In first thread there are comparation between VFA and CFA. Usually VFA optimized by THD. And we knew that CFA usually better in slew rate and bandwidth.
Then almost everyone propose CFA that optimized by THD and ended with complex amplifier.
And now we propose simple CFA like VSSA that optimized by THD (and ended with the complex one? 🙄).
I wonder why do we design simple CFA that optimized by slew rate or trade the THD for slew rate? I think we must maximize the superiority of CFA. Maybe 0,01% THD is enough for me.
Well, in my mind, it was *inside loops* and for signal path only.
This can explain the success of some very simple builds of Nelson Pass, Hirraga etc... But require a lot of sofisticated work around, PSU, cabling, components etc...
In my studio, we made extensive blind listening comparisons of analog ICs VS strait wires in order to improve a big Calrec mixing desk.
Where able to find some totally transparent sets. Four units in serial each time: +-1 X 0dB, 10dB gain. It makes a lot of transistors...
It happens they were all...Current feedback ones 🙂
I tried to show that everything equal, CFA don't add distortion by itself... on the contrary...
Agree on HD levels: i had during years an amp with 0.025% witch was my favorite.
Last edited:
Yes, and the point has been made very well. The CFA is very competitive for audio use.
I'm Ok with any CFA that has high SR (Slew Rate) to compare to more typical SR amps. Thus, my poking for making one's design have higher SR. Until I showed the one from 1993 Elantec and it was SIM'ed here at over 800v/usec 🙂 and very low HD... it is possible. But, might not have to go that high SR but a significant amount above what we usually listen to would be more conclusive. However, the THD can be kept much lower than .01-.025% threshold IMO for this development - ten times better or below .001% if possible (to be Blameless).
Thx-RNMarsh
Last edited:
Here we are, Richard 🙂
I believe that most of the tries exposed in this thread fulfill your requirements.
Now, what can we say ? At the end of my life, i had not the slightest idea of what can be the requested target for Slew rate or Distortion. How-can i explain i had tuned a low pass filter at the input of my ultrafast CFA personal amp around 80 or 100 KHz, if i remember well ?
There is so many things in action we don't know or measure.
The only judge i believe in is "my ears". Don't you ?
So, once we have designed an amp the best we can, and build-it, it tells everything, and only at this time we can tune-it to get the best of it, or give-it up, because it is crap, whatever the numbers (yes often experienced ;-)...
It would be interesting to make a serious study about all this: distortion slew rate, IM etc... But it should take years of full time work to be able to lighten general laws...
Of course, i would be very interested to know what is the HD number, or slew-rate witch do not need to be improved... while i believe it is asymptotic, or depending of other phenomenas.
I believe that most of the tries exposed in this thread fulfill your requirements.
Now, what can we say ? At the end of my life, i had not the slightest idea of what can be the requested target for Slew rate or Distortion. How-can i explain i had tuned a low pass filter at the input of my ultrafast CFA personal amp around 80 or 100 KHz, if i remember well ?
There is so many things in action we don't know or measure.
The only judge i believe in is "my ears". Don't you ?
So, once we have designed an amp the best we can, and build-it, it tells everything, and only at this time we can tune-it to get the best of it, or give-it up, because it is crap, whatever the numbers (yes often experienced ;-)...
It would be interesting to make a serious study about all this: distortion slew rate, IM etc... But it should take years of full time work to be able to lighten general laws...
Of course, i would be very interested to know what is the HD number, or slew-rate witch do not need to be improved... while i believe it is asymptotic, or depending of other phenomenas.
800 volts per microsecond SR -
I understand -- we can hear difference between 1% THD and .1% and lower. I never read anywhere about the slew rate compared by listening... all else the same. It was speculated years ago that 1v/usec/volt output was all that would be needed. It was based on a limited set of conditions and assumptions. So we get to 'move the needle' forward - perhaps by next trying various higher SR (all else equal or blameless) and see if there is any improvements to be heard.
I hope this is all useful to those not yet retired and can make use of this info and dispelled some rumors and fantasies and misunderstandings regarding CMA. They are easy to build/design and to stabilize and can use few parts and are low thd at high freqs and quiet and etc etc etc. Time to hand the baton to someone else to do more and do the follow-thu. And, get it published for the books.
Meanwhile, there are many CMA designs from past-present-future to be designed and built. And, dont forget to Enjoy the Music !
Thx-RNMarsh
😎🙂
I understand -- we can hear difference between 1% THD and .1% and lower. I never read anywhere about the slew rate compared by listening... all else the same. It was speculated years ago that 1v/usec/volt output was all that would be needed. It was based on a limited set of conditions and assumptions. So we get to 'move the needle' forward - perhaps by next trying various higher SR (all else equal or blameless) and see if there is any improvements to be heard.
I hope this is all useful to those not yet retired and can make use of this info and dispelled some rumors and fantasies and misunderstandings regarding CMA. They are easy to build/design and to stabilize and can use few parts and are low thd at high freqs and quiet and etc etc etc. Time to hand the baton to someone else to do more and do the follow-thu. And, get it published for the books.
Meanwhile, there are many CMA designs from past-present-future to be designed and built. And, dont forget to Enjoy the Music !
Thx-RNMarsh
😎🙂
Last edited:
Thank you RN-Marsh. That is what I mean. I have difficulty to describe it in English.
My VSSA sound better than my old VFA amp that has better distortion figure. I want to know why can VSSA sound better? Slew rate? IMD distortion?
My VSSA sound better than my old VFA amp that has better distortion figure. I want to know why can VSSA sound better? Slew rate? IMD distortion?
Slew rate, i presume, the "expansive" behavior of CFAs on transients, simplicity, miniaturisation due to SMD parts and careful design/choice of components made by a designer who listen at each step of his work...
About distortion figure, i think nobody till now have done any real measurement on this little amp.
Last edited:
I measured my old amp. It has much better distortion figure than the result of simulating VSSA. Than i asume that my old amp better distortion figure. Right now i don't have distortion meter. My old amp has 300khz power bandwidth, and 96db S/N ratio un-weighted, 180W into 8 ohm. I forgot the slew rate. I measured it around ten years ago. I hope someone measure VSSA.
Nothing of the above , that s pure psychologicaly motivated belief
but it has no foundation based either on measurement or real
comparative listening tests.
Harmonic distorsion analysis on a large bandwidth is related to the idea of waveform preservation of the audio signal.
Maximal slew-rate, as far it is sufficient to handle the highest variations of the signal (almost always the case), is not.
If an amp is prefered by a person to another amp having less THD (both supposed to have good stability and even no overshoot on square waves) , I suspect that it is only its non-linearities which are appealing for the person and no any other objective virtue of the design.
I do not criticize the preference (I may even agree, what is important is listening pleasure) but I abhor the claim that the circuit of the first amp is better for non-pertinent objective reasons.
Maximal slew-rate, as far it is sufficient to handle the highest variations of the signal (almost always the case), is not.
If an amp is prefered by a person to another amp having less THD (both supposed to have good stability and even no overshoot on square waves) , I suspect that it is only its non-linearities which are appealing for the person and no any other objective virtue of the design.
I do not criticize the preference (I may even agree, what is important is listening pleasure) but I abhor the claim that the circuit of the first amp is better for non-pertinent objective reasons.
forr and Wahab, we have been down this road many times and it leads nowhere.
Let people enjoy their subjective views and leave it at that.
Let people enjoy their subjective views and leave it at that.
Bonsai,
I clearly leave and never contest your subjective views. As I said I could even share them, there is no problem with that. I am only discussing of what is relevant to the technical aspects of the designs and objective facts about them.
Look at Morgan Jones on tubes or, in this thread, Dick Marsh.
Not many repetitive claims about the superior sonic quality of what they design. They just design for what they think as the best for the kind of technology which inspires them. Even if this technology is not my prefered one, I like to read them because I learn from them.
There certainly has been interesting contributions in this thread but the insistence of some claims appears as proselytism and does not do as good for the CFA technology as their promoters seem to think.
The slew rate of the B.B. King Gibson 335 is less impressive than the one of the Mark Knofler's Telecaster ;-)
Christpohe, you have such a elegant way of summing things up for me, as I listen to my TPA3100D2 amplifier.
forr, the gap between peoples subjective views and the the technical explanation will never be closed. Never.
So just leave it . . .
So just leave it . . .
Just to get this in - for the records ---
Slew Rate is possibly only a 'marker' of something else. Something else that we can hear. But we dont yet know what that 'something' is actually. But, you can still characterize theound by SR - if it turns out to have a character.
Let me make an analogy -- if we design for low distortion at the highest audio frequencies, we end up with a wide bandwidth amplifier. Now, if we used BW as a 'marker', we might assume the distortion could be low at high freqs. So, SR could be an indirect measurement of something we can hear. Might be good to find out and do a measurement that is more directly related/correlated to something else - to what we hear but that has not been done yet.
So. Just as jumping on freq higher than we can hear is the wrong issue - SR may be an indirect affect also... thus its called a 'marker'.
That is why i keep pushing for higher SR - to see if something else falls out from it. The CMA are the easiest way to get high SR and high linearity.
Another example, cables.... many people measured the cable and found nothing there - BW, THD etc etc. Thus, no one can hear differences in cables. Yet, that is/was the wrong argument --- interfacing with Zout of amps and cable C of phono cartridge/preamp interfacing affect frequency response of the "system". And, it is quit audible in many situations.
I think it is better to err on the side that a lot of people over a long time - across the world- say they hear - even if we dont understand what to measure or even which marker to test for.
Thx-Richard Marsh
Slew Rate is possibly only a 'marker' of something else. Something else that we can hear. But we dont yet know what that 'something' is actually. But, you can still characterize theound by SR - if it turns out to have a character.
Let me make an analogy -- if we design for low distortion at the highest audio frequencies, we end up with a wide bandwidth amplifier. Now, if we used BW as a 'marker', we might assume the distortion could be low at high freqs. So, SR could be an indirect measurement of something we can hear. Might be good to find out and do a measurement that is more directly related/correlated to something else - to what we hear but that has not been done yet.
So. Just as jumping on freq higher than we can hear is the wrong issue - SR may be an indirect affect also... thus its called a 'marker'.
That is why i keep pushing for higher SR - to see if something else falls out from it. The CMA are the easiest way to get high SR and high linearity.
Another example, cables.... many people measured the cable and found nothing there - BW, THD etc etc. Thus, no one can hear differences in cables. Yet, that is/was the wrong argument --- interfacing with Zout of amps and cable C of phono cartridge/preamp interfacing affect frequency response of the "system". And, it is quit audible in many situations.
I think it is better to err on the side that a lot of people over a long time - across the world- say they hear - even if we dont understand what to measure or even which marker to test for.
Thx-Richard Marsh
Last edited:
Could that be a hint that, in case of taking BW a design priority, when people claim they have enjoyed the listening pleasure from a wide band amplifier, they might as well have been experiencing the compromised distortion in a positive way, if such amp did trade distortion for speed?
There is no reason to make any such assumptions that wide BW designs are compromised in distortion. With CMA it can be the opposite at the higher freqs. At some point the (harmonic) distortion is not audible. Or, its Blameless. Got to look somewhere else.
-RNM
-RNM
Last edited:
Perfect, Richard. And thanks, one more, to put the things in the right perspective.
I mean, the earth is not exactly round.
it leave a little space for our feelings.
That does not excuse us from being as clairvoyants, scientific and smart as we can be.
Last edited: