Is the CFB topology superior, and why?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I'm glad that the topic attracted you and you're beginning to understand, nevertheles we could still have eternal theoretical discussion about VFB vs. CFB.

My suggestion would be more of a practical nature: try to build simple-basic 6 transistors VFB and simple-basic 6 transistors CFB amp and compare them sonically.

Lazy Cat,
Sorry I'm in it for the eternal theoretical discussion (I can barely tell the difference between speakers).

The supposed theme, what are the true differences and what advantages or disadvantages do they bring.
I can accept the cfb current on demand aspect resulting in much higher slew rates, as well as eliminating the current source pole which is what I think Catalin is referring to.
As for the remaining parts of a typical amp structure I still fail to see why they can't be made similar. A vfb can have a common base stage as well as current mirrors driving the voltage stage.
Typically it does seem that cfb op amps inherently use their input structure combined with the natural current mirrowing to minimize additional voltage gain phase shifts at the expense of lower frequency gain.

So does it comes down to typical implementations of cfb versus vfb and the comparison of higher open loop bw at the expense of lower frequency loop gain?

Thanks
-Antonio
 
Elvee, you want to bring a deviation to the true. You forget that the vfb is uncompensated.Therefore the group delay is the same.But an uncompensated vfb will not work ever.
So please compensate the vfb and I will bring also a cfb with good phase reserve.Let see than the real group delay.

Let 's do the things correctly and not loose time.

Thank you
In this case, there is no need to bring compensation, because the source CFB circuit is supposed to be stable (it is not the case, but that is nor my fault nor my problem) and the conversion into VFB changed practically nothing to the loop stability parameters.
I could always overcompensate the circuit, but then the comparison wouldn't be fair
 
In this case, there is no need to bring compensation, because the source CFB circuit is supposed to be stable (it is not the case, but that is nor my fault nor my problem) and the conversion into VFB changed practically nothing to the loop stability parameters.
I could always overcompensate the circuit, but then the comparison wouldn't be fair

Well exactly this is the main difference between the VFB and CFB .It will not work without compensation .And when the miller cap appear at vfb (discrete or parasitic) then the pole at low f appear .
 
Well exactly this is the main difference between the VFB and CFB .It will not work without compensation .And when the miller cap appear at vfb (discrete or parasitic) then the pole at low f appear .
The CFB you gave as an example will not work without compensation (it might work practically, but not in a stable and reliable manner).
The derived VFB version is almost identical in this respect: in theory, it should need a slightly heavier compensation, but in practice, the difference is smaller than normal tolerance on components.
This simply shows that the conversion from CFB to VFB is almost costless: it is true that a very high frequency pole is added by the follower, but it is minuscule compared to the rest of a multitransistor amplifier.

The difference could be felt for a single stage common emitter amplifier: then , the added phase shift wouldn't be negligible anymore, but in a multistage amplifier, the supplement in phase is very small, almost negligible.
 
Last edited:
Elvee I didn't said that the cfb from picture is working. The cfb circuit was only to see and understand how cfb works .

If you want a real working cfb amp there are a lot also in this forum also in the most expensive amps like accuphase ,goldmund,etc .The accuphase e 213 for example has a plot for the function transfer like no one else .Google it
 
Catalin, Elvee

I finally figured out (I think, yes it takes me a while) what you guys are talking about, anyway I'd have to agree with both of you.
I do agree the cfb will have slightly less phase shift at the higher frequencies and that in general this will hold for a cfb.

Thanks
-Antonio
 
But an uncompensated vfb will not work ever.

That s not rigorously true.

I used an uncompensated symmetrical differential amp for many
years and it was perfectly stable , used for almost 30 years
and still working perfectly to this day.

I published the schematic somewhere at DIY Audio as a comparison
to check the effect of increasing high frequency open loop gain in respect
of resulting low order intermodulation products.
 
The original three transistor circuit that the cfb configuration is based upon needed no compensation capacitor either, its just that when integration became the norm the pnp transistors in these were inferior types that needed compensation otherwise they would oscillate.

This is the point about the cfb scheme it is just a way of obtain a constant bandwidth with a fixed compensation capacitor, and it is nothing more than this, although some seem to raise it to the status of a holy grail, the last word in audio excellence, etc

You can't really hear it you know, and no you can't trust your ears, as every scientifically valid test has always shown.
rcw
 
www.hifisonix.com
Joined 2003
Paid Member
Having now built both, I think they have their advantages.

I like the simplicity of CFA topology (see my sx-Amp, or JLH's 10 watter for example) and the fact that you can easily get decent slew rates and wide bandwidth. The open loop gain in CFA's generally tend to be lower, so you dont have as much NFB to correct for cross over distortion, so they are as a result, a natural fit with class A output stages. Some designers believe PIMD is an important issue - well the cure for this is CFA IMO.

I will still continue to build VFA amps, because I think if they are well designed they sound superb, but I would never say one topology is better than the other: its all about the execution at the end of the day.
 
do you really mean 'sound stage' and 'mono' ? - please expand a little (I am a recent builder of a mono system)
I wasn't talking about stereo separation.
Yes, I meant soundstage, which is not impossible in monophonic.

If the equipment is good enough, one should be able to get long distance cohesion and high resolution to fill the entire house with very big, very clear audio from a (one) speaker. Lesser quality equipment will need more speakers for covering a large area.
 
The point I am trying to make is that with a single channel the perceived sound stage is due to the amount of reverberation in the sound. rcw
Assuming that reverberation and ambient sounds exist in the source recording, then are we not discussing the resolution of low amplitude signal that some amplifiers allow to pass and others do not?

Passing these signals, reasonably intact, is apparently not impossible for the VFB. But, why, with voltage feedback, is there a bigger workout with the compensations before such performance is realized?
 
In a word no.

If the various distortions and frequency response with a given signal are bellow the hearing threshold then the amplifier will have no audible effect upon the signal.
Those who claim a particular topology or local only rather than global feedback is superior, are never able to substantiate this in properly conducted double blind tests, and invariably go into a shoot the messenger mode and claim the test invalid.
rcw
 
Those who claim a particular... rather than... is superior, are never able to substantiate this in properly conducted double blind tests

Which one is superior: Italian chardonnay or New Zealand Chenin Blanc?

The first issue there, is that the taster must be able to differentiate the taste between the two. Then he should have an idea why one is superior than the other, and that could be just about preference.

Which one is your preference: Sumatra durian or Malaysia durian?
Which one is your preference: Expresso or Luwak coffee?
Worse/best, which one is your preference: EVIAN or PERRIER?

Even IF you can taste the difference, you need (a life) time to judge why one is superior than the other. As for mineral water, it is a unique test. The different ingredient does affect the taste, but they are just mineral water without taste, aren't they?

Try this: bring 2 glasses of different mineral water. Now try and tell yourself, can you succeeded in the double blind test? Find brands such that your answer is NO (well, of course your answer would be no).

Now, change the water consumption in your house from brand A to B, and notice how much you spend/drink for each brand in each one month or two. Repeat the test to make sure that there is no coincident.

My prediction with the above is: even tho you cannot differentiate the taste in a double blind test, "your body" might be able to prefer one of them (BTW, I have succeeded in double blind test of 2 local brands)

Human body is so complex that no human can fully understand them, or how they work.

Same with audio. Even tho I found that I'm good at double blind test, I will not use it to judge if 2 things are equal or one is better than the other, because I know for sure that even if I failed in the blind test, the 2 things may have different effect against me.

So I prefer to use longer time to select my preference of an audio system. That's why I (might) have the same motto as you: "DON'T TRUST YOUR EARS" (especially if you're just a "newbie")
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.