Absolutly, wahab. That's why i proposed to look in closed loop for the bandwidths and phase curves at the input of the second stage ( just after input signal has been mixed with NFB one).
You can compare that way the frequencies where feedback ratio begin to decrease (open loop flat frequency limit as an increase of the response curve level) and what happens at the limit of the frequency range of the amplifier, without modifying their behaviors.
This is a very good summary. The only problem is that it leaves not much to argue about, and arguing is the soull of this thread
jan
Are you saying, in other words, that one cannot build an amp with an UGBW (unity gain bandwidth) larger than 600kHz-1MHz without stability issues?
Imagine your amp is set as unity gain follower (other cases are worse), having 30dB GNFB at 20kHz means that its open loop gain at this frequency is 30dB. With a 6dB/octave roll-off, 0dB is achieved at 640kHz (the UGBW)
That doesn't sound that much to me, nor impossible without stability issues.
Last edited:
I cannot comment on this Michael. You may well be correct - I would have to sim this to see.
Where's Harry?
TPC will get you to much higher than 30-35 dB at 20 kHz and that's one of the tricks to get single digit ppm in VFA's. Similarly, TMC can allow about 15 dB of additional feedback at 20k.
I believe the correct statement is you cannot get much more than 30 or 35 dB of feedback at 20 kHz with standard MC.
Mosfets will also allow the ULGF to be pushed up so you may be able to claw a few more dB there as well
I agree with the idea but with different proposal considering that:
(1) You started with a bad cfa amp when you modded Crescendo into cfa.
(2) We cannot say that solid state is better than tube by trying to replace the transistors in a blameless amp with tubes. We treat components based on its character and design objectives.
(Instead of comparing measurement result we can otoh put the minimum standard measurement as the objective to achieve with both cfa and vfa and compare the result by listening).
With vssa i can see M400 or P101 (non symmetrical version) as the equivalent vfa. Here the vfa tend to have better bass and cfa better highs.
Now pay attention to ska gb150d. Cfp output is not suitable for fast amp, especially high input capacitance hexfet. This will be a great challenge to turn the dual ltp input to become a cf one.
Member
Joined 2009
Paid Member
Early on here, i would like to add some definitions that have been accepted in industry. I think it will help to seperate what is what for those not intimately familiar with the topology -- to be expanded upon, later ->
"True current-mode signal processing may be defined as the processing of current signals in an environment where voltage signals are irrelevent in detrmining circuit performance."
It is where/when this ideal deviates that the lines between CFB and VFB become blurred. Such as impedance increases with freq making the nodes less than ideal.
Thx-RNMarsh
"True current-mode signal processing may be defined as the processing of current signals in an environment where voltage signals are irrelevent in detrmining circuit performance."
It is where/when this ideal deviates that the lines between CFB and VFB become blurred. Such as impedance increases with freq making the nodes less than ideal.
Thx-RNMarsh
This begs the question: Would a BJT LTP used in an inverting topology be considered current mode operation, as both bases ideally never deviate from 0V?
Jay just made a statement about the response of a vfa vs a cfa and while I understand where the difference may be coming from at the 20khz range what would cause any difference at the other end of the scale down at let's say 20hz? Why would a vfa have better bass response where I would think these two topologies would be the most similar?
differences in Zo and PSRR can contribute to this. same question was asked when SS amps appeared....about why bass was tighter, etc compared to tube amps. Lower Zo etc.
However, you also need to be sure peak current output is same to compare 'similar' circuits.... as this also affects bass quality a lot.
-THX RNMarsh
Last edited:
Are you referring to my question about BJT LTPs used in inverting mode? If so, I find it a cheap answer as my question is directly related to your quoted argument. Along with the fact that current mode operation is at the core of CFA operation.