The beauty of triodes is that, if lightly loaded, in regions of ~constant mu they are in principle distortionless. For small signals the light loading is fairly easy to accomplish. For power amps, wretched excess always works, and is always energy-intensive.
Allen Wright used to talk about cathode followers being in need of extra help, and had advocated DMOS bootstrapped loads in the plate and so forth. While this further reduces the capacitance from plate to grid, it is not, generally, the lowest-distortion way to proceed.
Poor old cathode followers always seem to be abused. Maybe R. Cordell's (wonderful!) book will get them some relief. I remember reading long ago that women with large breasts were more likely to be mistreated (by men). Don't know what it means and don't know how it's related. Just a Zeitgeist thing maybe.
Vacuum tubes have a lot of performance advantages for linear amplifiers, but maybe the most important one is their tolerance for high temperatures. Who knew?~!
Always the best to you and yours,
Chris
ps: My younger and smarter friend has a blog/website: itisihifi.com
Big fun.
Vacuum tubes have a lot of performance advantages for linear amplifiers, but maybe the most important one is their tolerance for high temperatures. Who knew?~!
Always the best to you and yours,
Chris
ps: My younger and smarter friend has a blog/website: itisihifi.com
Big fun.
Last edited:
I made some tests using the usual sine and then adding harmonics to check their audibility, although speakers distorsion is quite high i got the surprise that not only 0.1% is clearly audible but that it sound horrible because these are not true musical harmonics...
It is extremely difficult to design a musical instrument that sound good because it must be shaped such that its H2, H3, H4 and so on are wrong in respect of the mathematical harmonics but right in respect of musical harmonics that are at a small distance of the mathematical ones.
By introducing these false harmonics one is litteraly reshaping the instrument caracteristics, tone controls have not the same effect as they increase the level of harmonics that are right musicaly speaking, so their effect amount to get at a closer distance of the player to better hear the higher frequencies of the instrument spectrum.
The test can be done with a well tuned piano, play a note that is low enough, say a C, as the sound slowly decay we can hear the musical harmonics, the first that will emerge is a G followed by a E and then by a B flat, all these harmonics are at a very small distance of the mathematical harmonics and the closer they are of the musical harmonics the better the instrument sound.
Badly manufactured instruments will produce too much perfect harmonics and will sound agressive and out of tune, and that s exactly the effect produced by high THD ratios, but for whom has not an ear musicaly trained it can sound better because of the higher tolerance to detuned sounds that are still perceived as right musicaly.
Wahab,
Single tone sine waves, or multiple?, a single tone sine wave is only conclusive for writing in your literature. It is well proven that a single tone is far more audible than the multiple complex waves of music, and the more complex and massed it becomes the higher the thd needs to be to be audible. To correlate thd directly to acoustic instruments is not entirely a fair comparision, as far as an acoustic instrument goes there is no thd, only complimentary resonances, and as far as that goes no two musicians have the same personal preference on absolute tonality.
Colin
Single tone sine waves, or multiple?, a single tone sine wave is only conclusive for writing in your literature. It is well proven that a single tone is far more audible than the multiple complex waves of music, and the more complex and massed it becomes the higher the thd needs to be to be audible. To correlate thd directly to acoustic instruments is not entirely a fair comparision, as far as an acoustic instrument goes there is no thd, only complimentary resonances, and as far as that goes no two musicians have the same personal preference on absolute tonality.
Colin
Last edited:
I made the test starting with a fundamental and then adding H3, H5 , H7.
I agree that in a musical signal it wouldnt be audible at this ratio.
As for instruments i made a mistake, H2 is no problem at all as this is exactly the same note as the fundamental, it s H3, H5, H6, H7 and in general any harmonic that is not a power of 2 that will be unmusical.
These are mathematical harmonics and are out of tune.
For instance and for a A that is at 880Hz H3 will be at 2640Hz while the musical harmonic is at 2637,02Hz and is a G note, the error is small and that s why H3 is considered as not harmfull.
H7 is at 3080Hz while the closer musical harmonics are at 3118.455Hz and that s why H7 is considered as a disaster.
Notice that an instrument will produce thoses out of tune harmonics along with the musicaly right harmonics, that s all the art of instruments manufacturers to minimize the relevant ones, and the reason why a Guarnerius or a Stradivarius sound better than most other violins.
Likewise with pianos Steinway did put the emphasis on a rich harmonic content that trade some musicality for brightness, at the opposite of Bosendorfer who have a much sweeter and consistent sound due to less of those false harmonics.
That s also the reason why accoustic instruments couldnt be emulated with old synthesisers set apart the very ones that had a lot of mathematical harmonics as sound signature, others instruments had to use a sample as basis to get a realistic result as a sample has the relevant harmonic structure that is impossible to create with additive synthesis unless you have 30 oscillators with induvidual VCFs for each note....
Unmusical? Some here are probably not as sensitive to pitch as others. But in the long run, we hardly need to accept equal temperament as the only game in town. Just intonation's days may be just around the corner, as we have the technology to make instruments that make "intelligent" decisions about the best just-intonational pitches in musical modulations---no need for 43-key-per-octave instruments a la Partch. Even if we go out to only the early primes and perhaps as an option omit 7, it could make a lot of difference once the ears get accustomed. Having a just major third would be nice, despite that, to conditioned ears and brains, it sounds flat (and there is an aversion to sounding flat, as opposed to a bit greater tolerance of playing sharp a lot [vide Kenny G]).
They are finding that in auditions by professional players, the much-vaunted difference between famous violins etc. and excellent modern ones vanishes. And I don't think it has anything to do with conformance with equal temperament versus rational frequency overtones.
The piano of course is a very specific and unusual example. I'd love to have the space and the money for a good one (I used to house-sit for a couple who had a nine-foot Bechstein), and the literature will never die, I hope.
It's remarkable to hear some old music on instruments with tuning for which the pieces were intended. Frescobaldi on period organs, various harpsichord works. And to this day one needs to remind some, even reputable musicologists, that J. S. Bach wrote The Well-Tempered Clavier, not The Equal-Tempered Clavier.
Accuracy?!
My 2 cents
about it:
Interpreting the correct physical concept, that's it: the output should be equal to the input.
Thus we get the desired sound from the sound board operator in the studio, right? "Maybe" is the correct answer 😕. You probably will not have the same acoustic studio and not the same speakers. Not even the same ears of the operator. The result can be frustrating, and for me it was. Then we can turn to answer correction features in a vain attempt to solve a problem of time (room acoustics) with amplitude (frequency response). Among other problems.
Your target may turn out to be the real reproduction of what the musicians are doing. This is a physical impossibility, and an offense to those who have some functional hearing!
For me the recordings are just a small sample of reality, and often doesn't deserve proper "correct" playback. Equalizers are allowed between the Orthodox people, but why not controllable distortion? For me, in SOME CASES, it serves as a "vaccine" against the shortcomings of certain recordings. I do not understand why the restriction of certain people. I seem to be the only one using both resources when needed, one or other in some cases, and none with good performance recordings. (as you fight out there, I laugh and test all things, kkkk😛😀🙂)
I have low THD amplifiers and also have a Single Ended "effects box" amp who pollute the signal less than the overwhelming majority of the recordings I know (and I am have 3TB audio from local pop recordings to ones said to "audiophile" and others "accurate"). True, some recordings really deserves the best linear amps.
Feedback? Nothing against it, but my personal preference is open loop class A stages, is very educational about devices and linearity. And more fun than trying to proper stabilize feedback amps.
All of this is only my opinion and preferences.
My 2 cents
about it:Interpreting the correct physical concept, that's it: the output should be equal to the input.
Thus we get the desired sound from the sound board operator in the studio, right? "Maybe" is the correct answer 😕. You probably will not have the same acoustic studio and not the same speakers. Not even the same ears of the operator. The result can be frustrating, and for me it was. Then we can turn to answer correction features in a vain attempt to solve a problem of time (room acoustics) with amplitude (frequency response). Among other problems.
Your target may turn out to be the real reproduction of what the musicians are doing. This is a physical impossibility, and an offense to those who have some functional hearing!
For me the recordings are just a small sample of reality, and often doesn't deserve proper "correct" playback. Equalizers are allowed between the Orthodox people, but why not controllable distortion? For me, in SOME CASES, it serves as a "vaccine" against the shortcomings of certain recordings. I do not understand why the restriction of certain people. I seem to be the only one using both resources when needed, one or other in some cases, and none with good performance recordings. (as you fight out there, I laugh and test all things, kkkk😛😀🙂)
I have low THD amplifiers and also have a Single Ended "effects box" amp who pollute the signal less than the overwhelming majority of the recordings I know (and I am have 3TB audio from local pop recordings to ones said to "audiophile" and others "accurate"). True, some recordings really deserves the best linear amps.
Feedback? Nothing against it, but my personal preference is open loop class A stages, is very educational about devices and linearity. And more fun than trying to proper stabilize feedback amps.
All of this is only my opinion and preferences.
It is exactly what I think, but only for class B amplifiers, where the audio output devices are "doing nothing" since it are with only a small quiescent current (low transconductance/gain). But I'd prefer to fix it at its roots, with more quiescent current (more class A), making the damping factor less dependent on feedback and more steadier during the signal excursion of the output devices. But if this will cause a sonic difference, only some tests will tell ...
For pure class A output devices excited by low impedance drivers stages, the stage will tame every loudspeaker EMF by itself, of course, in open loop amplifiers.
its much simpler to readjust your thinking: negative feedback is essential to having low Z, Class A devices too - from "100%" local feedback
negative feedback is the only practical means of achieving low output Z, high damping ratios, independence form back EMF ("back EMF" is another thinking impediment - its just the complex valued impedance of the moving cone acting through the electromechanical transformer = voice coil motor)
you can use output devices common emitter - the collector output is high Z even when biased fully Class A - triodes aren't a counter example - they do have internal local Plate V feedback - as evidenced by the mu factor
the only way to get low Z without any negative feedback is to have a shunt - for reasonable damping ratio it burns the damping ratio times the power delivered to the load - clearly not practical at audio power amp, loudspeaker load levels
under biased outputs have dead zones - feedback simply doesn't work when there's no forward gain
cut off transistor's local feedback isn't working either
negative feedback is the only practical means of achieving low output Z, high damping ratios, independence form back EMF ("back EMF" is another thinking impediment - its just the complex valued impedance of the moving cone acting through the electromechanical transformer = voice coil motor)
you can use output devices common emitter - the collector output is high Z even when biased fully Class A - triodes aren't a counter example - they do have internal local Plate V feedback - as evidenced by the mu factor
the only way to get low Z without any negative feedback is to have a shunt - for reasonable damping ratio it burns the damping ratio times the power delivered to the load - clearly not practical at audio power amp, loudspeaker load levels
under biased outputs have dead zones - feedback simply doesn't work when there's no forward gain
cut off transistor's local feedback isn't working either
Last edited:
Classic example is a valve guitar amplifier where "nice" distortion is required.
Clean doesn't sound very good.
Member
Joined 2009
Paid Member
Or if you configure your crossover to have constant impedance, then current or voltage drive doesn't matter. It's not really very hard (though big inductors can get expensive)
It was really awesome if a modern amplifier could be adjusted for the negative feedback on-the-go.
The cary 805 has this kind of adjustment, also some old tube amps like the Manley Ichiban.
This way we could listen to the tonal variation and adapt to the kind of effect we want to achieve, it's like choosing a triode operation with vocal music or a pentode for rock just by pressing a button 🙂
The cary 805 has this kind of adjustment, also some old tube amps like the Manley Ichiban.
This way we could listen to the tonal variation and adapt to the kind of effect we want to achieve, it's like choosing a triode operation with vocal music or a pentode for rock just by pressing a button 🙂
1) the GNF it’s used to low the distortion and to adjust the output impedance of the amplifier to an optimal defined level.
2) The GNF achieves determined curve responses with a goal to control the tonality.
3) in case 1, the GNF is independent from the frequency
4) in case 2, it depends on the frequency. So it will be stronger at the one that have to attenuate and weaker at the one that has to increase.
5) with the GNF it’s not possible to cancel ALL the distortion.
6) in these days (1962) the amplifiers use a value of 30 to 35dB of GNF, that means that the GNF signal it’s 96 to 98% of the input signal.
7) the GNF is not a medicine to make a bad amplifier into a good one, because it’s necessary that the circuit have excellent performance because the output signal has to be exactly at 180 degree out of phase with the input signal.
8) this out of phase signal depends also from the response curve. At the moment the response curve falls at the extremes of the band a out of phase is added to this 180 degree, so of it achieves 180 degrees the total will became 360 degrees, so instead of lowering it will increase at the input signal making the amplifier oscillate.
9) other important effect of the GNF is the reduction of the output impedance of the amplifier. This does not changes the output impedance load of the amplifier but changes the impedance provided by the speakers that follows to the amplifier. Th GNF can lower this impedance to zero keeping the undesirable cone excursions away.
10) the speakers do not only receive current from the amplifier, they also transmit.
11) if a huge impulse makes a big displacement of the cone, this will not return immediately to the initial rest position, and it will vibrate for some time according to the natural frequency.
12) this oscillation creates a tension proportional at the moving coil(making the speaker a generator) and with this at the secondary of the output transformer.
13) the lowest the impedance found by this current, the fastest the damping of the cone oscillation.
14) if the damping is too high, the bass reproduction could be weaker.
15) this is why there are amplifiers that we can adjust the damping factor to an optimal level, to match the speaker and the acoustic of the room we have.
16) there are 2 different systems of GNF that changes the distortion but that have different effects over the damping factor.
17) the first is the TENSION GNF. and the CURRENT GNF. With the second one, the output impedance is raised to an undesirable level.
18) by combining both, we can achieve the variable damping effect before told.
19) we can use the negative feedback at all the amplifier, or just in one of the stages, we can also combine the GNF with parcial ones.
20) also the hum of the amplifier is considerable reduced with the GNF.
21) if we include a capacitor or inductor to the GNF circuit we also make it sensible to the frequency.
22) the GNF used with capacitors or inductors to modify the response curves of the amplifier has some vantages over the ordinary tone control of the amplifiers, because it also decrease the distortion.
23) In every case the GNF is undoubtedly a method of improving the fidelity of the amplifier audio reproduction.
From HIFI-STEREO, 1962.
2) The GNF achieves determined curve responses with a goal to control the tonality.
3) in case 1, the GNF is independent from the frequency
4) in case 2, it depends on the frequency. So it will be stronger at the one that have to attenuate and weaker at the one that has to increase.
5) with the GNF it’s not possible to cancel ALL the distortion.
6) in these days (1962) the amplifiers use a value of 30 to 35dB of GNF, that means that the GNF signal it’s 96 to 98% of the input signal.
7) the GNF is not a medicine to make a bad amplifier into a good one, because it’s necessary that the circuit have excellent performance because the output signal has to be exactly at 180 degree out of phase with the input signal.
8) this out of phase signal depends also from the response curve. At the moment the response curve falls at the extremes of the band a out of phase is added to this 180 degree, so of it achieves 180 degrees the total will became 360 degrees, so instead of lowering it will increase at the input signal making the amplifier oscillate.
9) other important effect of the GNF is the reduction of the output impedance of the amplifier. This does not changes the output impedance load of the amplifier but changes the impedance provided by the speakers that follows to the amplifier. Th GNF can lower this impedance to zero keeping the undesirable cone excursions away.
10) the speakers do not only receive current from the amplifier, they also transmit.
11) if a huge impulse makes a big displacement of the cone, this will not return immediately to the initial rest position, and it will vibrate for some time according to the natural frequency.
12) this oscillation creates a tension proportional at the moving coil(making the speaker a generator) and with this at the secondary of the output transformer.
13) the lowest the impedance found by this current, the fastest the damping of the cone oscillation.
14) if the damping is too high, the bass reproduction could be weaker.
15) this is why there are amplifiers that we can adjust the damping factor to an optimal level, to match the speaker and the acoustic of the room we have.
16) there are 2 different systems of GNF that changes the distortion but that have different effects over the damping factor.
17) the first is the TENSION GNF. and the CURRENT GNF. With the second one, the output impedance is raised to an undesirable level.
18) by combining both, we can achieve the variable damping effect before told.
19) we can use the negative feedback at all the amplifier, or just in one of the stages, we can also combine the GNF with parcial ones.
20) also the hum of the amplifier is considerable reduced with the GNF.
21) if we include a capacitor or inductor to the GNF circuit we also make it sensible to the frequency.
22) the GNF used with capacitors or inductors to modify the response curves of the amplifier has some vantages over the ordinary tone control of the amplifiers, because it also decrease the distortion.
23) In every case the GNF is undoubtedly a method of improving the fidelity of the amplifier audio reproduction.
From HIFI-STEREO, 1962.
Last edited:
Jcx,
You can have open loop with low z, yes with local feedback, but if I'm wrong this thread was started with respect to a global loop negative feedback rhetorical question. Feedback is neccessary, whether it be local, degenerative or as I believe global as the very last resort of ultimate dependance, namely only to tie done as the final step to tie it neatly up together. What I'm more keen on is that does anyone have any ratios?, of course as with everything in life you can't have your cake and eat it too 😉.
Colin
Adaptive feedback in SMPSU is quite normal nowadays - but it has not found its way into mainstream audio yet - however the problems and operating requirements for PSU's are of course are totally different. For one, people spec'ing and designing PSU's will not brook any subjective b.s.
On my e-Amp I can use jumpers to select between MC, TPC and 'wide band' feedback (MC or TPC) where the loop gain is slightly lower, but extends to 40 kHz.
I have thought quite a bit about designing a power amp where you can switch between zero feedback, low feedback and high feedback but have not gotten around to it.
I suspect that in a good design, they 'may' sound different (ABX or DBT), but I somehow doubt that people will say 'Oh my dear what a calamity, with feedback the sound stage has collapsed, the music is flat . . . what a disaster'
On my e-Amp I can use jumpers to select between MC, TPC and 'wide band' feedback (MC or TPC) where the loop gain is slightly lower, but extends to 40 kHz.
I have thought quite a bit about designing a power amp where you can switch between zero feedback, low feedback and high feedback but have not gotten around to it.
I suspect that in a good design, they 'may' sound different (ABX or DBT), but I somehow doubt that people will say 'Oh my dear what a calamity, with feedback the sound stage has collapsed, the music is flat . . . what a disaster'
Last edited:
In my GainWire mk2 (pre amp with phone and line output, not power amp) it is possible to switch from high GNFB(CFA) to no GNFB, just to move one jumper. Some DIYers like GNFB and some like more non GNFB.
I could not decide wath to prefar more, could be old ears reason for that.
Also, in my experience, some loudspeakers heve less THD when driven by moderate impedance (eg.: DF=2). Some reduce at least by half, others become poorer. Beware of vented systems. Of course, a limit exists for proper damping the cones, but this limit is far higher for speakers units than this hi-end "high-DF" fallacy. Remember the coil dcr in the equation.
Even in simple tests like: Effects Of Source Impedance on Loudspeakers show the high limit needed for damage in impulse response. Others paper exists but I don't remeber them now...
And conventional feedback amps can be transformed:Variable Amplifier Impedance
Almost all conventional amps I have made in last years have this DF adjusting option.
So the "rock-solid low impedace" from high-NFB amps are not a cure for all diseases, after all...