To reiterate:
1) There is no need for excessive bandwith in audio applications.
2) "CFAs" deliver significantly poorer linearity than VFAs.
3) "CFAs" deliver poorer PSRR than well designed VFAs.
4) There is no need for servos with well designed VFAs; this not true of "CFAs".
There is no conceivable excuse for using "CFAs" in audio applications.
Completely wrong! 🙁
Yes. You can only achieve higher SR and BW by going for MIC, or if non MC, some suboptimal comp scheme. And, whatever scheme you choose, VFA's always require more circuit complexity.
Just why would you want more bandwidth or slew than you need?😕
And what on earth is MIC?
It is relatively easy to design a VFA with 100V/uS slew rate using two pole compensation.
And what on earth is MIC?
It is relatively easy to design a VFA with 100V/uS slew rate using two pole compensation.
Last edited:
There is no need for excessively low distortion in audio.
CFA's easily deliver distortion orders of magnitude below the threshold of hearing, and at least 2-4 orders of magnitude lower than speakers and analog source material (specifically vinyl)
CFA PSRR is below audibility threshold, and is in any event easily solved with simple supply rail filtering
I have no servos in my two current CFA designs as the DC performance is very good ( under 5mV offset). Both designs are fully DC coupled.
VFA's tend to be more complexed expensive than CFA's and yet offer no audible benefit
CFA's easily deliver distortion orders of magnitude below the threshold of hearing, and at least 2-4 orders of magnitude lower than speakers and analog source material (specifically vinyl)
CFA PSRR is below audibility threshold, and is in any event easily solved with simple supply rail filtering
I have no servos in my two current CFA designs as the DC performance is very good ( under 5mV offset). Both designs are fully DC coupled.
VFA's tend to be more complexed expensive than CFA's and yet offer no audible benefit
There is no need for excessively low distortion in audio.
There is no such thing as "excessively low distortion".
Firstly, there is no need for an excessively large closed loop bandwith (say, greater than 2Hz~200KHz) in audio applications.
I have to disagree with the above statement. An amp with 200khz at the canonical -3db does introduce phase shifts in the audio band and is therefore sub-optimal.
Point 3 I think was correct (PSRR).
For the very basic CFA topology that Michael seems to be stuck into, yes.
For a true high performance CFA (intended for audio), no. Look at this: http://www.ti.com/lit/ds/symlink/lme49871.pdf
102dB PSRR
1.2ppm distortions in 600ohm load
Under 1mV offset.
Quoting the SR and bandwidth numbers for this CFA would be, in the context of audio, obscene.
Point 3 I think was correct (PSRR).
Start using CFA with CCS and PSRR is leveled with noise floor.

I have to disagree with the above statement. An amp with 200khz at the canonical -3db does introduce phase shifts in the audio band and is therefore sub-optimal.
gotta love the things people think thay can get away with saying - no matter how easy to refute
200 kHz single pole is -0.043 dB, 8 ns delta group delay at 20 kHz re 1 kHz
its not the 5.6 degrees or the 800 ns total delay - it is the deviation from a pure delay that is the time "distortion", and again that is 8 ns
the reason 200 kHz LP is "cannonical" for audio is because it is technically and by even audio guru listening tests inaudible
Last edited:
Start using CFA with CCS and PSRR is leveled with noise floor.![]()
Not true , wether you re using resistors , basic current sources
R + zener + R , active sources or even perfect sources (in simulators)
the PSRR will be the same +-0.5db.
Taking Gabor s design as workhorse , it has basicaly 34db worse
PSRR that an equivalently biaised symetrical differential VFA ,
and your VSSA , wich is the same design , has the same mediocre
PSRR , the current source argument being just wishfull thinking
but has not the slightest technical ground.
Attachments
For a true high performance CFA (intended for audio), no. Look at this: http://www.ti.com/lit/ds/symlink/lme49871.pdf
Quoting the SR and bandwidth numbers for this CFA would be, in the context of audio, obscene.
Perhaps that the most obscene is that this chip is obsolete according
to its datasheet.
One has to wonder why....
Perhaps that the most obscene is that this chip is obsolete according
to its datasheet.
One has to wonder why....
Simply because this chip, originally designed by National, is superseded, performance wise, by other products designed by TI. I'm sure you'll find your way in the TI catalog.
Real circuit measurements are the ones to trust. Modulation of DC rails resulting in lower than noise floor suppression of the output voltage modulation in VSSA, in both cases, observing the scope-analyzer screen, as well as listening to speaker noise. Usually VFAs I ever tested had more problems with the same rail modulation signal (DC+AC). Simulations are so approximate that I trust only the real world's measurements results. 😉Not true , wether you re using resistors , basic current sources
BTW CFAs are always better sounding performer in all aspects.

Last edited:
No doubt that TI has their own equivalent products but overall ,
i think that those low impedance inputs opamps are hardly finding
their ways given improved processes.
Theses topologies are convenient to design high speed op amps
using slow devices but once the components frequency transistion
is increased (GHZs ft transistors) the speed advantage gradualy shrink.
i think that those low impedance inputs opamps are hardly finding
their ways given improved processes.
Theses topologies are convenient to design high speed op amps
using slow devices but once the components frequency transistion
is increased (GHZs ft transistors) the speed advantage gradualy shrink.
I am surprised you are so anti CFA Wahab. At least, they are symmetrical!
Ignore that false claims on both VFA and CFA sides and think about the additional scope you have in your design arsenal!
Ignore that false claims on both VFA and CFA sides and think about the additional scope you have in your design arsenal!
If you choose to use a separate power rail for the VAS than for the driver and output stage, it is of little consequence to include a small signal (highish frequency) voltage regulator to better the PSR figures for a CFA circuit.(and VFA for that matter)
I see no reason not to analyze the shortcomings of a design and build around it to mitigate those shortcomings. The result becomes more ideal, and that should be the goal, no?
The last amp I built is made using the CFA definition and in fact uses a half wave voltage doubler to power the VAS rails. Vripple is ~0.5V @ 60Hz, yet PSR is very good as there is no trace of the Vripple sawtooth signal on the rails or the output. However, if the regulator is disabled then the output is struck with 60Hz BZZZZzzzzz.
I see no reason not to analyze the shortcomings of a design and build around it to mitigate those shortcomings. The result becomes more ideal, and that should be the goal, no?
The last amp I built is made using the CFA definition and in fact uses a half wave voltage doubler to power the VAS rails. Vripple is ~0.5V @ 60Hz, yet PSR is very good as there is no trace of the Vripple sawtooth signal on the rails or the output. However, if the regulator is disabled then the output is struck with 60Hz BZZZZzzzzz.
- Status
- Not open for further replies.
- Home
- Amplifiers
- Solid State
- Pros and Cons of current feedback amplifier.