Your thread is an excellent source of info on the subject; helped to clarify what CFA is and what it is not. Thank you for helping to make CFA better understood. Yes, you have been quit neutral yourself. As a test and measurements guy, I understand both sides. However, in this instance, I take a position as JC does when he says - I have to go by 'what works' or what I hear.
THx-RNMarsh
Last edited:
Just before Scott activates the "ignore car talk" button.
It's not only high static pressure
George
I use a plasma system/plug myself.
THx-RNMars
By the way, can-you report the difference you noticed in his preamp between the no global feedback and CFA choices ?
I'm very curious to this, and believe... I can use your ears for a while ;-)
I will try to build the best OS one, to figure out something in the order of Simple/Complex, and see if I can hear any benefit for the highly reduced distortion.
What seems missing here, is the specific reasons why one input topology might be better than another for audio. It isn't just higher slew rate, as slew rate can be almost as fast with VFA's as much as CFA's. There is a LIMIT where slew rate by itself is useful. However, for a given amplifier topology, the higher the gain-bandwidth, the better the amp, (in general) and this reflects both in higher open loop bandwidth and slew rate. It appears that higher open loop bandwidth is more significant than increased slew rate, over 1V/us/Vpp. Matti Otala figured this out in the late 1970's, after we tried and tried to find the worst case dV/dT (slew rate) from virtually all audio sources that we could find. However, PIM is NOT completely understood or defined as of yet, and CVA's have higher open loop bandwidth over VFA's all else being essentially equal.
Personally I cannot agree that you can vastly improve Xover distortion with extremely high slew rates, although it might help a little. It is better to focus on designing out the Xover distortion in the output stage, that is what I do for sure.
Personally I cannot agree that you can vastly improve Xover distortion with extremely high slew rates, although it might help a little. It is better to focus on designing out the Xover distortion in the output stage, that is what I do for sure.
Last edited:
I notice something odd - the current level of disussion regarding VFA and CFA somehow assumes that CFA is a brand new all open field, a opposed to VFA which is somewhow assumed to be a done deal, everybody knows everything there is to know about it.I feel both premises are completely wrong - CFA has been around for sometime, although more sort of lurking, and there's plenty of juice left in the VFA.
A small example - i recently tried for cutiosity's sake to chnage the inverting input voltage and current from same value resistors at both noniverting and inverting inputs. Used to a big NO-NO. So my noninverting resistor stayed at 33k where it was, but my inverting resistor dropped down to 8k2, with a 330 Ohm voltage divider. And lo and behold, what I got was more stability, better bad load drivr (where bad load = 3 Ohms//2 uF), more than enough to reserach such a solution in much greater depth than just a one off attempt.
That in fact puts me somewhere in between VFA and CFA. Perhaps that is the way to go?
A small example - i recently tried for cutiosity's sake to chnage the inverting input voltage and current from same value resistors at both noniverting and inverting inputs. Used to a big NO-NO. So my noninverting resistor stayed at 33k where it was, but my inverting resistor dropped down to 8k2, with a 330 Ohm voltage divider. And lo and behold, what I got was more stability, better bad load drivr (where bad load = 3 Ohms//2 uF), more than enough to reserach such a solution in much greater depth than just a one off attempt.
That in fact puts me somewhere in between VFA and CFA. Perhaps that is the way to go?
(The peanut gallery, spitting the shells on the floor while never engaging in any technical discussions, is always right).
And, taking it from here, how much slew rate do we really need? At what point will be enogh, beyond which the nothing hew further helps? I never believed in one time, one sleigh of hand cure-alls. We all have our own ways of dealing with what John has mentioned, for example I use low voltage, but high gain BC devices (selected for 400 Hfe, usually BF 547B/BC557B od C types), knocked down to 10V, but lit up at 9+ mA of bias current, followed by an upper cascaded pair, as I suspect many of us do, and it works like a charm. This then allows me to run my VAS stage at like 27 or 28 mA, the price being a loal heat sink, but the payoff being an excellent slew rate.
My family had a 1989 Sable station wagon with the Vulcan engine. Took it to my schizophrenic uncle to fix one time. We couldn't be sure exactly what he was doing with it, but when we got it back it got 28MPG and would leave smaller cars in the dust on the highway. We still want to know what he did but we are nervous to pry.
Most amps are very nonlinear near the slew limit. Some margin is a good thing.
Since the thread loosely relates to vinyl source, slew rate of programme material equates to stylus acceleration equates to groove curvature...............all of which was controlled in the mastering process. Yet, if you measure playback performance, there's no significant diference between vinyl and digital sources as to slew rate- the limitation is typically in the source material.
Perhaps and some guardband is harmless/appropriate. But little programme content challenges decent amps, even when run at seriously unreasonably (high) levels.
Last edited:
Some remarks of Putzeys in The F-word I think may apply, although it's been a while since I read it.
And although there may be content arising from vinyl issues, you can knock it down close to the input(s). And remember to keep surfaces clean. CDs are limited in their ability to overload stuff, unless something is terribly wrong.
- Status
- Not open for further replies.