Re: Reply
Current feedback has been rather unkwon to me also until a year ago. I have had no reason to study it but there are lot's of info on the net in the subject.
I think you have missed the pedagogical side of the subject. Current feedback is rather unknown for most people. You can't really argue about it. It's just facts have you create the feedback. Nothing to discuss really.amp_man_1 said:
Current feedback has been rather unkwon to me also until a year ago. I have had no reason to study it but there are lot's of info on the net in the subject.
Hi peranders,
Because in VLSI, transistors are cheap and have fine matching,
whereas resistors are difficult and of poor reproduceability.
Building from discretes, matters are -sort of- the other way around.
Leaving all your settings intact and adding
.fourier 20k V(vout)
I get 0.0068% THD for both your original circuit and
when replacing the current sources with 11k resistors.
Regards,
Peter Jacobi
peranders said:
Because in VLSI, transistors are cheap and have fine matching,
whereas resistors are difficult and of poor reproduceability.
Building from discretes, matters are -sort of- the other way around.
peranders said:
Leaving all your settings intact and adding
.fourier 20k V(vout)
I get 0.0068% THD for both your original circuit and
when replacing the current sources with 11k resistors.
Regards,
Peter Jacobi
Re: Reply
Oh dear!!!
Maybe you should spend a little more time trying to understand
Per-Anders' design before you try to adopt ideas from it in
your quite different amplifier?
amp_man_1 said:WITH REFERENCE TO PERANDERS INFORMATION A NEW SOLUTION IS PREPOSED , which is very well illustrated in this schematic.
Oh dear!!!
Maybe you should spend a little more time trying to understand
Per-Anders' design before you try to adopt ideas from it in
your quite different amplifier?
A honest attempt to nomenclature clarification
In all references I have easily at hand, starting with the 1978 "Berechnung von Verstärkerschaltungen" by H. Römisch, a total of four (!) cases of feedback is differentiated:
A) Voltage controlled voltage feedback
B) Voltage controlled current feedback
C) Current controlled voltage feedback
D) Current controlled current feedback
Case A) is what is seen a most amplifiers today and in standard integrated OpAmps.
Case B) is seen in the CFA OpAmps (EL2020 etc), in amplifiers having a diamond transistor input stage, and -nearly forgotten- in all ye olde amps before the invention of the LTP, see for example "El Cheapo":
http://sound.westhost.com/project12.htm
Case C) and D), where the feedback is controlled by the load current, is certainly more rare, and I cannot comment on its virtues at all.
The amp_man_1 design is for sure intended to be case B), and
without the dreaded "true" in the thread title, less fighting may have resulted.
Whereas using the correct four case terminology would have avoided any confusion (but with a greater risk of carpal tunnel syndrome), proponents of the C/D designs feel that the B designs have unjustly hijacked the term "current feedback", and may want to call their designs "true current feedback".
In all references I have easily at hand, starting with the 1978 "Berechnung von Verstärkerschaltungen" by H. Römisch, a total of four (!) cases of feedback is differentiated:
A) Voltage controlled voltage feedback
B) Voltage controlled current feedback
C) Current controlled voltage feedback
D) Current controlled current feedback
Case A) is what is seen a most amplifiers today and in standard integrated OpAmps.
Case B) is seen in the CFA OpAmps (EL2020 etc), in amplifiers having a diamond transistor input stage, and -nearly forgotten- in all ye olde amps before the invention of the LTP, see for example "El Cheapo":
http://sound.westhost.com/project12.htm
Case C) and D), where the feedback is controlled by the load current, is certainly more rare, and I cannot comment on its virtues at all.
The amp_man_1 design is for sure intended to be case B), and
without the dreaded "true" in the thread title, less fighting may have resulted.
Whereas using the correct four case terminology would have avoided any confusion (but with a greater risk of carpal tunnel syndrome), proponents of the C/D designs feel that the B designs have unjustly hijacked the term "current feedback", and may want to call their designs "true current feedback".
Pjacobi,
I don't think the word "true" mattered much. There are
books and other references using the terminology differently.
Maybe it is only the older literature, though. For instance
Schilling&Belove, which I referred to earlier does distinguish
four cases too, but they use the terms voltage/current feedback
to refer to the load sensing rather than the error signal. I had
learnt the terminology this way and was thus very confused
for a long time until I eventually found some good app. notes
that made me realize the terminology is nowadays often
used differently. It is evident both from this thread and many
previous ones that there are many who have learnt to use the
terms in the same way as me, which is why they are confused,
and I am sure they did not all learn it from the same book. It
seems obvious that the terminology has been redefined or
"split" somewhere along the road.
It is interesting to hear that you have found textbooks that
define a "modern" and different terminology. I have asked
before what other peoples modern textbooks say about this,
but so far I can't remember anybody finding anything about
this in their books. So until now, it has only been a terminology redefined by the op amp manufacturers. An interesting
question is what terminology the universities teach on this
nowadays. Anybody out there having a fresh EE degree?
Definitions are unfortunately not always universal. They can
differ between countries, between universities, between
academia and industry etc. etc. That is confusing but something
we have to live with until somebody standardizes it. That may
not help either, though. I still remember a quote from
Tannenbaums book we had in the computer networks course
20 years ago:
"The best thing about standards is that there are so many
to choose from. Furthermore, if you don't like any of them,
just wait for next years model."
I don't think the word "true" mattered much. There are
books and other references using the terminology differently.
Maybe it is only the older literature, though. For instance
Schilling&Belove, which I referred to earlier does distinguish
four cases too, but they use the terms voltage/current feedback
to refer to the load sensing rather than the error signal. I had
learnt the terminology this way and was thus very confused
for a long time until I eventually found some good app. notes
that made me realize the terminology is nowadays often
used differently. It is evident both from this thread and many
previous ones that there are many who have learnt to use the
terms in the same way as me, which is why they are confused,
and I am sure they did not all learn it from the same book. It
seems obvious that the terminology has been redefined or
"split" somewhere along the road.
It is interesting to hear that you have found textbooks that
define a "modern" and different terminology. I have asked
before what other peoples modern textbooks say about this,
but so far I can't remember anybody finding anything about
this in their books. So until now, it has only been a terminology redefined by the op amp manufacturers. An interesting
question is what terminology the universities teach on this
nowadays. Anybody out there having a fresh EE degree?
Definitions are unfortunately not always universal. They can
differ between countries, between universities, between
academia and industry etc. etc. That is confusing but something
we have to live with until somebody standardizes it. That may
not help either, though. I still remember a quote from
Tannenbaums book we had in the computer networks course
20 years ago:
"The best thing about standards is that there are so many
to choose from. Furthermore, if you don't like any of them,
just wait for next years model."
Hi Christer, All,
1978 surely predates integrated CFA OpAmps, so it's only sort-of modern.
Googling for "current feedback", restricted to .edu domains, gives some empirical evidence. Most hits agree with the manufacturers.
Northern Illinois University Physics department agrees with you, giving following table for the four types:
Sample output - Type of input - Name
Voltage - Voltage - Voltage feedback
Current - Voltage - Current feedback
Voltage - Current - Operational voltage feedback
Current - Current - Operational current feedback
http://niufrm.physics.niu.edu/~labelec/lect/p375_lect121.pdf
To add another quote (unknown author, plagiarizing a quote attributed to Bismarck):
"Standards are like sausages. It's better not to be there, when they are made."
Regards,
Peter Jacobi
1978 surely predates integrated CFA OpAmps, so it's only sort-of modern.
Christer said:
Googling for "current feedback", restricted to .edu domains, gives some empirical evidence. Most hits agree with the manufacturers.
Northern Illinois University Physics department agrees with you, giving following table for the four types:
Sample output - Type of input - Name
Voltage - Voltage - Voltage feedback
Current - Voltage - Current feedback
Voltage - Current - Operational voltage feedback
Current - Current - Operational current feedback
http://niufrm.physics.niu.edu/~labelec/lect/p375_lect121.pdf
Christer said:I still remember a quote from
Tannenbaums book we had in the computer networks course
20 years ago:
"The best thing about standards is that there are so many
to choose from. Furthermore, if you don't like any of them,
just wait for next years model."
To add another quote (unknown author, plagiarizing a quote attributed to Bismarck):
"Standards are like sausages. It's better not to be there, when they are made."
Regards,
Peter Jacobi
Have you considered that current and voltage are inextricably related to each other by impedance? Feedback is feedback, whether you describe it as current feedback or voltage feedback doesn't matter a wet slap and you can correctly identify any practical system as using either voltage feedback or current feedback as you wish.
Re: Reply
Mr amp_man, why have you changed your design? Do you know any advantage over this new output stage? I'll guess that you don't know what you are doing here, right?amp_man_1 said:WITH REFERENCE TO PERANDERS INFORMATION A NEW SOLUTION IS PREPOSED , which is very well illustrated in this schematic.
traderbam said:Have you considered that current and voltage are inextricably related to each other by impedance? Feedback is feedback, whether you describe it as current feedback or voltage feedback doesn't matter a wet slap and you can correctly identify any practical system as using either voltage feedback or current feedback as you wish.
Don't agree. There is a clear way to differentiate between the two. If you disconnect your load, and the feedback signal goes to zero, that's current feedback (load current > zero, feedback > zero). If, on the other hand, you short the load, and the feedback goes to zero, that's voltage feedback (load voltage > zero, feedback > zero). You are right that I and V are related, but they ARE different. Otherwise we could do with just one word like vurrent (or coltage).
Jan Didden
pjacobi said:
Just to avoid more confusion, it does not agree with me, just
with my old textbook (and many others textbooks, it seems).
I had already changed my use of the terminology to the "modern"
one, which should be obvious since I "defended" amp_mans
design as being CFB.
However, since I had previously learnt the terminology in the
"old" way, I am aware of the ambiguity. Hence, I can now often
spot the reason for confusion in CFB discussions, which
is impossible for those who only knows one of the two
definitions. Until all of us who learnt it in the "old" way have
either relearnt or passed away, the same confusion will arise
over and over again.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.