Hi, I have a 20W class A solid state amplifier with 3 x 10,000uF 63V per channel in its power supply section. It works fine and sounds great but of course it runs hot.
What I'd like to know is has anyone tried using a bypass capacitor in parallel with a filter cap? Is it a good/bad idea? If affirmative:
1) Will it make an audible improvement?
2) I have a 10uF 400V electrolytic and a 0.1uF 800V film cap, which is better as bypass capacitor in this case?
3) If i use the 10uF electrolytic on each of the them I'd be adding 30uF in total per channel, is it better to use just one in parallel with the group?
Thanks and regards.
What I'd like to know is has anyone tried using a bypass capacitor in parallel with a filter cap? Is it a good/bad idea? If affirmative:
1) Will it make an audible improvement?
2) I have a 10uF 400V electrolytic and a 0.1uF 800V film cap, which is better as bypass capacitor in this case?
3) If i use the 10uF electrolytic on each of the them I'd be adding 30uF in total per channel, is it better to use just one in parallel with the group?
Thanks and regards.
If you use any bypass capacitor, install it near the circuit that is powered, rather than at
the power supply capacitors. If the supply feeds two circuits, install a bypass capacitor
at each circuit. The 10uF would probably be more effective than the 0.1uF. Whether it
will make any difference depends on many factors, so just try it.
the power supply capacitors. If the supply feeds two circuits, install a bypass capacitor
at each circuit. The 10uF would probably be more effective than the 0.1uF. Whether it
will make any difference depends on many factors, so just try it.
Last edited:
A lower value capacitor is sometimes used across very high value (~10,000uF) electrolytics because those big caps (although very effective at smoothing 100/120 Hz ripple) don't behave well at very high frequencies. AFAIK, the best approach is a 100nF film cap across the big cap; also (as rayma pointed out) putting a cap RIGHT AT the circuit will help.
Hi,
Search the Power Supplies forums for username Eva and keyword "bypass". Here are some very interesting threads:
Bypass capacitor theory
Power Supply Decoupling
Is bypassing PSU capacitors effective?
paralleling film caps with electrolytic caps
Search the Power Supplies forums for username Eva and keyword "bypass". Here are some very interesting threads:
Bypass capacitor theory
Power Supply Decoupling
Is bypassing PSU capacitors effective?
paralleling film caps with electrolytic caps
Thanks for the great links . After going through them and after inspecting my amp, I've decided to preserve design decisions made by its maker since I know nothing about circuit design 
. The 3 x 10K lytics had 0.47 ohm 2W resistors in series along the positive rail and a smaller 220uF lytic in parallel with last 10K before connecting directly to the first output transistor, a pair of 2SK1058s.
I've upgraded that 220uF to a Silmic II so I think I'll just stick to parts upgrade than mess with design for now
. Aside from other caps I upgraded I also replaced the spkr out cabling to Supra classic. The input signal cabling will be next and the last, trying litz silver wire for this.
. The 3 x 10K lytics had 0.47 ohm 2W resistors in series along the positive rail and a smaller 220uF lytic in parallel with last 10K before connecting directly to the first output transistor, a pair of 2SK1058s.I've upgraded that 220uF to a Silmic II so I think I'll just stick to parts upgrade than mess with design for now
. Aside from other caps I upgraded I also replaced the spkr out cabling to Supra classic. The input signal cabling will be next and the last, trying litz silver wire for this.For a class A amplifier, bypassing the supply won't get you much since by definition, the amp runs in constant current mode when operating in class A, so there won't be much current variation to bounce the supply around. There'd be nothing for the added cap to stabilize. Of course, if you drive the amp hard enough for it to go out of class A operation, then there would be supply current variation and added capacitance might have an effect then.
You got your definitions wrong. Very few class A amps run the output at constant current.
Are there not a million threads about bypassing PS caps and their effect on sound?
I have done several low-noise PSU and low-noise measurement devices (not ultra-low-noise but quite low noise) and had never seen that by-pass caps (parallel to large electrolytic) significantly helped. They help very few (unsignificantly) or usually don't help at all.
Last edited:
Not so much for reducing noise per se, but ensuring that the power supply impedance stays low as frequency rises. The amplifier will likely behave closer to its design ideal the closer the power supply is to zero AC impedance.
Many class A amps run the output stage at a constant current, many do not. Push pull class A and single ended class A are fundamentally different. But both would “prefer” a low impedance power supply, from DC to light.
Just adding bypass caps willy nilly doesn’t necessarily ensure the power supply impedance stays low. Sometimes you can introduce resonances. Unintended, but it happens. Two caps and an inductor between them - it goes high impedance *somewhere*. I suspect THAT is what is usually responsible for audible differences people hear between various cap types when they go swapping them out.
Many class A amps run the output stage at a constant current, many do not. Push pull class A and single ended class A are fundamentally different. But both would “prefer” a low impedance power supply, from DC to light.
Just adding bypass caps willy nilly doesn’t necessarily ensure the power supply impedance stays low. Sometimes you can introduce resonances. Unintended, but it happens. Two caps and an inductor between them - it goes high impedance *somewhere*. I suspect THAT is what is usually responsible for audible differences people hear between various cap types when they go swapping them out.
I'm talking about general system noise, or about power supply noise if it is regulated. So my aswer is "They don't help at all".
Last edited:
Many have and if you are sort of semi knowledgeable on how to apply these it will be an improvement unless you try on a high end type thingy, it's then more likely to have been optimized as is.