Why does Leestereo insist on increasing power supply/decoupling capacitors?

[ConradH]

IMO, most concerns about capacitor types for 50/60 Hz conventional audio power supplies aren't worth wasting a brain cell on. The only factor that affects performance is the capacitance value. Dissipation factor (or ESR) barely gets involved. Let's take a 6800 uF cap with no losses, a perfect cap. It would have an impedance of 195 milliohms at 120 Hz. Now lets say the dissipation factor is 0.3, which is more than three times greater than the datasheet says it should be. Impedance is now 204 milliohms and the 120 Hz ESR is 58 milliohms. Not worth worrying about.

Only slightly worth a brain cell is if they sent you a capacitor that was 10% low, which IME is quite likely. If it had no losses, the impedance would be 216 milliohms. Still of no concern, but it's a bigger factor than that made by low ESR caps designed for switching supplies. Instead, worry about lifetime, solvent resistance and if the lead spacing matches the PCB. Stressing the seal by spreading the leads is worth worrying about. If you need a low leakage cap for low level sections, that's also worth worrying about. FWIW, I've never measured an "Audio" capacitor that was one bit different than standard commercial grade parts.

I haven't worked on anything, much less bought new caps, in a few years. I see FG is now "Fine Gold" and recommended for audio. Last I knew it was just a good commercial part. The datasheet looks the same as before. Is it something different or the same old part with a fancier shrink sleeve?

 

[Leestereo]

...the fact that i replaced the caps with the nominal values, ill replace them with the bumped up values you mentioned in your AU-717 restore thread. Ill go with the Nichicon UPW series better designed for pwr ckts...

The regulated power supply board (F-2663) in the AU-717 runs quite hot due to the high parts density and relatively poor air flow cooling; and after >40 years of operation, many of the components on this board are failing or damaged from being "cooked". Hence, IMO, the installation of low ESR 105°C capacitors on this particular board should be the SOP.

 

[ConradH]

105°C - yes!

 

[bolero]

Are all 105+ degree rated caps low ESR?

 

[Shominy]

I upped my Technics SA-505 filter caps from the stock pair of 8200uF to a UCC 12,000uF pair. One did not fit down to the pcb due to increased diameter, so I made sleeved legs so that it sits about 3/4" off of the board, and hot glued it to the other one on the pcb.

It did not make a night/day sonic difference, but there is one. It's a bit punchier on the low end to my ear, and though I do not have the equipment to measure things like RMS power and ripple currents, I do feel that the amp is better ready for transients.

And the dial lights don't pulse any longer above 40% volume. This is a night/day difference.

 

[Leestereo]

Are all 105+ degree rated caps low ESR?

Not necessarily, e.g., Nichicon VZ are rated at 105°C, but are not low ESR types. However, the converse is usually true, i.e., low ESR capacitors are typically rated at 105°C.