I just read an interesting article about power supply design and would like to draw your attention to figure 6B and 6C.
Design Considerations - 2
The author describe the "spectrum of harmonics" from capacitor input and choke input filter is different, and that would affect the sound. I guess it is saying that the ripple does not only sound like 100/120hz hum but all different kinds of buzz/harsh at higher frequencies. I don't notice much elaboration on such view from the internet and the main focus on PSU design has usually been how much p-p ripple remains after filtering. Nevertheless I did noted a reply in a post in this forum that choke input sound "ease". That comment seems coherent to what the author suggests.
https://www.diyaudio.com/forums/tubes-valves/101064-motorboating-post1199057.html
What is your view or experience on choke-input filter on output stage / preamp / stage?
Design Considerations - 2
The author describe the "spectrum of harmonics" from capacitor input and choke input filter is different, and that would affect the sound. I guess it is saying that the ripple does not only sound like 100/120hz hum but all different kinds of buzz/harsh at higher frequencies. I don't notice much elaboration on such view from the internet and the main focus on PSU design has usually been how much p-p ripple remains after filtering. Nevertheless I did noted a reply in a post in this forum that choke input sound "ease". That comment seems coherent to what the author suggests.
https://www.diyaudio.com/forums/tubes-valves/101064-motorboating-post1199057.html
What is your view or experience on choke-input filter on output stage / preamp / stage?
> "spectrum of harmonics" from capacitor input and choke input filter is different
Be clear: his "choke input" is an L+C filter, his "cap input" is just a C filter. Two reactances is surely better than one?
Do you understand loudspeaker crossovers? Then you can picture what an increasingly complex woofer filter can do to the treble. L or C 6dB/oct, LC 12dB/oct, LCL 18 dB/oct, LCLC 24dB/oct....
Be clear: his "choke input" is an L+C filter, his "cap input" is just a C filter. Two reactances is surely better than one?
Do you understand loudspeaker crossovers? Then you can picture what an increasingly complex woofer filter can do to the treble. L or C 6dB/oct, LC 12dB/oct, LCL 18 dB/oct, LCLC 24dB/oct....
It is true that simple cap. I/P filters produce "triangular" ripple waveforms, while a simple choke I/P filter's waveform is nearly sinusoidal. However, the net ripple level of a simple choke I/P filter is unacceptably high. Older editions of the ARRL Handbook advise LCLC filtering, to obtain both good regulation and low net ripple. Can you say weight, space, and expense?
Large valued cap. I/P filters can yield OK regulation and tolerable net ripple levels, but lots of said ripple is nasty harmonic "hash". FWIW, I've repeatedly posted, on several "boards", regarding the suppression of "hash".
TANSTAAFL definitely applies to the subject of PSU filters.
Large valued cap. I/P filters can yield OK regulation and tolerable net ripple levels, but lots of said ripple is nasty harmonic "hash". FWIW, I've repeatedly posted, on several "boards", regarding the suppression of "hash".
TANSTAAFL definitely applies to the subject of PSU filters.
It is more than weight, space, and expense. It is also the power factor. Charging currents are much lower, but rectified voltage is lower as well.