You need to realize that he/she is of a different/relative universe, no guarantee laws of physics works the same way. His/her post is for the benefit of people of that universe, most likely bear no coherence to the rest of us.
These designs account for 90% of issues with DIYA projects. 20th century burning amps are often CFP's. I did not mean that the issue is taboo ,
But this is an issue of diminishing returns.
OS
That makes things clearer, thank you. I think this goes with the last part of the post from #19. It's got this difficulty that comes with it. So unless I know what I am doing, I should compensate for it. Like the loading it with a zobel and or not getting the input floating because I've read that it is prone to oscillation.
As for the burning amp, there was indeed a point in which when I was testing it, things started to smoke 😵. Really, the only reason as to why I chose CFP is because of the lower voltage bias when using diodes, that's it. But now, why bother if you can do darlington and Vbe Multiplier.
I had to start with EF2 , I was scared of EF3 instability until I found the Harmon Kardon OPS (EF3).
Current gain of 100K+ , 250V/uS (fast) . I looks a little more complicated , but will last many decades.
You can drive it with 3-4mA input stages .... run soooo cool. Since it does not load the input stage at all - super low distortion.
VERY high damping factor , it can take the craziest speaker loads with ease and drive 4R loads , as well.
You can also use outputs for the second stage , to drive 4-6 pairs of actual outputs. 500W wolverine OPS !!!
PS - If you layout (really)anally 🙂 - grounding/rail cancellation..... you can get part per million perfection.
Attachments
Last edited:
THE chance, sound related! by means of Sziklai is the symmetrization of the different tonal component characters: both parts of a pair of complementary transistors work in each half-wave. But a maybe even better idea is to build up balanced (2 x 2 amps): simply by transformer at the input. Also in this way each half-wave sees both transistors of a complementary transistor pair, and moreover the power supply characters are almost cancelled, these modulate the signal much less. Disadvantage of summing components, CFP or balanced: these components continue to modulate the signal, the sound. Another curtain of, most often, artificiality, fog, colorlessness, flatness is drawn into it.
;-)
;-)
See what I mean? They cumbb out of the woodwork. Which won’t help solve an oscillation problem, because that requires either exact models or at least a structured, logical progression of experiments in the case of less complete models.
The problem with pseudo-science is that sometimes it’s almost right or at least partially, often purely by chance. I think #27 is trying to refer to even harmonic cancellation, with no idea of what it actually is. And it’s not just a function of or strictly limited to the Sziklai connection.
It is when you try to optimize a CFP-based amplifier that you will find out how much of a pain in the *** it is. Build the same thing with an EF3 - you turn it on and it just works. And in the optimization process things behave more intuitively and according to models, even if they are basic and perhaps incomplete. There are only two cases where I’ll consider CFPs - one is if I just don’t want to give up that one volt of signal swing that you gain with it, and the other is if the two output transistors MUST be on separate heat sinks. And in neither case would I care about a little overshoot or excessive damping in the square wave, or if it actually gets unstable directly driving a 10 uF capacitor.
The problem with pseudo-science is that sometimes it’s almost right or at least partially, often purely by chance. I think #27 is trying to refer to even harmonic cancellation, with no idea of what it actually is. And it’s not just a function of or strictly limited to the Sziklai connection.
It is when you try to optimize a CFP-based amplifier that you will find out how much of a pain in the *** it is. Build the same thing with an EF3 - you turn it on and it just works. And in the optimization process things behave more intuitively and according to models, even if they are basic and perhaps incomplete. There are only two cases where I’ll consider CFPs - one is if I just don’t want to give up that one volt of signal swing that you gain with it, and the other is if the two output transistors MUST be on separate heat sinks. And in neither case would I care about a little overshoot or excessive damping in the square wave, or if it actually gets unstable directly driving a 10 uF capacitor.
I would only solve oscillation problems to the extent that they a) are not audible and b) do not destroy the circuit and more;-)
But I think it's obvious that I didn't want to talk about oscillation problems, but about the "experientially-scientifically" provable different sound characters - NOT harmonoics! - of different components;-)
Aside: Audio and audio-electronics are essentially experiential science, which far exceeds the complexity of the common training and ideas physicists and electrical engineers have of "science" and methods: switch on, lamp on, switch off, lamp off;-)
But I think it's obvious that I didn't want to talk about oscillation problems, but about the "experientially-scientifically" provable different sound characters - NOT harmonoics! - of different components;-)
Aside: Audio and audio-electronics are essentially experiential science, which far exceeds the complexity of the common training and ideas physicists and electrical engineers have of "science" and methods: switch on, lamp on, switch off, lamp off;-)
;-)
I would not recommend rebuilding a "Wolverine". Too complex to get even a single tone through cleanly, to get even a single watt through audibly undistorted, recoloured.
Regarding your complementary transistors push pull follower: The signal does pass half-wave different circuits: In addition to the half-wave different sound characteristics of the transistors, trimmer 2k, resistor 3k9 and Tr BD139 modulate the "upper" half-wave. That is audible: (more) "noise".
It might be more advantageous from a sonic point of view to feed back the power transistors by means of an emitter-base resistor instead to use a further pair of transistors.
Just try;-) Have much fun;-)
https://oshwlab.com/perseus/jat501-johnaudi-tech-discrete-amplifier
I would not recommend rebuilding a "Wolverine". Too complex to get even a single tone through cleanly, to get even a single watt through audibly undistorted, recoloured.
Regarding your complementary transistors push pull follower: The signal does pass half-wave different circuits: In addition to the half-wave different sound characteristics of the transistors, trimmer 2k, resistor 3k9 and Tr BD139 modulate the "upper" half-wave. That is audible: (more) "noise".
It might be more advantageous from a sonic point of view to feed back the power transistors by means of an emitter-base resistor instead to use a further pair of transistors.
Just try;-) Have much fun;-)
https://oshwlab.com/perseus/jat501-johnaudi-tech-discrete-amplifier
Last edited: