Self's article (EW march 1999) was reproduced p404 to p420 in his book "Self on audio" published in 2006, .
Re: To tell ya the truth guys, i feel very confused about those things
The difference is purely how much bias. All push-pull amps class AB
amplifiers operate class A up to some output current figure, generally
twice the bias current, and beyond that they are class AB (or B
depending on who you talk to).
Virtually all push-pull class A amplifiers become class AB (or B) at some
high output current.
destroyer X said:
The difference is purely how much bias. All push-pull amps class AB
amplifiers operate class A up to some output current figure, generally
twice the bias current, and beyond that they are class AB (or B
depending on who you talk to).
Virtually all push-pull class A amplifiers become class AB (or B) at some
high output current.
Now things are clarified to me.
Thank you Nelson, to put a ligth into the darkness i had in my mind..... all class A will go to AB, or B....... this is clear now.
No more confused.
Regards,
Carlos
Thank you Nelson, to put a ligth into the darkness i had in my mind..... all class A will go to AB, or B....... this is clear now.
No more confused.
Regards,
Carlos
I guess the next question is does it produce any audible benefit. Could be a good experiment, take an amp and reduce the bias to the point distortion becomes audible, then leaving the bias at this level add a current sink to the neg rail. Is it less objectionable.
I had always understood that if current flowed for 180 degress of the conduction cycle in a push pull pair it was class B, anything less it's class C (RF amps and the like), and anything more and it's AB up to conduction for 360 degrees when it becomes class A.
I had to smile looking at the "XD document" as the audio precision plots appear to show the "gain step" for class XD to be just as "bad" if thats the word as class AB. The class B distortion plot is the smoothest if a little higher at low amplitude.
Regards Karl
I had always understood that if current flowed for 180 degress of the conduction cycle in a push pull pair it was class B, anything less it's class C (RF amps and the like), and anything more and it's AB up to conduction for 360 degrees when it becomes class A.
I had to smile looking at the "XD document" as the audio precision plots appear to show the "gain step" for class XD to be just as "bad" if thats the word as class AB. The class B distortion plot is the smoothest if a little higher at low amplitude.
Regards Karl
My spectations is that will sound better into low volumes because Class A
No doubts, at least to me, that Class A sounds better.... the heat is the problem.
The difference you can see into the published graphics are into a distortion level that i cannot listen... say..... both distortions levels are something i just cannot listen.... so, invisible, inaudible, unlistenable difference to my own possibilities.
Something alike an amplifier that can go to 100 Megahertz and another that go to 50 Megahertz.... i cannot listen both frequencies, and the influences the faster one will provide into my audibility limit of 17.000 Hertz (I am 57 years old) is very small to be perceived.
Well, i have to listen, and i am crazy to listen this idea from Doctor Self, and i remember Graham Maynard made something that also could keep crossover distortions small into high volumes... and THAT i can say sounds excelent... the problem continues, the heat!
Into your environment, to North America and Europe, where average temperature is lower, hot amplifiers are not a problem... i keep windows full opened and 2 enormous fans pump air around the clock here in my place, where 32 degrees is not uncommon and very unconfortable... i am still thinking to substitute my PC tube monitor by a LSD just to remove one hot spot here in my home....so, the most i can love Class A, i have not the possibility to use it.... or i install my Class A amplifier outside home or i have to remove myself from home...there's no conditions to have both together here.
Any method to reduce heat dissipation will be welcome into my home.... i think to the big majority of South Americans, beeing the exception people from Andes and extreme south of my country, in places the average temperature goes to 18 degrées celsius almost even during summer, at nigth in special can go down to 5 degrées celsius and in some towns temperature reach zero degrees,with some ice and some snow (rarity even to our south borders),
Brazilian Gradiente made Super A, licensed by JVC down the seventies, but despite my people loves the sonics i felt it mufled, and used the strategy to reduce the bias (class A bias) when the audio level increased.... i perceived alike a sound that had to cross a pillow...thousands here felt it excelent.... sadly i cannot say the same.
regards,
Carlos
No doubts, at least to me, that Class A sounds better.... the heat is the problem.
The difference you can see into the published graphics are into a distortion level that i cannot listen... say..... both distortions levels are something i just cannot listen.... so, invisible, inaudible, unlistenable difference to my own possibilities.
Something alike an amplifier that can go to 100 Megahertz and another that go to 50 Megahertz.... i cannot listen both frequencies, and the influences the faster one will provide into my audibility limit of 17.000 Hertz (I am 57 years old) is very small to be perceived.
Well, i have to listen, and i am crazy to listen this idea from Doctor Self, and i remember Graham Maynard made something that also could keep crossover distortions small into high volumes... and THAT i can say sounds excelent... the problem continues, the heat!
Into your environment, to North America and Europe, where average temperature is lower, hot amplifiers are not a problem... i keep windows full opened and 2 enormous fans pump air around the clock here in my place, where 32 degrees is not uncommon and very unconfortable... i am still thinking to substitute my PC tube monitor by a LSD just to remove one hot spot here in my home....so, the most i can love Class A, i have not the possibility to use it.... or i install my Class A amplifier outside home or i have to remove myself from home...there's no conditions to have both together here.
Any method to reduce heat dissipation will be welcome into my home.... i think to the big majority of South Americans, beeing the exception people from Andes and extreme south of my country, in places the average temperature goes to 18 degrées celsius almost even during summer, at nigth in special can go down to 5 degrées celsius and in some towns temperature reach zero degrees,with some ice and some snow (rarity even to our south borders),
Brazilian Gradiente made Super A, licensed by JVC down the seventies, but despite my people loves the sonics i felt it mufled, and used the strategy to reduce the bias (class A bias) when the audio level increased.... i perceived alike a sound that had to cross a pillow...thousands here felt it excelent.... sadly i cannot say the same.
regards,
Carlos
Here is the schematic from that famous Gradiente i have mentioned
R21, Q11 and R24 are in series, but Q11 is a variable resistance triggered by the voltage that is also going to the power output driver transistors (Q14 and Q15)
Also R19, Q7 and R23 works the same way triggered by the audio signal that goes to the output drivers.
Q5 and Q8 will have their colector to emitter resistances changed by Q11 and Q7.
All that stuff means three resistances in parallel.... Q11 circuit is one of them, and variable resistance triggered by the signal, also Q7 is a variable resistance triggered by the audio signal.... and the one has almost fixed resistance (variates with heat) is R21, D1, D2, D3, D4 and R20.... this result into bias reduction when you have high volume.... the bias measurable from Q14 base to Q15 base, that is normal beeing sligtly bigger than 3.65 volts, is reduced by the previous dinamic bias circuit.
The unit is biased to 500 miliamps each rail.. and this is reduced to 80 miliamps when volume is higher than some threshold stablished my the designer.
The amplifier "character" is different when you play into low volumes and when you play it loud....but both of them mufled in my point of view... very good quality... but mufled.... i have to increase treble to 3 O'clock to find it fine... seems not flat to me.
regards,
Carlos
R21, Q11 and R24 are in series, but Q11 is a variable resistance triggered by the voltage that is also going to the power output driver transistors (Q14 and Q15)
Also R19, Q7 and R23 works the same way triggered by the audio signal that goes to the output drivers.
Q5 and Q8 will have their colector to emitter resistances changed by Q11 and Q7.
All that stuff means three resistances in parallel.... Q11 circuit is one of them, and variable resistance triggered by the signal, also Q7 is a variable resistance triggered by the audio signal.... and the one has almost fixed resistance (variates with heat) is R21, D1, D2, D3, D4 and R20.... this result into bias reduction when you have high volume.... the bias measurable from Q14 base to Q15 base, that is normal beeing sligtly bigger than 3.65 volts, is reduced by the previous dinamic bias circuit.
The unit is biased to 500 miliamps each rail.. and this is reduced to 80 miliamps when volume is higher than some threshold stablished my the designer.
The amplifier "character" is different when you play into low volumes and when you play it loud....but both of them mufled in my point of view... very good quality... but mufled.... i have to increase treble to 3 O'clock to find it fine... seems not flat to me.
regards,
Carlos
Here is a link for the last image with better resolution
http://picasaweb.google.com/Destroy...photo?authkey=I90AQ1Ax7V0#5195710728944448274
Carlos
http://picasaweb.google.com/Destroy...photo?authkey=I90AQ1Ax7V0#5195710728944448274
Carlos
Hi Carlos, The image is good 🙂 very clear. I sometimes wonder with all these schemes, whether they behave as well under transient conditions, driving a real reactive load as opposed to a test resistance. As your bias decreases as the drive increases the gain or transconductance of the output stage must change fairly abruptly I guess. I have long come to the conclusion that some amplifiers are much more than the sum of the parts. Was it Nelson who said that sound quality seems to be in inverse proportion to complexity. In many ways that is true, a simple design may be technically imperfect, yet can really sing, some designs just seem to "come together" musically.
Even in the U.K. we can get high temperatures, always with high humidity to. There is no denying Class A is technically superior to any other for linear "audio use" but I believe there are other, just as important issues that have as much if not more to do with subjective quality. I am constantly suprised at the differrence in "sound" by going away from the long tailed pair input stage and back to a single ended input stage, with a correctly designed servo. It is just so much more forgiving and stable, and to me, more musical.
Regards Karl
Even in the U.K. we can get high temperatures, always with high humidity to. There is no denying Class A is technically superior to any other for linear "audio use" but I believe there are other, just as important issues that have as much if not more to do with subjective quality. I am constantly suprised at the differrence in "sound" by going away from the long tailed pair input stage and back to a single ended input stage, with a correctly designed servo. It is just so much more forgiving and stable, and to me, more musical.
Regards Karl
Nelson is rigth about complexity... a single transistor sounds better
No doubts about that..... the heat continues to be my personall problem.... i think to North Europe this is not a problem...so.... living there i would use only Class A and Nelson Pass designs without any doubt... and will trow everything different than that to the garbage can having absolutelly sure i made the correct thing.
Even auxiliary sub circuits, CCS, mirrors, sinks, ..all that stuff sounds.... the stabilized supply sounds too...also VBE multiplier...the best possible is the unefficiency of single ended, BGT in my point of view (very subjective, of course, and subjected to changes in the future)... with enormous current crossing... very hot.... linear operation... and the most perfect sound we can have... each part contributes to the sound reproduction destruction, even the input potentiometer.
One day i will pick a double coil woofer and will use it in series with a Class A amplifier colector or emitter.... and will let the current cross.... the other coil i will put DC voltage and current to counter ballance the speaker diafragm movement...and them will listen something i believe may be very good... despite the waste of energy.
I have not made that.. i have not that double coil woofer able to accept a couple of amperes without burn.... the DC voltage in one of the coils will be modulated by the audio... a superb direct coupling...the other coil will only force the coil not to go forward because of the audio amplifier DC colector current.
regards,
Carlos
No doubts about that..... the heat continues to be my personall problem.... i think to North Europe this is not a problem...so.... living there i would use only Class A and Nelson Pass designs without any doubt... and will trow everything different than that to the garbage can having absolutelly sure i made the correct thing.
Even auxiliary sub circuits, CCS, mirrors, sinks, ..all that stuff sounds.... the stabilized supply sounds too...also VBE multiplier...the best possible is the unefficiency of single ended, BGT in my point of view (very subjective, of course, and subjected to changes in the future)... with enormous current crossing... very hot.... linear operation... and the most perfect sound we can have... each part contributes to the sound reproduction destruction, even the input potentiometer.
One day i will pick a double coil woofer and will use it in series with a Class A amplifier colector or emitter.... and will let the current cross.... the other coil i will put DC voltage and current to counter ballance the speaker diafragm movement...and them will listen something i believe may be very good... despite the waste of energy.
I have not made that.. i have not that double coil woofer able to accept a couple of amperes without burn.... the DC voltage in one of the coils will be modulated by the audio... a superb direct coupling...the other coil will only force the coil not to go forward because of the audio amplifier DC colector current.
regards,
Carlos
forr said:
Hi,
Well I think Selfs definition of Class B
(optimally biased) is self (sic) indulgent.
What he calls classes:
C, B, AB, A
I prefer
B, aB, AB, A
Why ? well because my understanding of typical class C from radio
circuitry is class C only has an active device on one half of the cycle.
Selfs definition of C, < 100% active, whilst true for radio, hardly
represents the ~ 50% truth, which is what I call class C.
Typical class C IMO has nothing to do with audio power amplifiers.
FWIW AB and A both have area of conduction overlap where there
is a flat area that gm is doubled and this what I call class A. A class
A amplifier on extremes can clip or enter AB operation. AB means
a central area where gm doubles and then halves at the extremes.
Class aB or Selfs B does not contain an area of gm doubling so
though it has a bias current it does not operate in class A at any
point. This may be contentious but IMO its true. Class A operation
IMO required a linear area, for aB in the central area the devices
are handing over from one to the other, not linearly in tandem.
Its class a in that conduction overlaps, but not linear class A.
Class B or Selfs C has a deep gm chasm at the crossover point.
🙂/sreten.
Hi Carlos
(sorry to hijack this a thread a bit)
I remember in another thread somewhere you mentioned using the electrocompaniet output stage, what do you think of its sound compaired EF. (it doesnt measure as well as EF).
Im waiting for some sankens from profusion to try it out.
I m gonna try it with current mirrors to bias the the drivers symmetrical.
Those cambridge amps sound very good, i have modded a couple too. With some tweaks to compensation they sound awesome.
Alex
(sorry to hijack this a thread a bit)
I remember in another thread somewhere you mentioned using the electrocompaniet output stage, what do you think of its sound compaired EF. (it doesnt measure as well as EF).
Im waiting for some sankens from profusion to try it out.
I m gonna try it with current mirrors to bias the the drivers symmetrical.
Those cambridge amps sound very good, i have modded a couple too. With some tweaks to compensation they sound awesome.
Alex
I have tried Electrocompaniet CCS to drivers, adjusted from 15 to 40 miliamperes
And i had less distortion and very huge bass... the speaker started to move alone into the room... almost self destroying.
The Precision I is already huge in bass....it is already a monster in bass... so i left this to be used into another amplifier that needs more bass.
I use to listen...i do not use to make measurements..only simulator, and both simulates great....ahahahah..everything is good into simulator...at least the ones i select to simulate and to build....maybe lucky i am.
regards,
Carlos
And i had less distortion and very huge bass... the speaker started to move alone into the room... almost self destroying.
The Precision I is already huge in bass....it is already a monster in bass... so i left this to be used into another amplifier that needs more bass.
I use to listen...i do not use to make measurements..only simulator, and both simulates great....ahahahah..everything is good into simulator...at least the ones i select to simulate and to build....maybe lucky i am.
regards,
Carlos
Carlos,
I must be doing something wrong then, i have only simulated this output stage with ltspice and gotten worse figures. Then i tried the current mirrors which helped.
Ive purchased an Elec aw250 a couple of months back second hand for 250 Euros and love the sound, yes it has big bass.
. My brother in law (sound engineer) recommended
it as well as a doxa. Ive listened to doxa too😉, hopefully i can find a good second hand unit to purchase.
Going back to drawing board
Thanks Alex
I must be doing something wrong then, i have only simulated this output stage with ltspice and gotten worse figures. Then i tried the current mirrors which helped.
Ive purchased an Elec aw250 a couple of months back second hand for 250 Euros and love the sound, yes it has big bass.
. My brother in law (sound engineer) recommendedit as well as a doxa. Ive listened to doxa too😉, hopefully i can find a good second hand unit to purchase.
Going back to drawing board
Thanks Alex
I hope you're having fun going back to the drawing board
gimme the schematics you have please.... all you have related Cambridge.
panzertoo@yahoo.com
I am extremelly curious about this brand... i have been using Doctor Self ideas and they sound great!.... after some tweak of course...well... we are audiomaniac!
regards,
Carlos
gimme the schematics you have please.... all you have related Cambridge.
panzertoo@yahoo.com
I am extremelly curious about this brand... i have been using Doctor Self ideas and they sound great!.... after some tweak of course...well... we are audiomaniac!
regards,
Carlos
Hi Sreten
---Typical class C IMO has nothing to do with audio power amplifiers---.
Self again... For him, Quad current dumping belongs to a sub-class he calls :
A + C.
Some amplifiers had their power output devices with no bias at all for low level signals : Phase Linear, ESS, Graham Nalty and one of the first amplifier published by Elektor.
---Typical class C IMO has nothing to do with audio power amplifiers---.
Self again... For him, Quad current dumping belongs to a sub-class he calls :
A + C.
Some amplifiers had their power output devices with no bias at all for low level signals : Phase Linear, ESS, Graham Nalty and one of the first amplifier published by Elektor.
- Status
- Not open for further replies.