If I understand this correctly, you are wondering aloud if you should use some form of sliding bias, rather than fixed point bias?
To this day, I have yet to hear an amp with any sliding bias scheme which will sound better because of that scheme. Every time I think I have just heard that, it turns out that the key reason is more likely something else, like an extra pair of output devices, a more powerful power supply, etc.
I have cme to believe that those arrangements are really a waste of time and an unnecessary complication.
On the other hand, 233 mA per output pair seems a bit too much - I could be wrong. I find most do what they do by around 130 mA per output pair, and moving on upwards just makes the sound softer and a little less focused, for lack of a better word. That too can be just as irritating as sharp and sibilant sound.
You really don't want to go the industry way, albeit in the opposite direction. They save by using low bias currents and thus have weedy heat sinks; you might use too much bias current and impose very restricting demands on the constructor regarding heat sinks (and I don't mean me, I also have 0.3 heatsinks).
As well as JVC s Super A amps , among other the awsome AX1/AX7 serie
and AX77..
The problem with most japanese specials , weak knees !
What about tube/MOSFET/bgt's ..... Best of all three ....!
What about tube/MOSFET/bgt's ..... Best of all three ....!
Hi,
Not sliding, but perhaps 2 or 3 Step, by either manual or automatic selection.
To burn off around 250W or more idle (with all losses in the chain) all the time seems a bit wasteful and anti-ecological (I just did my whole flat with LED Lights and cut the light power consumption massively compared to the dimmed 35/50W Halogen spots I would have used otherwise - strictly to be more green). Even my Tube Amp does not have that kind of idle consumption.
Could very well be.
However, I can likely get acceptable performance for every day use (watching movies, my daughter watching her cartoons, listening to music while working etc. which the Amp would be doung 70% of the time it is on, which tends from to be from 8am to well past midnight) at a fairly low bias (e.g. 80mA per pair and thus under 100W dissipation).
Yet I also feel that turning up the bias to the maximum I am willing to put up with thermally (in the interest of reliability) has benefits, something I know you agree with.
So either a "low/high" switch or a system that cranks up the bias when it gets loud may very well have merit outside strict audiophile concerns.
Note, the Amplifier is not meant as a general project for others, it is strictly a one off for my personal use, according to my personal ideas and concepts.
I am merely sharing the concepts and Ideas as they may benefit others even if they run counter to their views, as it allows us all to sharpen our wit in seeking good arguments to promote our own views and concepts in opposition, while others may find the ideas and concepts agreeable and benefit from them too, or may even like the whole design once done enough to replicate it wholesale.
Ciao T
Not sliding, but perhaps 2 or 3 Step, by either manual or automatic selection.
To burn off around 250W or more idle (with all losses in the chain) all the time seems a bit wasteful and anti-ecological (I just did my whole flat with LED Lights and cut the light power consumption massively compared to the dimmed 35/50W Halogen spots I would have used otherwise - strictly to be more green). Even my Tube Amp does not have that kind of idle consumption.
Could very well be.
However, I can likely get acceptable performance for every day use (watching movies, my daughter watching her cartoons, listening to music while working etc. which the Amp would be doung 70% of the time it is on, which tends from to be from 8am to well past midnight) at a fairly low bias (e.g. 80mA per pair and thus under 100W dissipation).
Yet I also feel that turning up the bias to the maximum I am willing to put up with thermally (in the interest of reliability) has benefits, something I know you agree with.
So either a "low/high" switch or a system that cranks up the bias when it gets loud may very well have merit outside strict audiophile concerns.
Note, the Amplifier is not meant as a general project for others, it is strictly a one off for my personal use, according to my personal ideas and concepts.
I am merely sharing the concepts and Ideas as they may benefit others even if they run counter to their views, as it allows us all to sharpen our wit in seeking good arguments to promote our own views and concepts in opposition, while others may find the ideas and concepts agreeable and benefit from them too, or may even like the whole design once done enough to replicate it wholesale.
Ciao T
Hi,
I think if we where to do that we would have to add a SIT somewhere and use laser biasing...
Don't get me wrong, I like hybrids as concepts, but if I am going to put up with tubes and all that implies, personally, I might as well use a tube amp.
My question is more to what degree the things I have learned with tube gear bear transplanting to solid state, with the aim of "good sound".
I'm perfectly prepared to fail, after all, I have a tube amp that works well and plays mostly loud enough...
The next after that may be to see if we can apply the same concepts to Class D...
Ciao T
I think if we where to do that we would have to add a SIT somewhere and use laser biasing...
Don't get me wrong, I like hybrids as concepts, but if I am going to put up with tubes and all that implies, personally, I might as well use a tube amp.
My question is more to what degree the things I have learned with tube gear bear transplanting to solid state, with the aim of "good sound".
I'm perfectly prepared to fail, after all, I have a tube amp that works well and plays mostly loud enough...
The next after that may be to see if we can apply the same concepts to Class D...
Ciao T
Last edited:
Yes , agree one can get good sound out of a total tube build and even thou they seem to add some kind of harmonic richness to the music there is a kind of drive delivered by SS amps that sound more like the source to me ......
Interesting to see your project of course I would like to scale up ...
Interesting to see your project of course I would like to scale up ...
Thorsten, let's talk some statistics.
About 99% of the current class AB amplifiers use less than 80 mA per output pair. Most are of no interest to people like they congregate here, we all want something better than average, average doesn't cut it for us.
Yes, I absolutely do agree that more bias, moving the output stage more towards class A, is beneficial to the sound we hear.
The bias switch scheme you mentioned is used in some form in my Marantz amps. In fact, at a point, you can barely discern a very low level "click" as it shifts into higher bias mode. All done with NTC devices. But, ...
No automatic system is foolproof. There will always be situations when the system misreacts, for whatever reason or set of cirumstances. While this may not be dangerous to anyone or the electronics, it never fails to **** me off. I just can't stand "ifs", "buts" and "maybes", not if I can help it. Less variables by default means better reliability.
Thus, having played with this effect, admittedly in all cases with factory made units, I have come to the conslusion that about 130 mA per output pair is the highest point at which I can actually HEAR the difference. And that's the point at which I call it a day. In aid of your concept, most gave up by about 100 mA, only the HK 680, the one with just 12 dB of global feedback, went on to 130 mA; at 140 mA, I could not hear any differences at all. So I reduced it back to 130 mA and left it there. Not because I had to, it has a massive heat sink which could ride out more, but because I heard no additional benefit.
True, it's a factory made unit, so it would probably require some modification do fare better at higher levels, because as was done, I really shifted the work points of the predriver, drivers and output transistors - NOT a terribly good idea, but far easier than modifying the whole thing. In other words, if you design straight off for it, obviously you will have better results, no doubt there.
You will of course do as you please, but I would advise you keep it on a constant level, even if it is elevated in comparison with the average. Keep it simple and ellegant, is what I say.
Use the minimum you have to use to achieve your goals, but no less. And no more.
About 99% of the current class AB amplifiers use less than 80 mA per output pair. Most are of no interest to people like they congregate here, we all want something better than average, average doesn't cut it for us.
Yes, I absolutely do agree that more bias, moving the output stage more towards class A, is beneficial to the sound we hear.
The bias switch scheme you mentioned is used in some form in my Marantz amps. In fact, at a point, you can barely discern a very low level "click" as it shifts into higher bias mode. All done with NTC devices. But, ...
No automatic system is foolproof. There will always be situations when the system misreacts, for whatever reason or set of cirumstances. While this may not be dangerous to anyone or the electronics, it never fails to **** me off. I just can't stand "ifs", "buts" and "maybes", not if I can help it. Less variables by default means better reliability.
Thus, having played with this effect, admittedly in all cases with factory made units, I have come to the conslusion that about 130 mA per output pair is the highest point at which I can actually HEAR the difference. And that's the point at which I call it a day. In aid of your concept, most gave up by about 100 mA, only the HK 680, the one with just 12 dB of global feedback, went on to 130 mA; at 140 mA, I could not hear any differences at all. So I reduced it back to 130 mA and left it there. Not because I had to, it has a massive heat sink which could ride out more, but because I heard no additional benefit.
True, it's a factory made unit, so it would probably require some modification do fare better at higher levels, because as was done, I really shifted the work points of the predriver, drivers and output transistors - NOT a terribly good idea, but far easier than modifying the whole thing. In other words, if you design straight off for it, obviously you will have better results, no doubt there.
You will of course do as you please, but I would advise you keep it on a constant level, even if it is elevated in comparison with the average. Keep it simple and ellegant, is what I say.
Use the minimum you have to use to achieve your goals, but no less.
And no more. It just struck me that Thorsten and I have come a full circle, in all this talk of bias levels.
Thorsten says more bias reduces high frequency distortion, which ellegantly brings us back to the basic question - yes, it can be mesured, he did measure it, but exactly how relevant is this to what we hear?
Will reducing 50 kHz distortion from say 0.4% to say 0.2% actually yield better sound as we hear it? Will we HEAR the difference?
An interesting question. I suppose the only way to know for sure is to make a sample and try both versions, say 130 and 230 mA, listen to what happens under actual, real life conditions.
I THINK (but do not know for sure) that 230 mA will not yield any audible improvements over 130 mA, or, if it does, it will be a minscule improvement. But as I sad, the only way to KNOW is to try it.
Thorsten says more bias reduces high frequency distortion, which ellegantly brings us back to the basic question - yes, it can be mesured, he did measure it, but exactly how relevant is this to what we hear?
Will reducing 50 kHz distortion from say 0.4% to say 0.2% actually yield better sound as we hear it? Will we HEAR the difference?
An interesting question. I suppose the only way to know for sure is to make a sample and try both versions, say 130 and 230 mA, listen to what happens under actual, real life conditions.
I THINK (but do not know for sure) that 230 mA will not yield any audible improvements over 130 mA, or, if it does, it will be a minscule improvement. But as I sad, the only way to KNOW is to try it.
HiFi no , life like realism ......True dat ..........
For life like realism, good buddy, you also need to have a front end capable of it, and a back end (speakers) capable of reproducing something like 114 dB SPL in peaks.
And the size and furnishing of your room are also highly relevant factors.
Personally, I just made it. I have a small room, some 14*12 feet, fairly full of everything. My speakers do 92 dB/2.83V/1 m, and my Karan amp is rated at 180/250W into 8/4 Ohms. 180W/8 Ohms works out to 22 dBW, plus a single speaker's efficiency of 92 dB equals 114 dB. Two speakers in tandem will of course increase the actual SPL, but the room will soak up some of it, so let's assume I'm even on that point.
My point is, I know from first hand experience excatly what you are saying. But I must now dissapoint you - in real life, peaks of 114 dB are EXTREMELY rare, and I do believe 99.9% of all audiophiles have system which don't even get near that, yet play music very well. Also, both the HK 680 integrated (rated at 85/130 W into 8/4 Ohms) and the Marantz 170 DC power amp (rated at 85/8 Ohms) are extremely convincing in ANY reproduction.
Ah, if it all boiled down just to power ...
Agreed - but the key word here is "if".
We all know that may happen, but we also all know it shouldn't happen. What's a CD/SACD player, or FM tuner, doing passing out 50 kHz signals? Should the amplifier now be charged with duties of mopping up after some other lemon in the system?
Dejan,
Let's, I did major work on stats for my second degree.
Let me give some stats of my own.
First, 99% of all statistics are made up on the spot...
Second, I only trust statistics I have faked myself...
They probably also use more than 0.22R emitter resistors.
I cut my teeth in my original praktikum on complex electronics intended to help fools who would neither train adequately nor Read The Fantastic Manual (RTFM) to still hit the target. After that I spend time in inductrail electronics (operators are usually fools and if you do not block them the costs of foolishness are high) before moving into pro audio and eventually into financial computer systems (you would not believe the fools you find in accounting)...
I may not be able to make the system foolproof, but heck, I'm the operator and I ain't no fool, just a bit "schräg"...
I do not disagree, but as we do not live in ideal world...
I think I go 2-level and use a lockable (my 2 YO Daughter otherwise may throw a spanner) toggle on the front panel.
My goals are differentiated and multidimensional... It may take more than simple bloodymindedness (I am good at simple bloodymindedness though).
Ciao T
Let's, I did major work on stats for my second degree.
Let me give some stats of my own.
First, 99% of all statistics are made up on the spot...
Second, I only trust statistics I have faked myself...
They probably also use more than 0.22R emitter resistors.
I cut my teeth in my original praktikum on complex electronics intended to help fools who would neither train adequately nor Read The Fantastic Manual (RTFM) to still hit the target. After that I spend time in inductrail electronics (operators are usually fools and if you do not block them the costs of foolishness are high) before moving into pro audio and eventually into financial computer systems (you would not believe the fools you find in accounting)...
I may not be able to make the system foolproof, but heck, I'm the operator and I ain't no fool, just a bit "schräg"...
I do not disagree, but as we do not live in ideal world...
I think I go 2-level and use a lockable (my 2 YO Daughter otherwise may throw a spanner) toggle on the front panel.
Use the minimum you have to use to achieve your goals, but no less.
My goals are differentiated and multidimensional... It may take more than simple bloodymindedness (I am good at simple bloodymindedness though).
Ciao T
Hi,
Good question. I use traditional terminology to express radical ideas.
My prior point was that what we "hear" are not individual harmonics but in essence the transfer function composite. We are gaining much better understanding of the concept.
For example, a "transfer function curvature" which gives rise to only low order harmonics is not objectionable, if at all audible, yet an opposite one may have much lower THD and yet be much more objectional on audibility, DESPITE the individual harmonics lying below the system (including listening room) noisefloor.
So I am rendering a newish concept in traditional terminology, rather than following the lead of L Ron & Scientology in just making up new volcabulary.
Ciao T
Good question. I use traditional terminology to express radical ideas.
My prior point was that what we "hear" are not individual harmonics but in essence the transfer function composite. We are gaining much better understanding of the concept.
For example, a "transfer function curvature" which gives rise to only low order harmonics is not objectionable, if at all audible, yet an opposite one may have much lower THD and yet be much more objectional on audibility, DESPITE the individual harmonics lying below the system (including listening room) noisefloor.
So I am rendering a newish concept in traditional terminology, rather than following the lead of L Ron & Scientology in just making up new volcabulary.
Ciao T
Hi,
I use Non-Filtered DAC's with sample rates up to 192KHz (but thankfully for the Amp's not SACD with around +10dB RF noise compared to signal levels) and LP. DMM LP's have a very high frequency pilot tone to make cutting easier, it is near -6dB IIRC.
Ciao T
I use Non-Filtered DAC's with sample rates up to 192KHz (but thankfully for the Amp's not SACD with around +10dB RF noise compared to signal levels) and LP. DMM LP's have a very high frequency pilot tone to make cutting easier, it is near -6dB IIRC.
Ciao T
Hmmm. Spice of the modified Hafler shows a drop at 50K with .154 V in from .19 to .17% At full power, ( 1.54 in) drop from .11 to .09%. At 1K and .154 in, it is some difference between .0000 something which I consider highly incorrect in the real world and most likely irrelevant. This is from the 110 it is set at to 250 ma. Audible? I plan on this test next time I set up for my wife to listen. She has not heard my Creek in the test. I am most interested in that. The .4 and .2 numbers are probably a lot more realistic.
Hi,
FM Tuners have 19KHz pilot at very high levels and a barely suppressed 38Khz pilot that makes it through the Stereo decoder.
DMM LP's utilise a > 50KHz pilot at near full level to make cutting in solid copper easier (kind of like these vibrating shavers).
SACD has around 10dB more HF noise than signal.
Traditionally filtered CD lacks high frequency stuff, remove the filter, another story...
Ciao T
FM Tuners have 19KHz pilot at very high levels and a barely suppressed 38Khz pilot that makes it through the Stereo decoder.
DMM LP's utilise a > 50KHz pilot at near full level to make cutting in solid copper easier (kind of like these vibrating shavers).
SACD has around 10dB more HF noise than signal.
Traditionally filtered CD lacks high frequency stuff, remove the filter, another story...
Ciao T
Some years ago, I recall Keith Johnson warning, at a presentation about DSD at an AES convention when he got the microphone, that these signals were going to get a lot of amplifiers in trouble. And at that, the Spectral amps were famously wide bandwidth (do I recall 2MHz? You could have a pirate radio AM station if you had room for a decent antenna ).
).- Status
- Not open for further replies.