Jan
Of course that any Amp has his own finite transient response limit ! , personally for me the faster - the better . ,
but my point was to tell you that one very fast Amp unit have No Chance to perform well without good designed PSU , which is capable to deliver DC stored energy in to Amp circuit very fast , almost without of any voltage sag .
And yes , I was not speaking about improving Amp transient response with more power supply caps , but about dividing one optimum big PSU capacitor value with few smaller values PSU capacitors , which capacitance sum is the same as was to the big one Elko , again to optimum capacitance value .
And one question for you :
What is more important for you , how one Amp reproduce some signal or how is sounds ?
Last edited:
Private joke. I had opened an answer to the previous input:
And my response was exactly: " ???".Originally Posted by magnoman
I thought it has already been argued and accepted that a cfa does not enter class ab on it's input stage and therefore has no slew advantage.
I hesitated and never clicked on "Submit Reply" for personal reasons: i was accused to had published a CFA in witch the input stage was accused to work in "class B and even C".
Last edited:
I understand what you mean, but this is not the case. The voltage sag has (almost) no impact on the amp transient response, as long as the amp doesn't clip.
I want it to reproduce the best as possible the signal at the input. If that doesn't sound good, it is bad music, or bad speakers, or I am not in the right mood today, or I just have bad taste
jan
If I have been following the discussion correctly the cfa topology is more sensitive to power supply noise and any ripple? Wouldn't this call for a better power supply due to this situation? What I am saying is wouldn't you want a better layout to the power supply to minimize any dc offset and other noise, perhaps better ground topology and all the other tricks that are pointed out by Gootee in his power supply thread?
I want it to reproduce the best as possible the signal at the input. If that doesn't sound good, it is bad music, or bad speakers, or I am not in the right mood today, or I just have bad taste
jan
Jan
I expected some simple answer, something as :
`` Both , good signal reproduction and good Amp sound`` !
Any way , I admit that my question was pretty absurd !
Regards !
Last edited:
Yes. You both need the best power supply and care, in your amplifier design and board layout to filter further any ripple and get lowest impedance as possible, specially for the input stage.
It seems obvious that, if your amp has "good signal reproduction", he will have a "no amp sound" characteristic.
Last edited:
then drop the "fanboy" paranoia, accusations of "attacks", by "naysayers", etc. from this thread too - to me such language indicates failures of logic, lack of interest in knowledge seeking
pointing out special conditions behind some claims of CFA "advantage" isn't an "attack" - without knowing such conditions no fair comparison can be made
and every circuit technique "competes" in the current technology with alternative possibilities - what designers need are valid comparisons, broad knowledge of strengths, weaknesses and interactions at multiple levels up to complete application circuit context
another unfortunate result of any "lets not be critical" plea is easily seen in the last few pages - "the usual" audiophoolish memes crawl forth and strut shamelessly throughout such threads
jcx,
I definitely don't know enough to be a fanboy, never have been that with anything anyway. Before I can do any comparison I need to understand the basic circuit difference besides the oft said current feedback vs voltage feedback. I'm just going to assume for now that Richard or whoever will point out some of the problems and shortcomings that have to be overcome, I don't think that anyone is going to say that this or any circuit is perfect, would be nice but I kind of doubt there is one.
I definitely don't know enough to be a fanboy, never have been that with anything anyway. Before I can do any comparison I need to understand the basic circuit difference besides the oft said current feedback vs voltage feedback. I'm just going to assume for now that Richard or whoever will point out some of the problems and shortcomings that have to be overcome, I don't think that anyone is going to say that this or any circuit is perfect, would be nice but I kind of doubt there is one.
Ok, no food recipe, but not always better in technical view is better sound. There is some other technical results that decreasing caused by increasing of other technical results.
For example faster configuration also gives lower gain. And in this case, bjt has its own sound characters that could be minimized with enhancement.
This view is wrong too, better technical specs result in better reproduction of incoming signal. The definition of amplifier is a piece of equipment that amplifies a input signal without distorting, adding or losing any information of that signal. It is not to amplify a signal in such a way that the resulting deviations from the input signal make it desirable or nice. If deviation is what you want get a graphic equalizer or a sound effect generator. You also forget the role that speakers and the source used to generate the signal for the amp has on the sound resulting from the system. If you have a wire with gain amp and dont like the sound rather blame the rest of your system than your near ideal amp.
A good example of this is accuphase amps. Ive heard these amps in many systems and in many they cause consternation because they are designed to be wire with gain amps and easily reveal dificiencies in of the rest of a system. In many cases the resulting sound is truly awefull but the deficiency is not from the amp. I mention these amps because they are CFA topology and represent the finest available commercial CFAs designs. There is also VFA amps that do so, a good example is the halcro range.
JCX, I apollogize to not answering to some of your posts, I mean no disrespect, I just have little time to do so and when I can follow the forum sometimes I lose track of the posts but Ill try.
I notice you are impartial and look at the issues from a obtaining ideal technical results rather than audiophool perspectives which is commendable.
I notice you are impartial and look at the issues from a obtaining ideal technical results rather than audiophool perspectives which is commendable.
Lets try to get back on track here.
So, we've established that the CFA input stage, if correctly biased up, does not transition from class A to class AB or B ( see Richards post #57).
Can we contrast a VFA and a CFA's behavior where the input is driven with an extremely steep rising edge without a front end band limiting filter?
We would expect the VFA front end to overload, and the result would be an error voltage (Vni- Vinv) in excess of the LTP linear input range I.e the amp would be slewing, with one half of the LTP OFF - analogous to class B operation.
A CFA input stage on the other hand remains in the class A region. But, what limits the rise time and ultimately the bandwidth? Does this in effect actually happen in the TIS because of the finite current available to charge the compensation capacitor, rather than the input stage as is the case with the VFA? What will happen if we replace the conventional current mirror in the CFA with a beta enhanced amplifier stage? Will this change the CFA amplifiers behavior fundamentally?
Some questions to explore . . .
So, we've established that the CFA input stage, if correctly biased up, does not transition from class A to class AB or B ( see Richards post #57).
Can we contrast a VFA and a CFA's behavior where the input is driven with an extremely steep rising edge without a front end band limiting filter?
We would expect the VFA front end to overload, and the result would be an error voltage (Vni- Vinv) in excess of the LTP linear input range I.e the amp would be slewing, with one half of the LTP OFF - analogous to class B operation.
A CFA input stage on the other hand remains in the class A region. But, what limits the rise time and ultimately the bandwidth? Does this in effect actually happen in the TIS because of the finite current available to charge the compensation capacitor, rather than the input stage as is the case with the VFA? What will happen if we replace the conventional current mirror in the CFA with a beta enhanced amplifier stage? Will this change the CFA amplifiers behavior fundamentally?
Some questions to explore . . .
Last edited:
Lets put the power supply sensitivity to bed for the time being. You might find it interesting to the extent that cascoding will improve CMR as well as PSR. If you have a very low distortion design (under .001% thd), then you dont need to use cascoding to lower the distortion . But you can use it to increase the CM (Common-Mode) Rejection aas well as the PS (Power Supply) Rejection.
That is what I tried after the 1980 design which ended up looking like this a.t.t. If you try to SIM it and you do not get the 'measured' results i got, then see what you need to correct in the SIM to make it match reality. [I think the outputs were 2N2905A and 2N2219A.]
Thx-RNmarsh
That is what I tried after the 1980 design which ended up looking like this a.t.t. If you try to SIM it and you do not get the 'measured' results i got, then see what you need to correct in the SIM to make it match reality. [I think the outputs were 2N2905A and 2N2219A.]
Thx-RNmarsh
Hi Bonsai,
This assessement (slewing of VAS limited by VAS input current to charge the miller) I never grasped because it's so contrary with my findings. Which are: The higher the VAS standingcurrent with a given miller cap, the better the VAS slewrate, while the available input current from the LTP never changed. This is in contrast to that assertion.
I think many view the miller comp as a "bypass" from the VAS input directly to the VAS output. While a fact is that the VAS output is an inverting node to the VAS input. It acts as a "counter bootstrap" to the input signal. A "brake" on the inputsignal when Cmiller's Z becomes low enough. Due to the inverting operation of the TIS transistor I feel the charge/discharge primarily comes from the VAS output, rather than from the VAS input.
Hello Magicbox, the LTP current in a conventional MC VAF is absolutely sets the slew rate. But, don't take my word got it - Self, Cordell and Prof Leach discuss this in depth in their texts with associated math. You can break this relationship and speed things up considerably of course by going to MIC - Bob Cordell discusses this in his book as well.
Odd, while I take their word for it, I wonder how the VAS output standing current comes into play with this. Mind you, I'm not talking about MIC.
Maybe it helps to reliealise that any current 'into' the Miller comp cap also has to 'come out' of it.
So on the Vas output side there is generally enough current to 'come out' of the miller cap, but that current also has to 'go into it' at the base side, and the only place it can come from is the diff input stage output collector. So, ultimately, it is this LTP standing current that limits the charge current and thus the slew rate.
jan
PS Of course as this is ac, 'come out of' and 'go into' are interchangeable.
So on the Vas output side there is generally enough current to 'come out' of the miller cap, but that current also has to 'go into it' at the base side, and the only place it can come from is the diff input stage output collector. So, ultimately, it is this LTP standing current that limits the charge current and thus the slew rate.
jan
PS Of course as this is ac, 'come out of' and 'go into' are interchangeable.
Jan
I expected some simple answer, something as :
`` Both , good signal reproduction and good Amp sound`` !
Any way , I admit that my question was pretty absurd !
Regards !
Asking questions with predefined answers that really are not opposites of each other is a well-known tactic that politicians are good at. And we all know what THAT leads to!
Its a current, it just flows between two points. As the parasitic capacitances, i look at it as a charge in the input, making a low pass filter with the previous stage output impedance.
Last edited: