A symmetrical differential has none of the blameless compromises
and best it in perfs particularly IMD wich is even more important
than THD.
The point is to have the best performing design without any
compromises and in this respect i dont buy the CFA argument ,
not even in matter of slew rate , you would be hard pressed
if asked what are the parameters besides THD that matters
and if CFAS are adequate to improve them or at least equal VFAs...
I think that goinf CFA you ll have to compromises all possible
parameters set apart slew rate , quite a meager trade.
Last edited:
Nonsense Wahab.
You are fixed on the symmetrical VFA and low THD.
You've produced sims of designs you never build, and claim distortion levels you cannot measure.
But, at least I build the stuff I sim, keep an open mind and don't sprout anti VFA diatribe.
You are fixed on the symmetrical VFA and low THD.
You've produced sims of designs you never build, and claim distortion levels you cannot measure.
But, at least I build the stuff I sim, keep an open mind and don't sprout anti VFA diatribe.
What are the parameters besides THD that matters
according to you and is a CFA the answer to yield
good results in thoses parameters.?.
Actualy i tested tons of the simulated circuits and i didnt
settle on a given topology by chance or by fashionism.
Why don't you tell us what the parameters are? Especially since, as I remarked earlier, other elements in the signal chain have distortion that is orders of magnitude higher. Why then do you focus endlessly on this metric?
Show us some of your circuits - real circuits, not just sims - that you have built to test out your claims.
Show us some of your circuits - real circuits, not just sims - that you have built to test out your claims.
THD of course but also IMD , PSRR , CMRR , DF , SR , well ,
all possible parameters and you can be sure that in none,
excepted SR , you ll have a CFA leading the lot.
I settled on the current mirrored symmetrical differential ,
the only part i posted was the input stage , the same
topology as published by Bob Cordell but with the current
mirror feeder enclosed in the differential local loop , i ll post
the complete schematic once i ll have a valuable soundcard
to at least produce some real world measurements.
As for simulations , well , since i cant build all possible
projects from everybody at least i took the time to draw
and simulate all possible amplifiers that hang by there ,
including a few of yours , as well as tons commercial
offerings , hence my usual skepticism once one come
with a schematic that is branded as a cure to all previous
shortcomings.
I have never claimed CFA is a cure all, as you claim VFA is.
Unless you build a few of these things and listen, don't go trashing them.
The problem when you try to define an amplifiers performance based on sims and guru-speak is you make statements that in practice either don't hold, or are of so little consequence as to be meaningless.
No one, after listening to one of Pass's amps with .025% distortion would say its inferior, would they? So why do it here?
Unless you build a few of these things and listen, don't go trashing them.
The problem when you try to define an amplifiers performance based on sims and guru-speak is you make statements that in practice either don't hold, or are of so little consequence as to be meaningless.
No one, after listening to one of Pass's amps with .025% distortion would say its inferior, would they? So why do it here?
I have been studying this thread and the earlier thread http://www.diyaudio.com/forums/soli...voltage-feedback-how-do-i-see-difference.html. I have also done a variety of simulations and builds of F5, F5-balanced, F6, F6CC, F4, and BA3FE. The simulation in this http://www.diyaudio.com/forums/soli...tage-feedback-how-do-i-see-difference-80.html posting provides an interesting insight about why CFA might outperform VFA. It boils down to the fact the feedback in CFA is coupled using a common-gate configuration whereas the feedback in VFA is coupled using a differential pair which combines common-source and common-gate. Run some sims of these two circuits and see the differences in both THD and in the AC analysis.
Yes, I would like to see this one.
It s a good starting base , the final schematic use a couple
more transistors as well as a lateral fet OS but believe it or not
this simplified version already worked in real world with careful
choices of transistors , actualy i used 2SA872A/2SC1775 for
the whole input stage excepted its CSs.
Btw , while we re at it , did you even provide the slightest
design by there , set apart your schematic of anthology when
computing a stage caracteristics.?..
This is from Texas Instruments by Xavier Ramus:
Distortion Issues
• At lower frequencies, the lowest distortion will be given by voltage feedback amplifiers
– We believe this is due to a linearity floor set by the error sensing point in the CFB
topology. The CFB inverting input linearity sets a floor to distortion much higher
than the best VFB designs. This is normally a 2nd harmonic term.
– CFB will, however, give relatively constant distortion vs. Gain setting and hold
better numbers to higher frequencies due to considerably more slew rate margin.
Unfortunately there is no reference to exactly what designs was used to make this statement.
Distortion Issues
• At lower frequencies, the lowest distortion will be given by voltage feedback amplifiers
– We believe this is due to a linearity floor set by the error sensing point in the CFB
topology. The CFB inverting input linearity sets a floor to distortion much higher
than the best VFB designs. This is normally a 2nd harmonic term.
– CFB will, however, give relatively constant distortion vs. Gain setting and hold
better numbers to higher frequencies due to considerably more slew rate margin.
Unfortunately there is no reference to exactly what designs was used to make this statement.
Nothing to do with audio applications. "Higher frequencies" means 1 MHz and over. E.g. for DSL drivers (where HF distortions are critical and gain is < 10), CFBs almost have exclusivity.
Most likely the conventional "diamond" complementary input stage, current mirrors, and unity-gain buffer output stage, the topology that Comlinear managed to patent years ago but which was either contested or paid little attention to.
The remark about 2nd harmonic almost certainly can be explained by the lack of ideal complementarity of PNP and NPN.
No, only a lack of even distortion cancellation mechanism. Doesn't require p/n ideal complementary, but only p/p and n/n. See the differential LTP loaded with a current mirror.
ASICs design use cells library and building blocks whose
caracteristics are fully modeled , hence it s rather copy
and paste of functional circuits , nothing impressive , mind you...😀
But anyway , have you something that is related to audio
rather than a low voltage circuit that work literaly only
in small signal conditions , far from what is discussed here ?.....
Yes, it is mine. Not so simple now though.
Here is a link to where the original DTMC lives.
PMP Amplifier
I took this scheme and slightly adapted it to my purposes.
The resistor + cap loading on the VAS/TIS output is applied to remove the peaking in response that the enhanced VAS/TIS seemed to suffer from in this amp. It also increased the GM / PM for the amp.
This VAS/TIS loading also allowed the use of the pre drivers as buffers for the DTMC network. In simulation it appeared to be stable and resulted in a small THD reduction. This compensation is also present on the MCP amplifier but at values that present a much lighter load to the VAS/TIS.
The fb (FB) node is actually the output of the amplifier. Sorry for the not so high quality ltspice schematic.
My miniscule brain is still working on the intricacies of this compensation method. The bit I'm really struggling with is the need for an RC network across the feedback resistor. It has a large affect on stability.
This DTMC compensation looked interesting and at that point there wasn't much out there about CFA compensation. Ended up simulating it and got reasonable results then built it and it eventually worked.
Bonsai,
He He. Thank you.
But it's not that complex really if you look at it as individual stages. E.g. remove the CCS components and substitute a perfect current source. It's probably not even that good a design, no measurements have been done yet as I don't have the equipment. I'm just happy to have designed an amp and got it working. It was one of my goals in life.
There is quite a bit of circuitry to cope with overload conditions, for example, clipping behaviour and to ensure it doesn't destroy itself if abused.
When I placed the orders for the PCBs and components I thought to myself "now I've done it". It wasn't as problematic as I was expecting to get working.
But after all of the building and some teething problems it does now work. Learnt a lot which is the main thing.
Hoping to do a mk2 version after this thread has evolved further.
Paul
Are you for real? Anyway, you made my evening, ROTFLMAO
That s not the place but let s say that sticking 10000
transistors in an ASIC and yielding a functional circuit
that will be still adequate when released in markets that
are evolving very fast forcibly require using building
blocks libraries of functional blocks that are constantly
improved but still use previous models as basis.
Or unless it is hand made..?...🙄
Edit ; still waiting for an audio related item
if ever there will be one...
Last edited:
Why do we dismiss our ear/brain reaction as a measure of good/bad sound and take what we see on the scope and distortion meter as a better measure?
It's like 'if there are nine zeroes after the dot and that square has so well defined perfect cliff like edges then the thing must deliver superb cymbals'. 😕
I mean, are we hearing sampled sinewaves or zeros or right angles through the speaker?
I am not adding to 'VFA vs CFA' fyi...
What I find to be interesting is that many people in the forum agrees that simple CFAs give superior sound quality than their VFAs. Also there are many who say the opposite. The latter guys make it their belief that VFA in general is the best suitable topology for audio. While I have no problem with that notion, it is kind of a tragedy to see many of them ridicule the guys who believe in the other topology to be the best as 'self-delusional' and 'distortion lovers'.
----------------------------------
Hello Bonsai.
By simple CFA which particular design do you mean? I do want to see the nominal CFA to become one of the top notch zero making machine and learn to build it myself. Please treat me as a student.
I guess it's the one that's shown in Mr.Self's book or maybe the one shown in the Wikipedia 'article' on CFB op-amps.
Whatever it is I'll follow this. Since there is a lot to learn in CFA which IMO has little to offer to typical 'consumers' and demands sharp ears, not eyes, as the audience.
On-topic. 😉
It's like 'if there are nine zeroes after the dot and that square has so well defined perfect cliff like edges then the thing must deliver superb cymbals'. 😕
I mean, are we hearing sampled sinewaves or zeros or right angles through the speaker?
I am not adding to 'VFA vs CFA' fyi...
What I find to be interesting is that many people in the forum agrees that simple CFAs give superior sound quality than their VFAs. Also there are many who say the opposite. The latter guys make it their belief that VFA in general is the best suitable topology for audio. While I have no problem with that notion, it is kind of a tragedy to see many of them ridicule the guys who believe in the other topology to be the best as 'self-delusional' and 'distortion lovers'.
----------------------------------
Hello Bonsai.
By simple CFA which particular design do you mean? I do want to see the nominal CFA to become one of the top notch zero making machine and learn to build it myself. Please treat me as a student.
I guess it's the one that's shown in Mr.Self's book or maybe the one shown in the Wikipedia 'article' on CFB op-amps.
Whatever it is I'll follow this. Since there is a lot to learn in CFA which IMO has little to offer to typical 'consumers' and demands sharp ears, not eyes, as the audience.
On-topic. 😉
Maybe I'm envisioning a different topology, but it would seem for the "standard" CF the asymmetry of P and N would account for it. If we're talking about voltage feedback and differential pairs, yes, if common-mode distortion is minimized then I agree we null even-order. I don't see how one achieves that with current feedback configurations of the typical sort, unless you go with a balanced pair of "diamond" input stages and preserve balanced throughout.
I do like the ideas around the very simple circuit with the independent feedback to each transistor, as that has the potential of a trim of certain components for ~nulling even-order. It suffers from a variety of other issues though.
Thinking about all of this, particularly the notion that the CFA is inherently compromised for distortion performance at lower frequencies compared to the VFA, I have some ideas I may present soon, that I haven't seen before but of course may well exist elsewhere. As I say I came to the thread late.