silly me, after years of hanging out here, you think i would have learned this by now ...
i'm sorry i sent us off on a tangent. can we get back to the tech discussion now?
i'm still looking forward to understanding more about how the experts build good sounding, non-overall feedback preamps so i can build something and see what i think for myself.
thanks.
mlloyd1
who promises he won't "fart in the theater" anymore
John,
I believe one poster said he had heard one of Charles's designs and didn't care for it. The conclusion he seems to have reached was that all no-feedback designs--including yours--were junk.
That's okay. I heard a high feedback design once and didn't like it. So by the same so-called logic, all high feedback designs are junk, right?
I've used feedback in the past. I reserve the right to use feedback now and in the future--however--I am currently trying to limit the amount I use in circuits, as it seems to detract from imaging (which actually does happen to be one of my goals). To that end, I am interested in what you have to say about open loop circuits.
Ditto for Charles Hansen.
Nelson has a way of approaching circuits that differs from yours and Charles's. It generally, though not always, involves feedback. That's okay by me. I'm interested what he has to say, too.
What most emphatically doesn't interest me is endless, repetitive nattering in the "feedback is the only way to Truth and Enlightenment" rut. It's a religion. Been there. Done that. Did the intervention thing and deprogrammed my brain. Any idiot can make a circuit with multiple zeros in the distortion specifications with enough feedback. That's nothing to brag about. I find the idea of seducing the circuit into giving me what I want far more interesting than hitting it over the head with a club. That approach is for cavemen.
And to think they brag about it!
"Ugh! Me caveman! Me hit circuit with club to make better. Oops! Crushed case. Broke circuit. Obviously need bigger club next time!"
Sorry guys. Go back to scratching on the wall of the cave with a lump of charcoal.
For the record, the power JFET circuits I'm working on use feedback. I'm still on the learning curve and have not achieved the wide bandwidth, etc. I want in an open loop configuration. Bummer. I'm still working on it.
John, my position is that neither you, nor Charles, nor anyone else should have to put up with rude people.
Period.
This thread isn't "yours" in the sense that you didn't personally start it. However, it is about one of your designs and accordingly you should be allowed to speak without having to endure a bunch of yik-yak from the peanut galleries. If they want to extol the presumed virtures of high feedback, let them start another thread. I, for one, don't want them polluting this one.
Grey
I believe one poster said he had heard one of Charles's designs and didn't care for it. The conclusion he seems to have reached was that all no-feedback designs--including yours--were junk.
That's okay. I heard a high feedback design once and didn't like it. So by the same so-called logic, all high feedback designs are junk, right?
I've used feedback in the past. I reserve the right to use feedback now and in the future--however--I am currently trying to limit the amount I use in circuits, as it seems to detract from imaging (which actually does happen to be one of my goals). To that end, I am interested in what you have to say about open loop circuits.
Ditto for Charles Hansen.
Nelson has a way of approaching circuits that differs from yours and Charles's. It generally, though not always, involves feedback. That's okay by me. I'm interested what he has to say, too.
What most emphatically doesn't interest me is endless, repetitive nattering in the "feedback is the only way to Truth and Enlightenment" rut. It's a religion. Been there. Done that. Did the intervention thing and deprogrammed my brain. Any idiot can make a circuit with multiple zeros in the distortion specifications with enough feedback. That's nothing to brag about. I find the idea of seducing the circuit into giving me what I want far more interesting than hitting it over the head with a club. That approach is for cavemen.
And to think they brag about it!
"Ugh! Me caveman! Me hit circuit with club to make better. Oops! Crushed case. Broke circuit. Obviously need bigger club next time!"
Sorry guys. Go back to scratching on the wall of the cave with a lump of charcoal.
For the record, the power JFET circuits I'm working on use feedback. I'm still on the learning curve and have not achieved the wide bandwidth, etc. I want in an open loop configuration. Bummer. I'm still working on it.
John, my position is that neither you, nor Charles, nor anyone else should have to put up with rude people.
Period.
This thread isn't "yours" in the sense that you didn't personally start it. However, it is about one of your designs and accordingly you should be allowed to speak without having to endure a bunch of yik-yak from the peanut galleries. If they want to extol the presumed virtures of high feedback, let them start another thread. I, for one, don't want them polluting this one.
Grey
GRollins said:
You probably mean me, though not named me. If yes, then you are wrong. The conclusion is the one that you have made, not me, I said nothing like this.
On the other hand I read here in several posts that more feedback means worse sound and that no feedback is the best. I do not agree with it and I do not agree with reasoning supporting this point of view. That is what I am saying.
Folks, it has been my experience that high negative feedback is not as good sounding as low or almost zero negative feedback, ALL ELSE BEING EQUAL! My opinion on this is based on my experience, rather than my initial assumptions. I used to think that negative feedback was the greatest thing going. It just did not appear to be so, once I actually experienced listening situations where negative feedback (or at least practical examples of it) seemed to lose out to more open loop designs. Then, I read Matti Otala's TIM papers and it started to make sense. Before that, the late Richard Heyser had told me that negative feedback was very problematic in audio circuits. He based this around his own listening tests.
On paper, at least at the undergraduate engineering level, negative feedback looks almost like the cure for almost everything. I have designed circuits with both positive and negative feedback in order to contour the output to any impedance that I want. Sometimes, current output is preferred over voltage feedback in motor drives.
Back in the late '60's while at the Ampex research department, I made designs that make most offerings here, simplistic. How about a 4 quadrant balanced bridge 2000W power amplifier with both positive and negative feedback loops in order to create a floating current drive, using multiple 2N5684's and 6's on an air blown heat sink? I really committed myself to feedback back then.
It was only when I LISTENED to what I and others preferred, that I realized that something was wrong. First, I thought it was TIM. Now I think it is PIM that is the deepest problem. This is just my opinion based on my experience. I let the marketplace tell me if I am going in the right direction.
On paper, at least at the undergraduate engineering level, negative feedback looks almost like the cure for almost everything. I have designed circuits with both positive and negative feedback in order to contour the output to any impedance that I want. Sometimes, current output is preferred over voltage feedback in motor drives.
Back in the late '60's while at the Ampex research department, I made designs that make most offerings here, simplistic. How about a 4 quadrant balanced bridge 2000W power amplifier with both positive and negative feedback loops in order to create a floating current drive, using multiple 2N5684's and 6's on an air blown heat sink? I really committed myself to feedback back then.
It was only when I LISTENED to what I and others preferred, that I realized that something was wrong. First, I thought it was TIM. Now I think it is PIM that is the deepest problem. This is just my opinion based on my experience. I let the marketplace tell me if I am going in the right direction.
John,
I arrived at a similar conclusion simply by wondering how it was that tube equipment with mediocre specifications frequently sounded better than solid state with oodles of zeros in the distortion department. Clearly the numbers didn't tell the whole story.
It is my hope that some clever lad (or lass) will come along and develop a spec that means something.
As of this writing I am unaware of anyone even trying--much less succeeding--at developing a specification that gives even a hint as to whether a piece of equipment images well. Yet imaging exists. Even a modest system can throw an image of some sort, at least width if not depth. Unfortunately, there seems to be an inverse relationship between "good" distortion measurements and a good image. My impression is that it is due to phase relationships being destroyed.
I measure distortion, but only as a casual thing. Until we have a more inclusive suite of distortion measurements, I'm not going to get all sweaty over a few points of THD or IMD. They just aren't enough to get the job done.
Something I'd like to see is a spec that measures real-time, frequency and level-dependent phase shifts. (For signal level, think of a standard Bode diagram in relation to dynamic level shifts--not steady state. I'm not sure what to do about frequency-related phase shifts, although I suspect that it is more critical to imaging in the long run.) I suspect that might get me closer to an understanding of what's happening in the imaging department.
Then there's dynamics and "tightness" in the bass (no, damping factor doesn't explain it, else why do some tube pieces whip solid state ones?) and about a dozen other things.
Grey
I arrived at a similar conclusion simply by wondering how it was that tube equipment with mediocre specifications frequently sounded better than solid state with oodles of zeros in the distortion department. Clearly the numbers didn't tell the whole story.
It is my hope that some clever lad (or lass) will come along and develop a spec that means something.
As of this writing I am unaware of anyone even trying--much less succeeding--at developing a specification that gives even a hint as to whether a piece of equipment images well. Yet imaging exists. Even a modest system can throw an image of some sort, at least width if not depth. Unfortunately, there seems to be an inverse relationship between "good" distortion measurements and a good image. My impression is that it is due to phase relationships being destroyed.
I measure distortion, but only as a casual thing. Until we have a more inclusive suite of distortion measurements, I'm not going to get all sweaty over a few points of THD or IMD. They just aren't enough to get the job done.
Something I'd like to see is a spec that measures real-time, frequency and level-dependent phase shifts. (For signal level, think of a standard Bode diagram in relation to dynamic level shifts--not steady state. I'm not sure what to do about frequency-related phase shifts, although I suspect that it is more critical to imaging in the long run.) I suspect that might get me closer to an understanding of what's happening in the imaging department.
Then there's dynamics and "tightness" in the bass (no, damping factor doesn't explain it, else why do some tube pieces whip solid state ones?) and about a dozen other things.
Grey
john curl said:
John,
I believe you, but nothing was said about circuit inside a closed loop. How did OLG look like? Was it wideband, with cutoff frequency high above audio band, or was it dominant pole design with integrating OLG decay, with cutoff somewhere like 100Hz? This is very important for PIM. If you only say that "ALL ELSE BEING EQUAL!" with no description of the circuit, we do not know much.
I know Ottala's paper well, but we live in different times now. It is not so big problem to design much faster circuits that are not significantly slew-rate limited.
According to my experience, global feedback is no evil for fast, wideband, locally linearized circuitry. I also made similar experiment - with very linear class A output stage inside and outside global feedback, everything else same. For me, it was sonically better inside feedback loop.
Some people like Porsches, some like Ferraris. Everyone takes what he likes. Both are great cars. What is there to argue ?
And what is one trying to achieve in going to a Porsche Owners Meeting with a Ferrari, trying to convince the other guys that their beloved Porsches are not as good ?
Start a new thread if you want to discuss the pros and cons of feedback. I am sure you would get great response.
Patrick
And what is one trying to achieve in going to a Porsche Owners Meeting with a Ferrari, trying to convince the other guys that their beloved Porsches are not as good ?
Start a new thread if you want to discuss the pros and cons of feedback. I am sure you would get great response.
Patrick
Grey, I had the same experience with the Grateful Dead back in 1970-71. They originally used Ampex tube mixers and Mac tube power amps. They wanted me to help them update to solid state. Well, most of our efforts for the first few years was essentially a failure. Open loop tubes actually sounded better than my IC op amp designs. It wasn't until I read Otala's papers that I was able to make successful solid state products with high open loop bandwidth, high slew rate and reasonably low distortion that pleased the band. Later, Mark Levinson built and used these circuits for the JC-2 preamp.
Nice article, proving that there is nothing like "pleasant distortion", "harmonic fullness" etc.
http://www.stereophile.com/reference/406howard/
http://www.stereophile.com/reference/406howard/
John,
Something I have found curious is that a solid state piece might take an open loop bandwidth of 200-300kHz to acheive similiar sonic qualities to a tube piece that has OL bandwidth of perhaps 75-100kHz. Granted, we're talking inherent differences between the circuits, but shooting from the hip I would say that, if anything, it should be the other way around, with the tube piece requiring greater bandwidth in order to push (coupling capacitor and/or transformer) phase shifts further out of the audible frequency range.
Got any ideas on that?
Grey
Something I have found curious is that a solid state piece might take an open loop bandwidth of 200-300kHz to acheive similiar sonic qualities to a tube piece that has OL bandwidth of perhaps 75-100kHz. Granted, we're talking inherent differences between the circuits, but shooting from the hip I would say that, if anything, it should be the other way around, with the tube piece requiring greater bandwidth in order to push (coupling capacitor and/or transformer) phase shifts further out of the audible frequency range.
Got any ideas on that?
Grey
Grey, this is a fairly narrow conclusion, but I would look at the effect of non-linear capacitance of the active devices for an answer. It is well known that transistors and fets have a lot of it, and this would be the primary limitation to open loop bandwidth. Less 'effective' capacitance means higher bandwidth.
Blowtorch in Europe
Hi Folks,
I've enjoyed reading the thread. I have just acquired a Blowtorch, something I've coveted for some time - ever since I heard Bob Crump's unit at his home. Nothing that I have ever heard equals the musical honesty of this preamp.
Early on, I noticed that several folks had observed that there apparently were no Blowtorchs in Europe. Well, that MAY change. I'd estimate roughly 50:50 odds that there will be one in London this year - mine. This would be my second time working in the UK, and is something I'm hoping works out. We shall see. Regardless, just wanted you to know that those in the area towards the end of the year would be welcome to drop by if they wish.
Mike
Hi Folks,
I've enjoyed reading the thread. I have just acquired a Blowtorch, something I've coveted for some time - ever since I heard Bob Crump's unit at his home. Nothing that I have ever heard equals the musical honesty of this preamp.
Early on, I noticed that several folks had observed that there apparently were no Blowtorchs in Europe. Well, that MAY change. I'd estimate roughly 50:50 odds that there will be one in London this year - mine. This would be my second time working in the UK, and is something I'm hoping works out. We shall see. Regardless, just wanted you to know that those in the area towards the end of the year would be welcome to drop by if they wish.
Mike
- Status
- Not open for further replies.