Hi pma...
Contrary to almost universal opinion, a bipolar junction transistor is in fact a voltage operated current source, even though its minority base charge draw is non-negligible...which is why BJT collector current is approximated by the following relationship:
Ic=Is e^(Vbe/Vt)
In other words, collector current is controlled by base-emitter voltage.
PMA said:
Contrary to almost universal opinion, a bipolar junction transistor is in fact a voltage operated current source, even though its minority base charge draw is non-negligible...which is why BJT collector current is approximated by the following relationship:
Ic=Is e^(Vbe/Vt)
In other words, collector current is controlled by base-emitter voltage.
Hi Andy...
I think i have been misunderstood.......good rules of thumb can only derive from a sound quantitative understanding of physical phenomena....
However, for a competent engineering mathematician, knowing when to apply mathematical formulae, algebraic derivations, etc... is just as important as recognising the many instances in practice, where it is futile to indulge in the self-flagellation
of deriving a mile long expression which yields near zero physical insight.
P.S: Michael Farady could hardly be described as a 'rule of thumb' practitioner methinks...
ALW said:
I think i have been misunderstood.......good rules of thumb can only derive from a sound quantitative understanding of physical phenomena....
However, for a competent engineering mathematician, knowing when to apply mathematical formulae, algebraic derivations, etc... is just as important as recognising the many instances in practice, where it is futile to indulge in the self-flagellation
of deriving a mile long expression which yields near zero physical insight.P.S: Michael Farady could hardly be described as a 'rule of thumb' practitioner methinks...
Hi pma...
AAAAHH i see your point.....you are assuming here, that the Vbe is established by Ib...which is incorrect. Ic is established by Vbe independent of Ib, but Ib is required to remove or replenish, (depending on BJT polarity), minority carriers resulting from said collector current flow....otherwise BJT transistor action would cease as soon as it is initiated....
PMA said:Mikek
that's right and Vbe=(kT/q)ln(Ic/Is)
AAAAHH i see your point.....you are assuming here, that the Vbe is established by Ib...which is incorrect. Ic is established by Vbe independent of Ib, but Ib is required to remove or replenish, (depending on BJT polarity), minority carriers resulting from said collector current flow....otherwise BJT transistor action would cease as soon as it is initiated....
Mikek
BJT is defined by the equations
Ic=Is.exp(qVbe/kT) or Vbe=(kT/q)ln(Ic/Is)
and Ib=Ic/beta.
Vbe voltage depends on Ic current. Ib is Ic divided by beta, that is why I speak about Vbe derived from Ib. In every real transistor circuit you control Ib, Ib = (Vin - Vbe)/(R + h11), Vbe = 0.6V approx. You never control Vbe directly, such a circuit would not work.
BJT is defined by the equations
Ic=Is.exp(qVbe/kT) or Vbe=(kT/q)ln(Ic/Is)
and Ib=Ic/beta.
Vbe voltage depends on Ic current. Ib is Ic divided by beta, that is why I speak about Vbe derived from Ib. In every real transistor circuit you control Ib, Ib = (Vin - Vbe)/(R + h11), Vbe = 0.6V approx. You never control Vbe directly, such a circuit would not work.
Err,
Long-tail pair?
You're both right. But, allow me to quote this from Doug Self's site :
By "transistor law", he means the relationship between Ic and VbeTHE TRANSISTOR LAW. (in preparation)
Every transistor obeys it with startling accuracy over nine or ten decades of Ic, which is a pretty broad hint that we are looking at the fundamental mechanism. In contrast, beta varies with Ic, temperature, and just about everything else. The collector current is to a first approximation independent of collector voltage- in other words it is a current-source output. The qualifications to this are:
1: This only holds for Vce above, say, 2 Volts.
2: It is not a perfect current-source; even with a high Vce, Ic remains a weak function of Vce. This is called Early Effect, after Dr Early, and has nothing to do with timing or punctuality. The same effect when the transistor is operated in reverse mode- a perversion that will not concern us here- has sometimes been called Late Effect. Ho ho
I think it is important to always remember the point the mikek made - Ib is just a consequence of Vbe. But, the exponential relationship between Vbe and Ic is a pain when you are trying to make a linear circuit, which is why LTP's are such a good circuit.
An interesting experiment - make a simple CE amplifier - bias it correctly, and feed the base from a low-impedance signal generator (thereby modifying Vbe directly). Your output won't look much like a sine wave due to the exponential transfer function. But, add a 47K resistor between the signal generator and transistor base, which will approximate current drive... Suddenly, the output waveform looks like a sine wave! Look at the base waveform and you'll see a distorted sine wave again, the exponential Vbe-Ic relationship still holds true. The linear Ib-Ic relationship is working here... But, you wouldn't want to rely on it - it will be affected by temperature, transistor type, etc, etc.
Replace the transitor with a diode connected between the added resistor and ground, and you'll see the same distorted waveform - proving that the BE junction looks like a diode...
Good discussion.
I don't think it matters whether you consider Ic a direct consequence of Vbe or a direct consequence of Ib since Vbe and Ib are inextricably related by the highly non-linear input Z - a bit of a chicken and egg thing. However, I have found it more fruitful to understanding what is going on to think of transistors as current and charge devices: this applies to both BJTs and FETs. It is less familiar to think of them in this way since most people are brought up on voltages across resistances causing currents and then focus on what the voltages are doing. You have to be very careful because transistor transconductance is very non-linear. As mhennessy mentions the input interface to a standard power amp is voltage-defined, which is converted into (usually) a linearly related current by operating a LTP with as small a signal current variation as possible (and/or large emitter resistors) to maximize the linearity of the transconductance. Then we amplify this current and the integral of this current (charge) using several transistor stages and feed this into the speaker. Because (for some historical reason related to the EMF of voicecoils) the output is voltage-defined we then have a job to control this current in order to get the desired output voltage across the unpredictable speaker load. This is done by some combination of feedback and load impedances. Some prefer the impedance route (usually involving high currents and heat) and others the feedback route (requiring close attention to linearities and stability).
My advice to the budding designers here is to remember to think about the current and charge. Do the maths to compare the equation of a darlington and a pseudo-darlington or CFP in circuit and see what is different - if anything. If you just look at the diagram on paper they look different and one looks like it has a loop and one doesn't and soon you are making visual conclusions. This is fine for the Tate Modern but electrons don't see the big picture. Electronics is one field, along with mechanics and astrophysics and others I suppose, where the mathematics is highly correlated with real behaviour and an essential tool for understanding how systems behave. Don't be fooled by the schematic - it is just a diagram. Don't assume something that looks pretty and symmetrical performs better: the electrons don't give a wet slap. Conversely, just because the diagram isn't pretty doesn't mean the circuit behaves poorly. Remember N-channel and P-channel devices are NOT exact opposites (even if the mfr writes "complementary" on the datasheet) because electrons don't have opposites (except in Star Trek and briefly in some underground tunnel in Switzerland).
I don't think it matters whether you consider Ic a direct consequence of Vbe or a direct consequence of Ib since Vbe and Ib are inextricably related by the highly non-linear input Z - a bit of a chicken and egg thing. However, I have found it more fruitful to understanding what is going on to think of transistors as current and charge devices: this applies to both BJTs and FETs. It is less familiar to think of them in this way since most people are brought up on voltages across resistances causing currents and then focus on what the voltages are doing. You have to be very careful because transistor transconductance is very non-linear. As mhennessy mentions the input interface to a standard power amp is voltage-defined, which is converted into (usually) a linearly related current by operating a LTP with as small a signal current variation as possible (and/or large emitter resistors) to maximize the linearity of the transconductance. Then we amplify this current and the integral of this current (charge) using several transistor stages and feed this into the speaker. Because (for some historical reason related to the EMF of voicecoils) the output is voltage-defined we then have a job to control this current in order to get the desired output voltage across the unpredictable speaker load. This is done by some combination of feedback and load impedances. Some prefer the impedance route (usually involving high currents and heat) and others the feedback route (requiring close attention to linearities and stability).
My advice to the budding designers here is to remember to think about the current and charge. Do the maths to compare the equation of a darlington and a pseudo-darlington or CFP in circuit and see what is different - if anything. If you just look at the diagram on paper they look different and one looks like it has a loop and one doesn't and soon you are making visual conclusions. This is fine for the Tate Modern but electrons don't see the big picture. Electronics is one field, along with mechanics and astrophysics and others I suppose, where the mathematics is highly correlated with real behaviour and an essential tool for understanding how systems behave. Don't be fooled by the schematic - it is just a diagram. Don't assume something that looks pretty and symmetrical performs better: the electrons don't give a wet slap. Conversely, just because the diagram isn't pretty doesn't mean the circuit behaves poorly. Remember N-channel and P-channel devices are NOT exact opposites (even if the mfr writes "complementary" on the datasheet) because electrons don't have opposites (except in Star Trek and briefly in some underground tunnel in Switzerland).
Misunderstood - not at all
Couldn't agree more, but I'm not certain that many 'competent' engineers can do that, simply by looking at a circuit, physically, ro via a schematic.
It's far too easy sometimes, to make assumptions, simply because one circuit looks like another.
Andy (who likes rules of thumb and hates math)
Couldn't agree more, but I'm not certain that many 'competent' engineers can do that, simply by looking at a circuit, physically, ro via a schematic.
It's far too easy sometimes, to make assumptions, simply because one circuit looks like another.
Andy (who likes rules of thumb and hates math)
Bam made this comment:
"It is naive to blame the CFP topology without understanding the cause. Judging by some of the nonesense I'm reading here more homework is required."
By his categoric statement he damns all practitioners, who work hard by building and listening, within the exacting limits imposed by the engineering, to produce a better amplifier. An analogous argument would be to dismiss Nigella Lawson's cooking skills because she understands none of the organic chemistry of taste (or the maths of chemical equilibrium), and all like her, who work in an area properly rergarded as more as an art than a science.
I work with a detailed understanding of Ohms and Kirchoff's laws, and the simpler equations of semiconductor operation, including beta, output impedance, and a string of others I refer to in Boylestad and Nashevsky's excellent Prentice Hall text on the matter.
Yet in this text, replete with equations deriving gain, Zin, Zout, frequency response and others, nowhere does it mention the intrinsic instability of the unity gain CFP/Sziklai, which I investigated by building and sonically comparing in identical circumstances alongside a stock Type II Self emitter follower.
To be honest, I don't give a hoot why the CFP cannot deliver the goods. I made the observation empirically, looked at the mechanisms, briefly considered the complex math, and moved on. Like all practititioners, I'm more interested in what does work, so I can make and sell them. My customers don't care either; they spend a lot of time and energy listening to music, and they want the best possible for the least dollars. All this iconoclastic ******** about the math whys and wherefores is inconsequential; highly academic from where I stand, and only useful if there is no solution. In fact there is. If we are considering the joys of cold fusion, or the technology of energy storage, and seek a solution, then yes, it's undoubtedly useful to pore over the math, but when the double emitter follower is a tried and tested configuration, is well understood and easy to implement and works unbelievably well, then only the committed academic and a few engineers would be prepared to dissect the maths models to find why the CFP falls short of the mark.
Like Faraday, and possibly a couple of others here, I refute utterly the intellectually barren argument that failure to fully understand the mathematics of semiconductor operation precludes any worthwhile contribution to the art because it '...leads to misunderstanding'. This is arrant snobbery, and serves only to highlight that Bam understands this stuff and others presumably don't. Do we forbid the electrician to enter our home because he does not understand the hyperbolic cosh functions which describe the dissipation of power in an AC transmission line? Do we roundly condemn his work because he does not understand the math which describes the inductive behaviour of a fluorescent lamp ballast? An absurd notion...... he can still be productive even though he may understand little of the math. Fer Crissakes, mon, this is the reason we have researchers and academics - to examine these phenomenon with the tools of maths - and empiricism I might add - so that eventually it CAN be explained, and subsequently exploited commercially by people to whom the math is as meaningful as colour to a blind man.......
One final point, before I give up on this crazy thread. Mathematics is a model, an attempt to explain quantitatively a physical phenomenon. It is a tool. It is certainly not the actual phenomenon. An example is Ohms law; it works very well in 99% of cases. However, the property of negative resistance is an important exception, and arguably, superconductivity is another. It is unreasonable to postulate that full understanding of a phenomenon is based entirely on maths; this suggests that people who have a 'feel' for circuits are just quacks, and yet this is not true. There are people who can look at a circuit of an amplifier, and without knowing precisely why, suggest ways of making it sound better. I have met plenty of such people. You ignore them at your peril if you are attempting to make a better mousetrap.
Bam's propositions reflect that he is good at maths, intellectually capable, and very proud of it. This thinking is classic academic ego, and he should realize that many important developments in the history of technology - nitroglycerine to name a comical example - are pure serendipity.
Cheers,
Hugh
www.aksaonline.com
"It is naive to blame the CFP topology without understanding the cause. Judging by some of the nonesense I'm reading here more homework is required."
By his categoric statement he damns all practitioners, who work hard by building and listening, within the exacting limits imposed by the engineering, to produce a better amplifier. An analogous argument would be to dismiss Nigella Lawson's cooking skills because she understands none of the organic chemistry of taste (or the maths of chemical equilibrium), and all like her, who work in an area properly rergarded as more as an art than a science.
I work with a detailed understanding of Ohms and Kirchoff's laws, and the simpler equations of semiconductor operation, including beta, output impedance, and a string of others I refer to in Boylestad and Nashevsky's excellent Prentice Hall text on the matter.
Yet in this text, replete with equations deriving gain, Zin, Zout, frequency response and others, nowhere does it mention the intrinsic instability of the unity gain CFP/Sziklai, which I investigated by building and sonically comparing in identical circumstances alongside a stock Type II Self emitter follower.
To be honest, I don't give a hoot why the CFP cannot deliver the goods. I made the observation empirically, looked at the mechanisms, briefly considered the complex math, and moved on. Like all practititioners, I'm more interested in what does work, so I can make and sell them. My customers don't care either; they spend a lot of time and energy listening to music, and they want the best possible for the least dollars. All this iconoclastic ******** about the math whys and wherefores is inconsequential; highly academic from where I stand, and only useful if there is no solution. In fact there is. If we are considering the joys of cold fusion, or the technology of energy storage, and seek a solution, then yes, it's undoubtedly useful to pore over the math, but when the double emitter follower is a tried and tested configuration, is well understood and easy to implement and works unbelievably well, then only the committed academic and a few engineers would be prepared to dissect the maths models to find why the CFP falls short of the mark.
Like Faraday, and possibly a couple of others here, I refute utterly the intellectually barren argument that failure to fully understand the mathematics of semiconductor operation precludes any worthwhile contribution to the art because it '...leads to misunderstanding'. This is arrant snobbery, and serves only to highlight that Bam understands this stuff and others presumably don't. Do we forbid the electrician to enter our home because he does not understand the hyperbolic cosh functions which describe the dissipation of power in an AC transmission line? Do we roundly condemn his work because he does not understand the math which describes the inductive behaviour of a fluorescent lamp ballast? An absurd notion...... he can still be productive even though he may understand little of the math. Fer Crissakes, mon, this is the reason we have researchers and academics - to examine these phenomenon with the tools of maths - and empiricism I might add - so that eventually it CAN be explained, and subsequently exploited commercially by people to whom the math is as meaningful as colour to a blind man.......
One final point, before I give up on this crazy thread. Mathematics is a model, an attempt to explain quantitatively a physical phenomenon. It is a tool. It is certainly not the actual phenomenon. An example is Ohms law; it works very well in 99% of cases. However, the property of negative resistance is an important exception, and arguably, superconductivity is another. It is unreasonable to postulate that full understanding of a phenomenon is based entirely on maths; this suggests that people who have a 'feel' for circuits are just quacks, and yet this is not true. There are people who can look at a circuit of an amplifier, and without knowing precisely why, suggest ways of making it sound better. I have met plenty of such people. You ignore them at your peril if you are attempting to make a better mousetrap.
Bam's propositions reflect that he is good at maths, intellectually capable, and very proud of it. This thinking is classic academic ego, and he should realize that many important developments in the history of technology - nitroglycerine to name a comical example - are pure serendipity.
Cheers,
Hugh
www.aksaonline.com
Bravo Hugh Dean....
....could'nt agree more....and yes, i consider myself an applied mathematician.....have derived CFP transfer function, (to second order approx), to determine location and relevance of singularities, complex or otherwise....etc....
Alas, managed to prove nothing of value in practice that i had'nt already established on stripboard, and pspice....i.e: requires lag compensation to gaurantee unconditional stability,....is prone to cross-conduction at HF, particulaly when driving low impedance loads with paralleled devices.....is infact not really that linear in closed-loop class-B/AB...etc.......In general, looks elegant but notworth the hassle in practice.....
....could'nt agree more....and yes, i consider myself an applied mathematician..
...have derived CFP transfer function, (to second order approx), to determine location and relevance of singularities, complex or otherwise....etc....Alas, managed to prove nothing of value in practice that i had'nt already established on stripboard, and pspice....i.e: requires lag compensation to gaurantee unconditional stability,....is prone to cross-conduction at HF, particulaly when driving low impedance loads with paralleled devices.....is infact not really that linear in closed-loop class-B/AB...etc.......In general, looks elegant but notworth the hassle in practice.....
What about just running some gain?
http://www.diyaudio.com/forums/attachment.php?postid=37626
I've run this with 3X gain and not seen a problem.
And what about this:
http://home.kimo.com.tw/skychutw/ampzilla/schematics/ampzillaIII_sch.gif
I have a well respected amplifier with this output stage except instead of 100% feedback in the driver to pre-driver it has a voltage gain of 5X (and the same resistor and cap scheme as the first).
http://www.diyaudio.com/forums/attachment.php?postid=37626
I've run this with 3X gain and not seen a problem.
And what about this:
http://home.kimo.com.tw/skychutw/ampzilla/schematics/ampzillaIII_sch.gif
I have a well respected amplifier with this output stage except instead of 100% feedback in the driver to pre-driver it has a voltage gain of 5X (and the same resistor and cap scheme as the first).
stereotypes R us
"By his categoric statement he damns all practitioners...
"This thinking is classic academic ego..."
Hugh I really don't understand why you make stuff up like this and go in for personal attacks and why you have some chip on your shoulder about engineers/academics. It doesn't strengthen your arguments and I don't think this is relevant to this thread nor in the spirit of the diyaudio. So why not concentrate on the technical discussion; you might even learn something.
"By his categoric statement he damns all practitioners...
"This thinking is classic academic ego..."
Hugh I really don't understand why you make stuff up like this and go in for personal attacks and why you have some chip on your shoulder about engineers/academics. It doesn't strengthen your arguments and I don't think this is relevant to this thread nor in the spirit of the diyaudio. So why not concentrate on the technical discussion; you might even learn something.
Bam,
It's not strictly personal. That is in error. I'm perfectly willing to accept you as a decent, considerate guy. By your personal email to me you have demonstrated that you are sensitive, and thoughtful. I merely rail against the purely mathematical approach.
It is the philosophy, which contaminates the entire industry, and has delivered a horrific red herring. I have a beef about how math and measurement has delivered amplifiers which measure well, and sound very ordinary. I feel passionate about this, and would stress that nothing in technology, just like art, is achieved without passion.
I concede that if I listen carefully, I might learn something, and I thank you for pointing this out. However, there is not enough time. I fight daily for survival as this is my living. So, knowing what does work, I have little choice.
I've just had word from an AKSA constructor that my 100W amp, which I had believed sounded marvellous and measured badly, in fact measures very well indeed! I had never fully measured it because of my obsession with sonics and my lack of expensive measurement equipment! It turns out that at 1W into 8R at 20KHz the distortion measurement is just 0.045%!
I wish you, your family, and those of people reading this, a Happy New Year. Not long now!
Let's hope that 2003 brings new revelations, and successes!
Cheers,
Hugh
It's not strictly personal. That is in error. I'm perfectly willing to accept you as a decent, considerate guy. By your personal email to me you have demonstrated that you are sensitive, and thoughtful. I merely rail against the purely mathematical approach.
It is the philosophy, which contaminates the entire industry, and has delivered a horrific red herring. I have a beef about how math and measurement has delivered amplifiers which measure well, and sound very ordinary. I feel passionate about this, and would stress that nothing in technology, just like art, is achieved without passion.
I concede that if I listen carefully, I might learn something, and I thank you for pointing this out. However, there is not enough time. I fight daily for survival as this is my living. So, knowing what does work, I have little choice.
I've just had word from an AKSA constructor that my 100W amp, which I had believed sounded marvellous and measured badly, in fact measures very well indeed! I had never fully measured it because of my obsession with sonics and my lack of expensive measurement equipment! It turns out that at 1W into 8R at 20KHz the distortion measurement is just 0.045%!
I wish you, your family, and those of people reading this, a Happy New Year. Not long now!
Let's hope that 2003 brings new revelations, and successes!
Cheers,
Hugh
Re: stereotypes R us
Bam does have a liking it would appear, for unsavoury phrases and words like 'nonsense'...'you might learn something', etc...just look at the Naim thread for instance.......It would appear it has'nt occured to the worthy that these may apply particularly to his good self......
traderbam said:"By his categoric statement he damns all practitioners...
"This thinking is classic academic ego..."
Hugh I really don't understand why you make stuff up like this and go in for personal attacks and why you have some chip on your shoulder about engineers/academics. It doesn't strengthen your arguments and I don't think this is relevant to this thread nor in the spirit of the diyaudio. So why not concentrate on the technical discussion; you might even learn something.
Bam does have a liking it would appear, for unsavoury phrases and words like 'nonsense'...'you might learn something', etc...just look at the Naim thread for instance.......It would appear it has'nt occured to the worthy that these may apply particularly to his good self......
Mikek: I'm glad you mentioned the Naim site. If you speak with some of the Naim experts there you will discover that the Naim amps use both a darlington and CFP in their outputs. You may gan some insights into why they chose to do this and how they dealt with the "problems" of CFPs that you have claimed; perhaps you'll share with this forum what you learn.
Hi traderbam
In other words Naim uses Baxandals' modification of Lins' quasi-complementary SEPP output stage.
The Baxandal mod. in the low-side CFP portion of the circuit, consists of a diode junction in parallel with a resistor. Notice that this resistor is the same size as the collector resistor in the drivers collector....This drastically reduces the local feedback factor of the CFP, which strictly speaking, cannot be accurately called a 'sziklai' compund pair, as the later is a 100% local voltage feedback concoction.
The reduction in local feedback for the low-side pair, makes it markedly less linear than the sziklai, while ameliorating the laters propensity for parasitic oscillation.
The high-side compund pair on the other hand, is a simple darlington, without the sziklais comparatively better linearity,and saddled with the later's handicap with respect to charge extraction from the output device, particularly at high freqs.
Such an output stage constitutes a symphony of trade-offs, which are only justifiable and consistent with a tight budget. I.e.:low-fi to mid-fi.
For those who feel moved to drive low impedance loads with multiple paralled BJT's, this compromise is sub-optimal....and the whole is likely to be extremely difficult to rid of intermittent low-level oscillation.
It is nigh on imposible to learn audio electronic design from subjectivists, some of whom for instance, swear by the cascode, while others deride it for 'veiling' the 'tonality of the sound', etc..etc.....(refer to cascode thread on this forum). Consistency, and repeatability are singularly lacking in the theology of the subjectivist camp.......This is hardly conducive for serious research, or indeed, the Traderbam-defined 'learning' process.
traderbam said:
In other words Naim uses Baxandals' modification of Lins' quasi-complementary SEPP output stage.
The Baxandal mod. in the low-side CFP portion of the circuit, consists of a diode junction in parallel with a resistor. Notice that this resistor is the same size as the collector resistor in the drivers collector....This drastically reduces the local feedback factor of the CFP, which strictly speaking, cannot be accurately called a 'sziklai' compund pair, as the later is a 100% local voltage feedback concoction.
The reduction in local feedback for the low-side pair, makes it markedly less linear than the sziklai, while ameliorating the laters propensity for parasitic oscillation.
The high-side compund pair on the other hand, is a simple darlington, without the sziklais comparatively better linearity,and saddled with the later's handicap with respect to charge extraction from the output device, particularly at high freqs.
Such an output stage constitutes a symphony of trade-offs, which are only justifiable and consistent with a tight budget. I.e.:low-fi to mid-fi.
For those who feel moved to drive low impedance loads with multiple paralled BJT's, this compromise is sub-optimal....and the whole is likely to be extremely difficult to rid of intermittent low-level oscillation.
It is nigh on imposible to learn audio electronic design from subjectivists, some of whom for instance, swear by the cascode, while others deride it for 'veiling' the 'tonality of the sound', etc..etc.....(refer to cascode thread on this forum). Consistency, and repeatability are singularly lacking in the theology of the subjectivist camp.......This is hardly conducive for serious research, or indeed, the Traderbam-defined 'learning' process.
Hi PMA
I have no personal experiance of it myself, but i am reliably informed that this too is prone to cross-conduction-induced overheating if tested at even modest audio freqs., eg 10Khz.
The triple cross-coupled follower as used for instance in Leachs' so-called 'Low TIM' design, and Harman Kardons late lamented HK610/620 series, really is near ideal in this respect.
I have no personal experiance of it myself, but i am reliably informed that this too is prone to cross-conduction-induced overheating if tested at even modest audio freqs., eg 10Khz.
The triple cross-coupled follower as used for instance in Leachs' so-called 'Low TIM' design, and Harman Kardons late lamented HK610/620 series, really is near ideal in this respect.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.