No, I am investigating the subject mainly from a discrete implementation perspective, where you are at liberty to tailor the compensations, currents, etc for the best perfomance tradeoff in a given situation.
I understand CFB does have an advantage in the >automatic compensation/one size suits all case<, typically a general purpose integrated video amplifier, but what I am interested in is: above and beyond that, does the CFB offer striking advantages over an equivalent (BOM, cost, or otherwise) VFB.
Also, is it always possible to classify one type of circuit into one or the other category, or is it always possible to emulate either of them, for example by adding parallel and series degeneration resistors to a normally VFB or CFB amplifier.
The old rule of thumb is that when you reached a frequency of ft/b, then the input and output time constants of the individual devices start to become significant and feedback over more than one stage starts to become a problem.
I have not simulated the circuit bellow but it works very well and is for a project that needs a buffer line driver in the 10-20MHz. region
The overall gain is enough just using the common base stage with local feedback and bootstrapping, thus avoiding the problems with common emitter amplifiers.
Since the input and output are emitter followers with current source loads you can look at it as having 100% voltage feedback in these two stages, and feedback across the common base stage base resistor.
rcw
I have not simulated the circuit bellow but it works very well and is for a project that needs a buffer line driver in the 10-20MHz. region
The overall gain is enough just using the common base stage with local feedback and bootstrapping, thus avoiding the problems with common emitter amplifiers.
Since the input and output are emitter followers with current source loads you can look at it as having 100% voltage feedback in these two stages, and feedback across the common base stage base resistor.
rcw
Attachments
For audio purposes a so called CFB amp not only yield no advantages
over the usal VFB but will more surely bring a lot of drawbacks compared
to the latter.
Some will argue for higher slew rates without explaining how this will
translate in sound quality since theses kind of amps are not more linear
than the more traditionnal VFBs wich seems to be better in respect of
global perfs.
over the usal VFB but will more surely bring a lot of drawbacks compared
to the latter.
Some will argue for higher slew rates without explaining how this will
translate in sound quality since theses kind of amps are not more linear
than the more traditionnal VFBs wich seems to be better in respect of
global perfs.
Yes this is so.
The fact is that the 072 and 5534 op amps are all you need for audio. and anybody who says they can hear the inferiority of these is speaking through an orifice lower and to the rear of their mouth.
Much of the symmetrical and local only feedback topologies were originally developed for instrumentation etc. where there is a need for it, and the only need for it in audio is esthetic and marketing hype, not practical.
The circuit I posted for instance uses a CA3083 transistor array and is currently buffering a 10MHz. crystal oscillator and driving a 50 Ohm line with 500mV. rms.
I could have used a current feedback op amp but I wanted to design a discrete circuit for old times sake, and had a CA3083 on hand.
rcw
The fact is that the 072 and 5534 op amps are all you need for audio. and anybody who says they can hear the inferiority of these is speaking through an orifice lower and to the rear of their mouth.
Much of the symmetrical and local only feedback topologies were originally developed for instrumentation etc. where there is a need for it, and the only need for it in audio is esthetic and marketing hype, not practical.
The circuit I posted for instance uses a CA3083 transistor array and is currently buffering a 10MHz. crystal oscillator and driving a 50 Ohm line with 500mV. rms.
I could have used a current feedback op amp but I wanted to design a discrete circuit for old times sake, and had a CA3083 on hand.
rcw
This confirms my own impressions: an integrated CFB amplifier is a useful general purpose building block, but if you brew your own circuit, you are completely free to tailor the compensations as you wish, and to add internal degenerations if necessary to attain the best possible performance with a VFB circuit.
Interesting opinion. And it would be more interesting to know what amplifier that you can suggest for the top class amplifier that can be built as DIY amplifier. I have no idea what would it be. A Lynx amplifier? Musical Fidelity clone? MBL clone? Audio Analogue Puccini? The only idea that I have is that you have none. But it can't be that simple. The answer must be interesting I suppose.
The first statement is incorrect, VFB has as many advantages as it does have disadvantages over CFB. Common Wahab you know that.
How does slower slewrate argue for sound quality ??? Until Edmond showed TMC compensation it was impossible for a VFB amp to come close to THD20 performance that can be obtained by CFB. I havent looked but it might just be possible to use TMC in CFB amp. Better THD was obtained by CFB amps and this has been the case ever since CFB made its appearence, only recently performance is much closer.
Such a claim can be made only by a person that instead of brain, for thinking and listening purposes, uses the stuff that is regularly found inside the orifice that you mention.
The difference is easily heard, even by untrained ear. Two conditions are required to accomplish this:
1. Person must have undamaged hearing
2. Person must listen for the difference on, at least, half-decent system that can be bought in an average hi-fi store for 1k$ (source, amp, speakers). The higher the quality of the reproduction chain, the difference becomes more pronounced.
The only thing I can say Juma is that post a documented and verified account of your assertions. otherwise you are making noises out of the orifice that you purport to be your mouth, but is in fact located in another position.
As to hearing that is undamaged, my own was subjected to the usual 60's ansd 70's rock n' roll , plus playing a bass guitar, and also a small stint in a boiler making shop.
If you can actually hear the difference between a 741 op amp and your pet device with a statistically significant figure of more than fifty percent then you might convince me and the bulk of other scientists on the planet, until then your assertions are entertaining but irrelevant.
rce
As to hearing that is undamaged, my own was subjected to the usual 60's ansd 70's rock n' roll , plus playing a bass guitar, and also a small stint in a boiler making shop.
If you can actually hear the difference between a 741 op amp and your pet device with a statistically significant figure of more than fifty percent then you might convince me and the bulk of other scientists on the planet, until then your assertions are entertaining but irrelevant.
rce
It is in principle always possible to convert any working CFB amplifier configuration into a VFB one having at least the performances of the CFB original (that's in theory at least).
Doing the opposite might also be possible, but it looks more difficult.
Hmmm some notable guys in Natsemi seem to had been listening to their audio chips for evaluating far more subtle variations between their advanced designs. Older relative post.
The first statement is incorrect, VFB has as many advantages as it does have disadvantages over CFB. Common Wahab you know that.
VFB has more advantages than disadvantages over CFB.
I beg to differ in this matter.
CFBs logicaly gave gave better perfs than the old first kind of VFB
amps because it had two gain stage instead of a single inverting
gain stage.
Latter topologies using input differential did equal and even best
the classsical CFBs.
The slew rate limitation is often due to poor design but in principle,
slew rates of hundreds V/uS are quite feasible with VFB topology,
although it is somewhat pointeless in audio gears.
.[/QUOTE]
The point is that if we take the slew rate we need as being where, f = 2pifVp, then if we want to design an amplifier to put out a 10V peak signal at 100kHz. we need a slew rate of 6V/us.
Exactly why one would need more than this is not clear to me, and the vast majority of others, and the 5534 has a slew rate that is more than twice this.
To adequately amplify the 10MHz. sine wave I am with the posted circuit you need at least 45V/us, now current feedback devices start to make sense, other wise it is technical nonsense to say that you need this sort of figure to amplify audio signals, however vehement your assertions to the contrary might be.
rcw
Exactly why one would need more than this is not clear to me, and the vast majority of others, and the 5534 has a slew rate that is more than twice this.
To adequately amplify the 10MHz. sine wave I am with the posted circuit you need at least 45V/us, now current feedback devices start to make sense, other wise it is technical nonsense to say that you need this sort of figure to amplify audio signals, however vehement your assertions to the contrary might be.
rcw
I think both topologies have their merits. However one advantage (and plenty will disagree with this) is that in CFB th loop gain bandwidth is generally greater than the audio bandwidth, so you can apply the same amount of feedback across the whole audio band. If you believe PIMD is an issue, you can either design you VFB for very high loop gain (this reduces PIMD - see Cordell for example) or you can go the CFB route.
I will publish a little amplifier in a week or so on my site - CFB. Slew rate 250V/us, 700kHz -3dB and loop gain -3dB of circa 60kHz. Figures like his are not possible with VFB without a lot of effort.
I think slew rate is important, but beyond a certain point, does it add any value.
Which one is best? Both of course!
I will publish a little amplifier in a week or so on my site - CFB. Slew rate 250V/us, 700kHz -3dB and loop gain -3dB of circa 60kHz. Figures like his are not possible with VFB without a lot of effort.
I think slew rate is important, but beyond a certain point, does it add any value.
Which one is best? Both of course!
The CFP is an interesting circuit. It can give much more current gain than a Darlington pair if you are doing an impedance transformation. It is prone to UHF oscillation and it is a good idea to use a stopper resistor at the base of the PNP transistor. It is also a bit noisy. A very good basic 2 transistor circuit is the bootstrapped amplifier. There are quite a few tricks you can pull with that circuit.
In its basic form it is rather unusual because it is an example of positive feedback actually increasing linearity. You can connect an emitter resistor to the first transistor without actually reducing the gain very much, improving linearity even more. You can also use negative feedback around the outside of the amplifier giving even greater linearity.
The use of both positive and negative feedback in an amplifier can be both viable and very useful. If I had time I would provide a clear circuit example but maybe you can understand from some of the examples you can download from here: EvoSpice 4.1 & 4.2
In its basic form it is rather unusual because it is an example of positive feedback actually increasing linearity. You can connect an emitter resistor to the first transistor without actually reducing the gain very much, improving linearity even more. You can also use negative feedback around the outside of the amplifier giving even greater linearity.
The use of both positive and negative feedback in an amplifier can be both viable and very useful. If I had time I would provide a clear circuit example but maybe you can understand from some of the examples you can download from here: EvoSpice 4.1 & 4.2
[/QUOTE]VFB has more advantages than disadvantages over CFB.
I beg to differ in this matter.
CFBs logicaly gave gave better perfs than the old first kind of VFB
amps because it had two gain stage instead of a single inverting
gain stage.
Latter topologies using input differential did equal and even best
the classsical CFBs.
The slew rate limitation is often due to poor design but in principle,
slew rates of hundreds V/uS are quite feasible with VFB topology,
although it is somewhat pointeless in audio gears.
.
Name the advantages and disadvantages of VFB, Ill add the disadvantages youre not thinking of.
CFB outperformed all generations of VFB, even now the latest outperform the VFB if we look at opamps. Please name me one VFB opamp that could equal even the old 1980s comlinear opamps. Two gain stages ???, could you demonstrate.
Traditional VFB topology cant and never will reach the speed that can be attained by CFB, theory dictates. VFB can have larger slewrate at the cost of very high currents in the LTP, CFB on the other hand can probably reach 10 times that high if the same current is used.
Again the 'classical' mistake of assuming that the to-be-reproduced-audio-signal is all that the amplifier is working with. No! Any active feedback system needs to and runs at much higher frequencies than required for the signal-to-be-created-by-feed-back.
This means: The better the slew-rate, the faster the loop control frequency is. The faster loop control frequency, the better harmonics and other interference suppression.
In theory, any CFB amplifier can be converted into a VFB one having at least identical performances.
The only thing I can say Magic Box is that making bold assertions of that sort withought any technical justification for them is really not good enough.
Exactly where is my "classical mistake" and what references do you have to back up your assertions, there are non that I can find. This is the sort of thing you sometimes hear being asserted, withought any shred of evidence at all, in the more lunatic fringes of audio.
rcw
Exactly where is my "classical mistake" and what references do you have to back up your assertions, there are non that I can find. This is the sort of thing you sometimes hear being asserted, withought any shred of evidence at all, in the more lunatic fringes of audio.
rcw
Of course there's nothing wrong with the basics. What happened usually is that real world situation is always more complex (than implementing one or two basic formula).
Your bold assertion, explicitly or not, is that the scientists, or the vast majority of the more clever scientists, are having exactly the same thinking as you...
Please show us your high level of technical knowledge to justify that...
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.