Member
Joined 2009
Paid Member
This is what most people have not done - compare the *same* amplifier with both CFA and VFA.
Did you modify your VSSA or variant? Any shots? I have so many close-shot photos of my lovely amplifiers both common and unique type.
Actually both CFA and VFA could reach better sound quality.
Ad nauseam repeating the audible superiority for the CFA structure certainly has certainly been more detrimental to the discussion than the criticisms.
Suggestions have been emitted that some here lack the basic knowledge of electronics and transducers. The supposed weaknesses had not been highlighted so can't be discussed. What's a forum for if not to debate ?
A technical point of view
Current conveyors (CCs) have the same "diamond" input as the most common CFA structure. When CCs appear in IC form circa 2000, it was thought they were promised to a great future because circuits using them are very smart, easier to understand and to build and providing better performances at high frequencies than those based on OPAs with differential input. As far as I know, it's an historical fact by now that CCs did not encounter the expected success. I make the hypothesis that the cause relies in the reluctance towards using circuits which has DC operating conditions not as firmly defined as the differential pairs and which are more expensive.
Member
Joined 2009
Paid Member
it has many uses, not least is to provide a place where you can observe some old human behaviours whenever there is a collection of people - you will notice that some folk around here like to collect a following, some others like to become followers, some like to challenge those being followed and others just sit back and laugh 😎
I ll post the schematics in a relevant thread once i have
a working implementation in real life but i can already
tell you that it s not CFA compatible by the virtue of the
input stage wich is a differential whose two outputs are used
and no , i m sure that simpler wouldnt be possible.
In the waiting the THD1 , THD20 and IMD1920 numbers below,
30V Pk/8R with PSU +-50V , slew rate is 50V/us.....
I already pointed , and i m not alone , that this is not true at all.
If we take the most simple CFA that hang by there , that is Gaborbela
25 year old amplifier whose schematic has been recycled by Lazycat ,
adding two transistors will make this topology a symmetrical differential,
so more complexe by two transistors but this is without counting the
increased complexity of said CFA power supply to counter the inherently
lower PSRR , in short what is removed from the amplifier is transfered
to the PSU and regulation side making the whole amplifier more complexe...
Attachments
Why don't you post the *.ASC of this circuit so we can check your analysis with different tools and also apply PSR tests to both circuits.
It may be possible to improve VSSA without adding complexity too.
I am extremely happy with my CFA designs. I've built both and listened to both topologies over an extended time period.
I do not recognize that CFA's are inferior.
What I do see is that for the most part they have traded uneccessarily low distortion, for better slew rates, wider loop gain bandwidths and less complexity than VFA's with equivalent performance.
I do not recognize that CFA's are inferior.
What I do see is that for the most part they have traded uneccessarily low distortion, for better slew rates, wider loop gain bandwidths and less complexity than VFA's with equivalent performance.
I've long come to the conclusion that the IPS/VAS topology doesn't matter when you're using a real outputstage, either 3EF or MOSFETs. It doesn't matter. And from that point on, set your own design goals like level of complexity, amounts of THD, stability margins etc.
Facts, baby, facts hehe
Are we supposed to believe to such non sense that you quoted.?
So there are amplifiers that can magicaly remove a part of the signal
with far better precision than any DSP since they are still surely
linear enough to be hifi , i guess that they have not enough
slew rate to only reproduce thoses "voices" accurately....
This kind of "argument" is just plain insult to people intelligence
and i m not exagerating , i hope that the less trained members
will do the effort to grasp how an amplifier work with enough
insight that they ll be immune from such marketing and indeed
interest vested claims...
As is visible with the graphs i posted i dont use LTspice but Simetrix
and the files are not compatible , for the record i had to draw all schematics
that i simulated , including yours of course.😉
That said there s no meaningfull difference between said schematic
and its differentialized extension linearity wise for simple implementation,
the obvious differences will lie in PSRR and also SR to some extent.
Once things are complexified a little CFA become less easy to improve.
That's the essential, indisputable point.
It can be said just the inverse : CFA trade low distorsion and high PSRR to non necessary wide loop gain bandwidths and better maximal slew-rates.
I think the aesthetics of the symmetry in CFA schematics has quite an appealing influence on people who think CFA technology is better.
I do not agree about complexity. CFAs need more reservoir capacitance which is cumbersone. If IS and VAS need stabilized power supplies by capacitance multipliers, it adds complexity, more test points have to be controlled, and it's time consuming.
A large difference between VFAs and CFAs is that the CFA voltage stages have their DC currents and operating points set by the same devices which also handle the AC signals.
In VFAs, the voltage stages are DC controlled "externally" by CCSs, which make the circuits more dependable and needing less tests.
Bonsai, that is a good summary but I would go further.
You, Dadod, Esperado & others have shown that CFAs do not HAVE to trade low THD for better slew etc. They can achieve low THD AND zillion V/us slew AND good PSR ... and use less complex circuits to do so.
If you are prepared to use evil electrolytics, you can have lower noise too.
________________
Wahab & forr, my offer is still open. Post an ASC or a pic of a simple VFA which you feel has good performance. You specify what performance aspects are important including PSR if you wish.
I will confirm your SPICE world performance and attempt to better them with a simpler CFA .. at least in SPICE world.
________________
If you don't feel sound is important or that the claims for sonic superiority are rubbish, just ignore them and concentrate on the objective measures in your proposed simple VFA as above. 🙂
If you don't feel 1 pp zillion THD or zillion V/us slew is important, then propose a VFA with a lesser spec.
But please don't complain if we achieve better specs all round with a simpler circuit.
_________________
BTW, I don't think CFAs are perfect. They have important faults that I will only have a good handle on if I was making them in at least small scale production. But most of these faults are cos the symmetrical topology rather than cos they are CFAs.
Are-you really serious ? Did-you listen music only across simulators ?
When even the sens of a power chord, the wiring of a line signal cable can change the "definition" of a system...
Now, i understand everything, we are living in a different planet...
I wonder what kind of enclosures and source you use...
You are kidding. The response is the most referenced amp circuit in this forum, and it is both the most ever simulated and the most ever built.
To me, simulation is only to prepare building, its virtue is of an indicative, not decisive, nature . I have always suspicions towards simulation which have not been confirmed by finished, working, and measured circuits.
Simulation is nothing else than a game delivering a virtual sound which does not move air particles. I get a bit bored by the fact it has so much overcome fora at the detriment of realizations.
As seen on your web site, you have knowledge in photography.
So you can't deny that, in photography, "definition" is an objective concept.
How can it be not the same in audio ?
Member
Joined 2009
Paid Member
There are comments that are not technical and appear to read anti-CFA. But there are also plenty of comments that are not technical espousing how much better CFAs are than everything else. Each side encourages the other to respond and few if any are impartial. However, a technical-only discussion relies on simulations and measurements but this leaves you with a dead-end since there is the issue of how technical results relate to what we experience when we listen to the amplifiers.
How can the argument be reduced to VFA vs CFA? It is well known that within the Class of VFA, especially the input stage determines the subjective sound character to a large extent, viz., LTP, Rush Cascode, Singleton, Complementary etc.,
Similarly, if the CFA topology is better understood with more reasonable dialogue than witnessed in this thread, we may be able to understand better, the effect of different input stage configurations on CFA sound character.
As Bonsai has been repeatedly pointing out, it may finally boil down to a particular circuit and its parameters such as stability, thermal issues etc., rather than a generic CFA vs VFA debate.
Both should have equivalent performance, all other things being optimized per circuit/schematic and ancilliary circuits like PSU, Protection etc.,
Similarly, if the CFA topology is better understood with more reasonable dialogue than witnessed in this thread, we may be able to understand better, the effect of different input stage configurations on CFA sound character.
As Bonsai has been repeatedly pointing out, it may finally boil down to a particular circuit and its parameters such as stability, thermal issues etc., rather than a generic CFA vs VFA debate.
Both should have equivalent performance, all other things being optimized per circuit/schematic and ancilliary circuits like PSU, Protection etc.,
Member
Joined 2009
Paid Member
Thanks for the insight but i m also musician when
i have some time left for the last 30 years or so ,
no need to come with thoses useless arguments
to try to make an easy point , as you can see
it doesnt work and you would be even hard pressed
if keeping on this debate , better to forget about it.