[1UP]

I recall it's been said Sanyo OSCON caps are well thought of for PSU decoupling in digital circuits.

Is that the case? If so, what's special about them vs any respectable electrolytic - say Panasonics/Nichicons?

I just can't grasp why any particular electrolytic should be preferred for PSU decoupling in analogue vs digital circuits.

 

[Tomo]

I think it probably means it is designed to operate at lower voltage. Digital signal ... probably means on motherboards ... operate at low voltage well below 15 volts. Digital signals are designed to defeat noise but it does have vulnerabilities to noise. OSCON is trying to sell by marketing that it is for digital implying it has low noise level. However, this is probably isn't strictly true.

I would pay attention to electrolytic caps that are rated for "low leakage". OSCON is sometimes rated high ripple current, I thought 1A/usec or some absurd amount. Which is why I got my stock only to realize later I am pretty happy with Panasonic HFQ (discon).

Anyone know direct replacement for these?

T

 

[majkel]

Quote:

Originally Posted by 1UP /img/forum/go_quote.gif I just can't grasp why any particular electrolytic should be preferred for PSU decoupling in analogue vs digital circuits.

Because it's not always optimal to have low ESR capacitors in analog section power supply circuits. For the digital this assumption is more correct, together with low ESL. The analog supply section benefit from low THD at higher audible frequencies (10kHz is often reported) as these capacitors contribute to the signal path as the power supply is for the music the impedance closing the signal path.

 

[hopeless]

Quote:

Originally Posted by majkel /img/forum/go_quote.gif Because it's not always optimal to have low ESR capacitors in analog section power supply circuits. For the digital this assumption is more correct, together with low ESL. The analog supply section benefit from low THD at higher audible frequencies (10kHz is often reported) as these capacitors contribute to the signal path as the power supply is for the music the impedance closing the signal path.

You simply bypass them with a small polyester cap (.33 uF + .1 uF Wima MKS2 in my case) and that should provide the low distortion at high frequency...

 

[guzzler]

You can't really measure "THD" for a power supply, and paralleling capacitors like that is asking for trouble in high speed digital circuits. Again, noise isn't something you can measure for a capacitor in this instance. OSCONs are excellent for digital in that they have fairly constant ESR, ESL, and capacitance across a large frequency range. The large ripple current rating is also related, as digital circuits can have enormous instantaneous current demands if a large number of gates are simultaneously switched (each gate has an associated capacitace...). They're not well thought of as coupling capacitors for audio purposes.

Regards parelleling, have a look back through some of my previous links, in particular:

UltraCAD datasheet

If you can manage the maths, great, if not, just look at the diagrams. Very informative...

 

[Tomo]

Hey I found it. What Guzzlar said is right here!

DIY Paradise - Sanyo OS-CON

 

[hopeless]

Quote:

Originally Posted by guzzler /img/forum/go_quote.gif You can't really measure "THD" for a power supply, and paralleling capacitors like that is asking for trouble in high speed digital circuits.

I was commenting majkel, who was talking of analog power supplies. Specifically, the parallel of a Rubycon ZL 680uF + Wima 330nF + Wima 100nF is the power reservoir for the output stage opamp inside my Super Pro. It is before the regulators that bring energy to DAC and receiver (and USB interface).