Pros and Cons of current feedback amplifier.

[Osvaldo de Banfield]

I see the circuit in an old printed manual, in this page it is say to be "low cost" because of lower component bill, but I never try this wiring. As I say in other thread, it is anti-theory to drive a loudspeaker in high impedance.

[forr]

Originally Posted by hahfran

That entirely misses the point

So do Gilbert, Wurcer, Self, Horowtiz and Hill and many other audio designers who all certainly agree with the Wikipedia extract ?

[JMFahey]

Originally Posted by Waly

The schematic, page 2 in the ST datasheet.

You must have a different datasheet, because the one I checked http://www.datasheetcatalog.org/data...cs/mXutuqt.pdf
says nothing of the kind in page 2.
The actual "official" application circuit (which by the way is the one followed by 99.999% of users) is the one on page 6, labelled as "Figure 16. Typical application circuit" and it clearly shows

The typical low value resistor network shown provides Voltage fedback.

composed of a 220 ohms resistor and a 2.2 ohms one.

Of course you may use it *without* NFB , a.k.a. "open loop" for some unfathomable reason which is up to you.

That's not what Bruno Murari, the original TDA2002 designer wanted, by the way, and he went to great lengths to avoid it, that's why the low impedance *voltage* feedback network was implemented.

[abraxalito]

Originally Posted by JMFahey

You must have a different datasheet, because the one I checked http://www.datasheetcatalog.org/data...cs/mXutuqt.pdf
says nothing of the kind in page 2.

Perhaps the amp's operation mode is opaque to you? If you care to examine the schematic, the negative feedback is provided to Q4's emitter. That's by nature a low impedance node, hence the feedback is by means of current. For a voltage feedback amplifier, the -ve input is high impedance, normally the base of an LTP.

[jan.didden]

There might be a semantic issue here. Traditionally, the name 'current feedback' has been given to the kind of feedback that returns a sample of the output current. This could be for instance the voltage across a small resistor in series with the load. Current feedback thus makes the output impedance very high.

With the new so-called 'CFA' opamps that appears to have taken on a new meaning (blame it on the marketeers) namely: feedback to a low impedance node. Personally I think this is an ambiguous way to name it because, what's a low impedance node? The emitter impedance of a starved BJT can be pretty high, while 'voltage feedback' to an invering input fed from a low-impedance source may be pretty low impedance.

But whatever you fancy, you should at least be clear which definition you want to use in this particular discussion.

As to pros and cons: like always in engineering, you can make very good or very bad amps with either technology.....

jan

[wg_ski]

A lot of earlier (and nice-sounding) amps used second collector to first emitter feedback. Output followers of course don't count. They didn't need a well-matched and balanced diff pair to see all the benefit of the feedback applied. I'd imagine that the thermal feedback on a TDA20x0 is pretty severe and relying on a diff pair staying balanced was pretty iffy when that device was designed. Doesn't surprise me the choice of input toploogy.

[Waly]

Originally Posted by jan.didden

There might be a semantic issue here. Traditionally, the name 'current feedback' has been given to the kind of feedback that returns a sample of the output current. This could be for instance the voltage across a small resistor in series with the load. Current feedback thus makes the output impedance very high.

With the new so-called 'CFA' opamps that appears to have taken on a new meaning (blame it on the marketeers) namely: feedback to a low impedance node. Personally I think this is an ambiguous way to name it because, what's a low impedance node? The emitter impedance of a starved BJT can be pretty high, while 'voltage feedback' to an invering input fed from a low-impedance source may be pretty low impedance.

No, nothing to do with marketing. It's an illustration of the four types of feedback circuits. Each of these types has it's own formalism, in either h, g, y or m parameters. These are chosen "naturally", based on the impedances, "low" means current, "high" means voltage (at the input) and the other way around for the output. Actually, these formalisms are not bringing anything special, you can analytically evaluate any of the four topologies using e.g. only voltages and impedances (but, you or the computer, be prepared to handle some stupid numbers).

Your simulator doesn't give a rat's *** if the curcuit is voltage or current feedback, it just determines the transfer function you need (U/U, I/U, U/I, I/I).

There are only circuits with inputs and outputs. Some of them can be conveniently analyzed using the feedback concept. For humans, when analyzing (or designing) a new circuit, it is easier to go through an euristic phase and perhaps identify if the circuit functionlity could be understood using some pre-fabricated methodology (like feedback theory).

[jan.didden]

Originally Posted by Waly

These are chosen "naturally", based on the impedances, "low" means current, "high" means voltage (at the input) and the other way around for the output. .

My point was that the 'traditional' terms were named after what sample you returned, iow voltage feedback meant you took (a sample of) the output voltage and returned that as feedback, while if you sampled the output current you would call that current feedback. It had nothing to do with the impedance of the node to where you send the feedback.

Your interpretation is the newer interpretation that came into swing after the 'CFA' came onto the market.

Neither is intrinsically right or wrong - you agree to the definition and that's it. It's just that there are two types of definitions afoot and I just wanted to flag that.

jan

[ontoaba]

Current feedback uses different way compared to VFB. The output comes from current measurement at the input devices, so there is no market name there. Also current feedback basically work in half (+ or - side only).

[Ken Newton]

Hi, Guys:

The distinction between voltage feedback (VFB), and current feedback (CFB) seems interesting to ponder. We could say that a distinction is that CFB samples the output current back to the input, while VFB does not. Except that, VFB using a long-tailed input pair constructed of bipolar transistors also samples the output current - since bipolar transistors operate on current conducted through their base electrodes.

Alternately, we could argue that VFB samples the output voltage, while CFB does not. Except that, one can measure a small, but present, signal voltage presented to the non-zero ohm feedback input electrode. The impedances of the respective feedback networks of VFB and CFB reflect the relative impedances of their respective feedback input electrodes. So, I don't feel that distinction get's us at the root of the distinction either.

To me, perhaps, the most relevant operational distinction is that in VFB, the feedback input electrode typically drives the base/gate of a grounded emitter/source stage. In CFB, the feedback input electrode typically drives the emitter/source of a grounded base/gate stage. So, with CFB, we get all of the expected benefits of grounded base/gate operation, such as wide bandwidth and better open loop linearity. However, we also get the expected difficulties, such as low input impedance. We are all describing this core distinction, I think, but just from varying perspectives.