CEF = complementary emitter follower?
Mitek,
I always thought that hf-crossconduction was basically caused by a combination of relatively slow output transistors that cannot turnoff quick enough, and a driver circuit that has not enough capability to pull the charge out of the base of the turrning-off output transistor.
Surely the same fast output transistors can be used in any config, so this would not make any difference. So, when you say that the CEF stage inherently is worse at xconduction, what would be the reason?
Jan Didden
Mitek,
I always thought that hf-crossconduction was basically caused by a combination of relatively slow output transistors that cannot turnoff quick enough, and a driver circuit that has not enough capability to pull the charge out of the base of the turrning-off output transistor.
Surely the same fast output transistors can be used in any config, so this would not make any difference. So, when you say that the CEF stage inherently is worse at xconduction, what would be the reason?
Jan Didden
Mikek wrote:
"In other words Naim uses Baxandals' modification of Lins' quasi-complementary SEPP output stage."
Naim might take issue with which came first. I believe the configuration was published in the RCA Power Transistor book in the early 1970s, well before Linn were making amps. It is a pretty standard configuration for using two npn output transistors in push-pull.
"...which strictly speaking, cannot be accurately called a 'sziklai' compund pair, as the later is a 100% local voltage feedback concoction."
The way I look at it the darlington, Sziklai, CFP, etc, etc are not fundamentally different wrt their feedback. I don't really understand what you mean by "100% local voltage feedback".
"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."
You are missing the point. I infer you think Naim chose this to overcome 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. "
Charge extraction is not a benefit of the compound darlington over the ordinary darlington.
"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. "
Oh dear.
I think you had better placate the Naim owners out there who don't consider their equipment to be low-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."
Do you believe Krell's output stages are sub-optimal?
"It is nigh on imposible to learn audio electronic design from subjectivists..."
I agree in the sense that those who can hear differences are not always the best people to deduce the root cause or make design choices. Eg: 'I added a cascode and it sounded worse therefore cascodes are bad' would be naive reasoning. Is this what you mean?
"In other words Naim uses Baxandals' modification of Lins' quasi-complementary SEPP output stage."
Naim might take issue with which came first. I believe the configuration was published in the RCA Power Transistor book in the early 1970s, well before Linn were making amps. It is a pretty standard configuration for using two npn output transistors in push-pull.
"...which strictly speaking, cannot be accurately called a 'sziklai' compund pair, as the later is a 100% local voltage feedback concoction."
The way I look at it the darlington, Sziklai, CFP, etc, etc are not fundamentally different wrt their feedback. I don't really understand what you mean by "100% local voltage feedback".
"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."
You are missing the point. I infer you think Naim chose this to overcome 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. "
Charge extraction is not a benefit of the compound darlington over the ordinary darlington.
"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. "
Oh dear.
I think you had better placate the Naim owners out there who don't consider their equipment to be low-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."
Do you believe Krell's output stages are sub-optimal?
"It is nigh on imposible to learn audio electronic design from subjectivists..."
I agree in the sense that those who can hear differences are not always the best people to deduce the root cause or make design choices. Eg: 'I added a cascode and it sounded worse therefore cascodes are bad' would be naive reasoning. Is this what you mean?
I haven't tried this topology out yet but will definitely do so with a BJT driver and a MOSFET output device as soon as I have the time.
There are indeed two reasons I could imagine, that make this circuit more prone to cross-conduction:
1.) Common emitter circuits are much more influenced by the Miller effect than emitter-followers (the same applies to the corresponding MOSFET circuits).
2.) The output devices are not turned-off by an active device but by a resistor only.
The only remedy for 1.) is the choice of apropriately fast devices IMO. While 2.) might be overcome by a more refined circuit topology (I still have to think about this one).
Regards
Charles
There are indeed two reasons I could imagine, that make this circuit more prone to cross-conduction:
1.) Common emitter circuits are much more influenced by the Miller effect than emitter-followers (the same applies to the corresponding MOSFET circuits).
2.) The output devices are not turned-off by an active device but by a resistor only.
The only remedy for 1.) is the choice of apropriately fast devices IMO. While 2.) might be overcome by a more refined circuit topology (I still have to think about this one).
Regards
Charles
Phase accurate wrote: "1.) Common emitter circuits are much more influenced by the Miller effect than emitter-followers (the same applies to the corresponding MOSFET circuits).
"
Are you sure? The Miller effect is CB (collector-base) current due to CB voltage change. Consider the relative voltage changes across the CB in common-emitter output vs common collector output. How different is it? Where does the CB Miller current flow to in either case?
"2.) The output devices are not turned-off by an active device but by a resistor only."
The BE of the output transistor always 'sees' the resistor in parallel with the drivers output Z. When the driver is no longer amplifying its output Z rises and becomes dominated by its parasitic BE Z in series with the Z of the preceding stage.
"
Are you sure? The Miller effect is CB (collector-base) current due to CB voltage change. Consider the relative voltage changes across the CB in common-emitter output vs common collector output. How different is it? Where does the CB Miller current flow to in either case?
"2.) The output devices are not turned-off by an active device but by a resistor only."
The BE of the output transistor always 'sees' the resistor in parallel with the drivers output Z. When the driver is no longer amplifying its output Z rises and becomes dominated by its parasitic BE Z in series with the Z of the preceding stage.
Hi traderbam
I definitely agree with you that everything looks the same if you only look at the Miller capacity. But the Miller effect is also influenced by the Base-Emitter capacity (or Gate to Source capacity) and therfore things may look different for both circuits.
I do also definitely agree with your model of the base-drive circuit, i.e. the driver as a current source parallelled with the resistor. But that doesn't avoid the fact that the circuit is able to turn on faster than turn off.
BTW: I like the CFP as well, but sometimes it is wise to check all advantages and disadvantages before implementing a circuit. And suddenly you come to an elegant solution .......... !
Regards
Charles
I definitely agree with you that everything looks the same if you only look at the Miller capacity. But the Miller effect is also influenced by the Base-Emitter capacity (or Gate to Source capacity) and therfore things may look different for both circuits.
I do also definitely agree with your model of the base-drive circuit, i.e. the driver as a current source parallelled with the resistor. But that doesn't avoid the fact that the circuit is able to turn on faster than turn off.
BTW: I like the CFP as well, but sometimes it is wise to check all advantages and disadvantages before implementing a circuit. And suddenly you come to an elegant solution .......... !
Regards
Charles
Naim dropping
Thanks ALW. I just returned from a visit to Specsavers having destroyed my glasses last week. It is true! Can't say I've ever heard of Lin...have I missed out on anything profound?
Phase:
I'm not sure what you mean by the miller effect being influenced by the BE capacitance. The BE capacitance will vary (moreso for a FET) which will affect the proportion of miller current that ends up returning to the base for different Vbe/Ic. Is this what you mean? I believe Ccb is primarily an inverse function of Vbc due to the variation in width of the depletion layer of the reverse-biased bc junction.
I agree that push-pull drivers, for example, will speed up both injection and removal of charge in the output device. But is it not a matter of what maximum base current is needed for normal operation? So, you could replace the BE resistor with a CCS with a high enough value and get perfectly symmetrical charge and discharge rates. With a BJT output the BE voltage is pretty steady at 0.7V or so until things start turning off so a resistor is a practical alternative to a CCS. The same is not true, however, of FETs because their Vgs is much more variable and so a resistor CCS will introduce more non-linear current through the driver. Granted the FET has virtually zero dc current demand but it also has a big input capacitance that is very non-linear and needs a correspondingly complex driver to control it, at least in source-follower configuration.
Thanks ALW. I just returned from a visit to Specsavers having destroyed my glasses last week. It is true! Can't say I've ever heard of Lin...have I missed out on anything profound?
Phase:
I'm not sure what you mean by the miller effect being influenced by the BE capacitance. The BE capacitance will vary (moreso for a FET) which will affect the proportion of miller current that ends up returning to the base for different Vbe/Ic. Is this what you mean? I believe Ccb is primarily an inverse function of Vbc due to the variation in width of the depletion layer of the reverse-biased bc junction.
I agree that push-pull drivers, for example, will speed up both injection and removal of charge in the output device. But is it not a matter of what maximum base current is needed for normal operation? So, you could replace the BE resistor with a CCS with a high enough value and get perfectly symmetrical charge and discharge rates. With a BJT output the BE voltage is pretty steady at 0.7V or so until things start turning off so a resistor is a practical alternative to a CCS. The same is not true, however, of FETs because their Vgs is much more variable and so a resistor CCS will introduce more non-linear current through the driver. Granted the FET has virtually zero dc current demand but it also has a big input capacitance that is very non-linear and needs a correspondingly complex driver to control it, at least in source-follower configuration.
Hi traderbam
I fear you have misunderstood.....you are mistaking Linn the scottish audio manufacturing company for H.C Lin, the inventor of the quasi complementary SEPP output stage, (circa 1956),...see attached file..
....each sziklai pair, (low-side or high-side), in fact consists of two discrete and complementary common emitter stages in series, with the output from the second stage fed back to the emitter of the first in its entirety...in otherwords, the feedback signal is a voltage derived from the collector of the second member, (shunt-derived),and applied at the emitter of the first, (series-applied).
...what precisely is the point i appear to be missing?
.....correct, the darlington, compound or otherwise thereof, provides no singular improvement with respect to speed of charge extraction compared to the sziklai.......
However, i think you'll find that i specifically refered to the double complementary emitter follower, (modified Darlington if you like), with cross-coupled, (shared) emitter resistors for driver(s), as used, (for instance), in the Leach 'low-TIM'.
The shared emitter resistor in the complementary driver stage facilitates the reverse bias of the base-emitter junction of the output BJT being turned off..(see D. Selfs' book), which improves the rate of charge extraction significantly....
...nothing personal.....we are all entitled to our opinions...
...correct me if i am wrong, ....but i do'nt think Krell uses the quasi-complementary SEPP output stage, (Lin or variations thereof), in their designs.
...Yes..
traderbam said:
I fear you have misunderstood.....you are mistaking Linn the scottish audio manufacturing company for H.C Lin, the inventor of the quasi complementary SEPP output stage, (circa 1956),...see attached file..
traderbam said:
....each sziklai pair, (low-side or high-side), in fact consists of two discrete and complementary common emitter stages in series, with the output from the second stage fed back to the emitter of the first in its entirety...in otherwords, the feedback signal is a voltage derived from the collector of the second member, (shunt-derived),and applied at the emitter of the first, (series-applied).
traderbam said:
...what precisely is the point i appear to be missing?
traderbam said:
.....correct, the darlington, compound or otherwise thereof, provides no singular improvement with respect to speed of charge extraction compared to the sziklai.......
However, i think you'll find that i specifically refered to the double complementary emitter follower, (modified Darlington if you like), with cross-coupled, (shared) emitter resistors for driver(s), as used, (for instance), in the Leach 'low-TIM'.
The shared emitter resistor in the complementary driver stage facilitates the reverse bias of the base-emitter junction of the output BJT being turned off..(see D. Selfs' book), which improves the rate of charge extraction significantly....
traderbam said:Mikek wrote:
"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. "
Oh dear.
...nothing personal.....we are all entitled to our opinions...
traderbam said:
...correct me if i am wrong, ....but i do'nt think Krell uses the quasi-complementary SEPP output stage, (Lin or variations thereof), in their designs.
traderbam said:
...Yes..
Re: CEF = complementary emitter follower?
...correct...
.....Not quite, the darlington, compound or otherwise thereof, provides no singular improvement with respect to speed of charge extraction compared to the sziklai.......
However, the double complementary emitter follower, (modified Darlington if you like), as used, (for instance), In Selfs' so-called 'blameless'design, incorparates a shared emitter resistor in the complementary driver stage, which facilitate the reverse bias of the base-emitter junction of the output BJT being turned off.....This improves the rate of charge extraction significantly over that achievable with simple darlington,or sziklai derivatives........
P.S: said charge storage has nothing to do with Miller effect...., which cannot exist in an emitter follower anyway.....
janneman said:
...correct...
janneman said:
.....Not quite, the darlington, compound or otherwise thereof, provides no singular improvement with respect to speed of charge extraction compared to the sziklai.......
However, the double complementary emitter follower, (modified Darlington if you like), as used, (for instance), In Selfs' so-called 'blameless'design, incorparates a shared emitter resistor in the complementary driver stage, which facilitate the reverse bias of the base-emitter junction of the output BJT being turned off.....This improves the rate of charge extraction significantly over that achievable with simple darlington,or sziklai derivatives........
P.S: said charge storage has nothing to do with Miller effect...., which cannot exist in an emitter follower anyway.....
CFP and Doug Self
It is very popular to quote Douglas Self here. Let me do the same, this one is from his web article "Distortion in Power Amplifiers":
--------------
The other major type of bipolar complementary output is the Complementary Feedback Pair (CFP) or Sziklai Pair, seen in Fig 13c. The drivers now compare the output voltage with that at the stage input. Wrapping the outputs in a local NFB loop gives better linearity than EF versions with 100% feedback applied separately to driver and output transistors. The CFP topology is generally considered to show better thermal stability than the EF, because the Vbe of the output devices is inside the local NFB loop, and only the driver Vbe affects the quiescent conditions. The true situation is rather more complex. [10],[11],[12] The output gain plot is shown in Fig 16; Fourier analysis shows the CFP generates less than half the LSN of an emitter-follower stage. (See Table 4) It is hard to see why this topology is not more popular. The crossover region is much narrower, at about 1V. When under- biased, this appears on the distortion residual as narrower spikes than those from an emitter-follower output. Optimal bias here is 1.296V.
It is very popular to quote Douglas Self here. Let me do the same, this one is from his web article "Distortion in Power Amplifiers":
--------------
The other major type of bipolar complementary output is the Complementary Feedback Pair (CFP) or Sziklai Pair, seen in Fig 13c. The drivers now compare the output voltage with that at the stage input. Wrapping the outputs in a local NFB loop gives better linearity than EF versions with 100% feedback applied separately to driver and output transistors. The CFP topology is generally considered to show better thermal stability than the EF, because the Vbe of the output devices is inside the local NFB loop, and only the driver Vbe affects the quiescent conditions. The true situation is rather more complex. [10],[11],[12] The output gain plot is shown in Fig 16; Fourier analysis shows the CFP generates less than half the LSN of an emitter-follower stage. (See Table 4) It is hard to see why this topology is not more popular. The crossover region is much narrower, at about 1V. When under- biased, this appears on the distortion residual as narrower spikes than those from an emitter-follower output. Optimal bias here is 1.296V.
Mikek,
Thanks for those historical articles. Now I know who Lin is. I think the Bax article is the most useful of the three, even though Bax encourages thinking of transistors as voltage devices! None of the articles appear to address the stability problems that you and AKSA have suggested.
"P.S: said charge storage has nothing to do with Miller effect...., which cannot exist in an emitter follower anyway....."
I think the "miller effect" is a consideration in ALL configurations. It is the impact of parasitic current between capacitive junctions due to junction voltage change. This is why I pointed out that the effect of Ccb is just as important for a common-emitter as for a common-collector because in either case the Vcb variation is the same.
Thanks for those historical articles. Now I know who Lin is. I think the Bax article is the most useful of the three, even though Bax encourages thinking of transistors as voltage devices! None of the articles appear to address the stability problems that you and AKSA have suggested.
"P.S: said charge storage has nothing to do with Miller effect...., which cannot exist in an emitter follower anyway....."
I think the "miller effect" is a consideration in ALL configurations. It is the impact of parasitic current between capacitive junctions due to junction voltage change. This is why I pointed out that the effect of Ccb is just as important for a common-emitter as for a common-collector because in either case the Vcb variation is the same.
Hi pma
D. Selfs' stuff is pretty good,...but do'nt take everything he, (or indeed anyone else, including yours truly ) says at face value....
i think you'll find that he makes a couple of belated admissions in his book viz a viz the proneness to parasitic oscillation of the CFP based SEPP....see for instance, his recent article in EW on class G amps......can also be found in the third edition of his book....
D. Selfs' stuff is pretty good,...but do'nt take everything he, (or indeed anyone else, including yours truly
) says at face value....i think you'll find that he makes a couple of belated admissions in his book viz a viz the proneness to parasitic oscillation of the CFP based SEPP....see for instance, his recent article in EW on class G amps......can also be found in the third edition of his book....
Diametrically opposed listening
I think sometimes contributors here come from very different angles.
Having built various Self blameless configurations when they appeared in EW&WW years ago they all sounded s**t musically. They measured well, but I listen to music with my ears.
Having taken an interest recently, I'm just starting to get a handle on why. Self has contributed to my education, and for that he deserves thanks, but when it comes to designing amps that sound any good, I feel he has a long way to travel.
Andy.
I think sometimes contributors here come from very different angles.
Having built various Self blameless configurations when they appeared in EW&WW years ago they all sounded s**t musically. They measured well, but I listen to music with my ears.
Having taken an interest recently, I'm just starting to get a handle on why. Self has contributed to my education, and for that he deserves thanks, but when it comes to designing amps that sound any good, I feel he has a long way to travel.
Andy.
Hi alw..
i think the most significant short coming of Selfs' designs as published, is the extremly poor single slope S.O.A protection network, which will almost certainly be activated in normal use....and is very...very audible.........
Also, i would steer well clear of the CFP output stage .....
i think the most significant short coming of Selfs' designs as published, is the extremly poor single slope S.O.A protection network, which will almost certainly be activated in normal use....and is very...very audible.........
Also, i would steer well clear of the CFP output stage .....
Andy,
For once I seem to agree with you ...... I don't think Self listens to what he designs. Having built a coulpe of his designs myself, he sould be working for Pioneer or Yahama in their reciever department.
Trouble is, too many people seem to think he is God's answer to
amplifier design. Lucky for us, Mr. Pass is kind enough to share his designs with us and we don't have to blame our blameless amplifiers for bad sound.
Jam
P.S. And a happy New Year to you all.
For once I seem to agree with you ...... I don't think Self listens to what he designs. Having built a coulpe of his designs myself, he sould be working for Pioneer or Yahama in their reciever department.
Trouble is, too many people seem to think he is God's answer to
amplifier design. Lucky for us, Mr. Pass is kind enough to share his designs with us and we don't have to blame our blameless amplifiers for bad sound.
Jam
P.S. And a happy New Year to you all.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.