As far as precise points of my ultra-basic simulations are not denied, I've not great interest to look at complex circonvolutions which tend towards drowning the fish (french expression) and incidentally meet disagreements with other persons than me.
Some here seems to think that it's unnecessary with a Ccomp in a CFA and one person thinks that it's not necessary to use a Ccomp in a H-bridge amp because it is based on a CFA.
Nothing new it appears all the time like the JLH amp with 2N3055's the devices are so slow the amp is stable by accident not design.
Sine 1000 Hz is not DC. You do not seem aware that it is the signal I used (1 Vrms), the DC conditions being set by ideal voltage and current sources.
Adding a grounded 100 pF in the collector of the second stage (VAS) begins to show dissimilar currents in the CFA and VFA inputs. But the topic of the thread is not the behavior of these circuits at high frequencies.
This topic is about the meaning which can be given to current feedback for a circuit where the involved currents are controlled by voltages.
A few months after the first publication of its class A amp in Wireless World (1969), JLH proposed to add R and C stabilizing components.
Big Cob of caveman silicon is the compensation , a good way to skimp on
components. Then like microsoft , the public abuses (uses) the product. Then skimping
corporation decides in a design change to "save face".
OS
If you don't consider them, how can you characterize them? And if you paid any attention, you'd find no one disagreed; it became clear finally that they thought them not relevant, because the point they made was already proven.
Regardless, you ignored them for quite a time before their relevance was challenged.
But I've really accomplished something - I've managed to support the claim that CFAs are VFAs while simultaneously producing complex circumlocutions! Pretty impressive, huh?
Last edited:
Much higher THC , much slower slew. You do low Z FB from the OP ??
PS - I still do not understand the reason for this thread ... another "blowtorch"
clone. I thought V1 of the CFA thread would settled this "debate". ??
Differences in sound, compensation , and other behaviors generally prove
CFA is different from VFA , all this leaves is a pi$$ing contest in mathmatics
.OS
Differences in sound, compensation , and other behaviors generally prove
CFA is different from VFA , all this leaves is a pi$$ing contest in mathmatics
That exercise was basically useless in any practical sense, just an example of how the current matters.
Last edited:
No , bode plots , real world behaviors under load , compensation schemes.
"rolling component values" is BS. making sub ppm VFA's and CFA's is NOT.
OS
In the mythology of Control Theory, the world was created from vibrations of the highly revered sacred Feedback Mantra. Feedback is everything, feedback is everywhere, feedback is greater than God.
IanHegglun,
No feedback is occurring.
Any resistor in any position reduces amplitude (voltage) and the amount of excited charges (current) by presenting an opposing force, thereby is degenerative. You would not go as far as to claim that resistors generally introduce feedback, would you? It would hardly mitigate the prevalent conceptual chaos increasing with every new post.
It can - but it is not.
By the way, proportionality (linearity) is a mathematical relationship. Maybe for you, like for all engineers and most scientists, physics and mathematics are the same thing. No discrepancies arising, but nothing being understood or explained either.
I refute that emitter degeneration implies feedback. The base and collector resistor are degenerating bias elements as well.
No. A connection to the output of subsequent stage must exist.
No. Feedback is established by R5, it is a prerequisite of CFP and CFA, the emitter resistor is not. The emitter resistor is a way to determine a desired stable operating point and to influence transitions properties, but need not to be there for the definition of common-emitter amplifier, CFP or CFA.
I was veiledly suggesting that the expression "local feedback" is nonsense. The topological terms "feedback, input, output" require the presence of active load.
The P-N junction diode makes up (almost) all transistors. It is downright reprehensible to consider the single transistor bias arrangement and interelectrode capacitances as feedback. What is, in that view, not feedback?
The base-emitter capacitance is not feedback, nor is the base-collector (Miller / Early) capacitance.
IanHegglun,
No feedback is occurring.
Any resistor in any position reduces amplitude (voltage) and the amount of excited charges (current) by presenting an opposing force, thereby is degenerative. You would not go as far as to claim that resistors generally introduce feedback, would you? It would hardly mitigate the prevalent conceptual chaos increasing with every new post.
It can - but it is not.
By the way, proportionality (linearity) is a mathematical relationship. Maybe for you, like for all engineers and most scientists, physics and mathematics are the same thing. No discrepancies arising, but nothing being understood or explained either.
I refute that emitter degeneration implies feedback. The base and collector resistor are degenerating bias elements as well.
No. A connection to the output of subsequent stage must exist.
No. Feedback is established by R5, it is a prerequisite of CFP and CFA, the emitter resistor is not. The emitter resistor is a way to determine a desired stable operating point and to influence transitions properties, but need not to be there for the definition of common-emitter amplifier, CFP or CFA.
I was veiledly suggesting that the expression "local feedback" is nonsense. The topological terms "feedback, input, output" require the presence of active load.
The P-N junction diode makes up (almost) all transistors. It is downright reprehensible to consider the single transistor bias arrangement and interelectrode capacitances as feedback. What is, in that view, not feedback?
The base-emitter capacitance is not feedback, nor is the base-collector (Miller / Early) capacitance.
It is true that the word feedback implies a loop. There is no loop on a straight line. The circuit matrix of a degenerated transistor has no loop. The loop equation is valid for a straight line, but it would be pointless, like a 3D rendering of a photograph. The instant you try to do this you notice half of the process is simply redundant.
You could call degeneration feedback, the equations apply, but we choose to make a categorical distinction here because using feedback theory to explain degeneration is devolutionary and harmful to our purposes (namely efficiency).
But the distinction is ONLY categorical. When thinking, the thing is more important than the category. When communicating, the category is equally important but not more so. But if thinking and communicating are to be done together, the category should never be more important than the thing.
You could call degeneration feedback, the equations apply, but we choose to make a categorical distinction here because using feedback theory to explain degeneration is devolutionary and harmful to our purposes (namely efficiency).
But the distinction is ONLY categorical. When thinking, the thing is more important than the category. When communicating, the category is equally important but not more so. But if thinking and communicating are to be done together, the category should never be more important than the thing.
Good Morning
Yes I know, but in 2018 there should be no excuse for designing amps that are stable by pure luck and insist that it's not necessary to use compensation.
If you look at post #1413 Stradivarius answers "Last but not least, there is no compensation C. It is not a classic VFA with a Miller cap, the topology is based om a CFA." which in my opinion is pure BS.
He is not mentioning anything about slow devices that he can rely on to make a "stable" design, he is only talking about the topology (H-Bridge) and that a H-bridge VFA dont need compensation C. See post # 1392 for the circuit.
I dont care for my own part, but there might be others that follow this thread that try to learn something and are getting bad design advices and a lot of handwaving.
It was unexpected, and quite fun, for me to read that a degenerated transistor is not a feedback circuit. However, I think this hypothesis is worth to be exposed.
Take the most simple circuit using a transistor in Common Emitter operation.
The current across it will decrease if you add an emitter degenerative resistor.
The feedback loop is the whole circuit.
Last edited: