Pertinent question !Is an emitter follower a CFA ?
Mona
If the answer is yes, the input circuit of a VFA (long tail pair) is practicallly built with two CFAs !
The entire subject HERE is a mess. and some in the marketing is a mess. Just read the book(s) I suggested. They should answer all your questions best.
THx-RNMarsh
Wait a week, and somebody will start a new topic, "How to build a CFA"
By ear, that would be much appreciated by certain contributors here, with an allergy to mathematics.
I found this on Bonsai’s website Hifisonix:
He has a series of tests to determine if a circuit is a VFA or a CFA. From his article:
“H bridge input amplifiers appear topologically similar to a classic CFA with the inverting feedback network input buffered by a second diamond stage, mirroring the non-inverting input diamond buffer. A resistor connected between the summing junction of the two buffers sets the front end stage gm, allowing very wide bandwidths and high slew rates. The H bridge input stage would test out using the postulates in Table 1 as a CFA – the peak TIS current is set by the buffer coupling resistor, and it is not constant gain bandwidth limited like a VFA – a fact I easily confirmed in simulation. However, the IC industry classifies it as a VFA and so we will leave it at that – in certain cases the debate as to whether an amplifier circuit is VFA or CFA will remain a contentious one.”
He is actually saying that the H-bridge (which is my preferred VFA IPS) in fact “in his opinion” is a CFA.
His opinion is of course completely wrong. The H-bridge is a VFA IPS.
BTW: in an earlier post I should not have said “floating” but “driven” sorry for the confusion.
S
He has a series of tests to determine if a circuit is a VFA or a CFA. From his article:
“H bridge input amplifiers appear topologically similar to a classic CFA with the inverting feedback network input buffered by a second diamond stage, mirroring the non-inverting input diamond buffer. A resistor connected between the summing junction of the two buffers sets the front end stage gm, allowing very wide bandwidths and high slew rates. The H bridge input stage would test out using the postulates in Table 1 as a CFA – the peak TIS current is set by the buffer coupling resistor, and it is not constant gain bandwidth limited like a VFA – a fact I easily confirmed in simulation. However, the IC industry classifies it as a VFA and so we will leave it at that – in certain cases the debate as to whether an amplifier circuit is VFA or CFA will remain a contentious one.”
He is actually saying that the H-bridge (which is my preferred VFA IPS) in fact “in his opinion” is a CFA.
His opinion is of course completely wrong. The H-bridge is a VFA IPS.
BTW: in an earlier post I should not have said “floating” but “driven” sorry for the confusion.
S
Oh my! The H-bridge has a fixed transconductance set by the bridging resistor and the re's of the input devices and a fixed Ccomp, i.e. the GBW product is constant (VFA). The slew on demand has nothing at all to do with VFA vs. CFA.
@Waly - Harvard radio does music orgies during reading period, they used to include a more diverse selection of genres, I have 16 hr's of nothing but Eye somewhere in my archives.
Come on man. I did some sims and H bridge is fixed bandwidth - you can vary the gain over an appreciable range and the bandwidth is constant and that's why I said it behaved like a CFA.
I also go on in the same document a few lines further down, if you bothered to read it, to say the semiconductor industry classes it as VFA and so just go with the flow.
The only difference between a standard CFA and an H bridge is that the feedback node is buffered - so the peak current into the VAS input is then set by the bridge tie resistor - its not limited by a current source as it is in a VFA.
And here you see exactly the problem of blindly applying canonical feedback forms and expecting to understand the fundamental differences between the two: in one case the peak current into the VAS input is set by a current source (VFA) and in the other, its set by the peak output voltage and the value of the feedback resistor.
I'll don my asbestos fire suit on and await the sh1t storm that will surely envelop me.
Have a good evening all. I'm going to pour myself a whiskey.
Of course it's a mess, that's what you get when a term is reused to mean something different from what it originally meant, but something that still has a relation to the original meaning.
Anyway, there are terms that are even messier, such as linear.
Come on Mr. Wurcer, I can't believe they were playing Hataranashi then. I am told they were banned here, together with a bunch of other bands, US GG Allin and UK Screwdriver included.
I have an acquaintance claiming he saw Hataranashi live, and barely escaped from a bulldozer wielding on the stage.
Last edited:
There should be a large agreement that the CFPs belong to the toplogy called CFA. The simplicity of the circuit is an invitation for the members of each camp to expose their own analysis to justify or to refute the concept of current feedback applied to a kind of topology.
Michael Kiwanuka used it in its 'Refutation of Jung Et All' :
Current Feedback and Voltage Feedback Dispute.
Yes a buffered CFA is a VFA, I think I have posted something like that earlier in this thread and many years ago in other threads and forums.
"the peak current into the VAS input is set by a current source (VFA) and in the other, its set by the peak output voltage and the value of the feedback resistor."
This is completely wrong. BTW a VFA is not the same as a LTP with a CCS and a so called VAS and an OPS.
"I'm going to pour myself a whiskey."
Cheers enjoy your whiskey
S
The problem is not that I haven't read the literature but that I am not sure exactly what it is that you claim.
If you believe it is explained in a book then can you provide a reference that is useful?
That is - which book, what section, p. reference preferably.
Best wishes
David
Kiwanuka said:
This nicely summarizes MK's position. While indeed calling the CFAs "current feedback" is in violation of the basic negative feedback topologies (as many as 4), being in fact exactly "voltage feedback", calling the CFAs a "pointless topology" and the authors "deluded" is absolutely garbage, and he should be ashamed of.
I would 100% agree though with his evaluation when it comes to audio amplifiers.
Don't quote out of context.
It's quite acceptable for analysis purposes to treat an amplifier as an input stage block, a VAS (or TIS/TAS) stage and an OP stage. I had to explain this to another member a few pages back. See Putzeys, Solomon, Franco, Middleton, Friederiksen etc etc etc
I'll get another whiskey.
Last edited:
No, you only quoted that.
And then you are talking about "trolling and obfuscating".