power supply design question 832A amp

[trondareo]

I am hoping someone can help me understand the design of this Power supply: audio diy

I have been running PSU designer simulations of the B+ supply for this amp and unsurprisingly a bridge rect. followed by a simple C filter of 500 - 800uF ends up with a lot of ripple current. I have built the amp with 800uF and it sounds nice on a breadboard. No detectable hum.

A Pi filter would obviously be smoother, and I can make this work in the simulator with my amps 218V transformer secondary, and 156mA current at B+. ALso I need to bring town the B+ voltage a bit.

So why is the single 800uF filter chosen and why have other builders reported using even greater values of C?

[prairiemystic]

I find people going overboard on filter capacitors in tube gear, with the resultant very high ripple currents and transformer buzz. If you have a wimpy power transformer, then a large filter cap can hold up during transients in the music. My experience is the rectifier switching transients make a lot of noise/RF and bigger filter caps can attenuate this better. So a bigger filter cap sounds cleaner for the wrong reasons. It's easier to add a film cap 100-470nF across it, and 10nF across the transformer secondary, and use fast recovery rectifier diodes, like UF4007's. Try a pi-filter with smaller, 100-220uF's and see. In the old days they used 10x less, or 80uF... note 800uF at 300V is equivalent energy storage to 30,000uF at 50V !

P.S. that PSU has no bleeder resistor, which can be dangerous, those caps can bite!

[SemperFi]

+1 to what prairiemystic said.

I am one of those who am now trying to downscale my capacitor batteries. In the mid 90s the craze was oversized transformers and massive capacitor banks. I started off like most student into audio building solidstate amps. My first amp had 10watt output, 1000VA trannies and >100 000UF capacitans! It was important to be bigger than the other dude. When I got into tubes, the same idea of big is better followed. I have finally realized it's a rather crude way (not very good engineering) of getting low noise, with large amounts of strain on the supply and even HF noise due to extreme ripple currents.
In the 'good old' days capacitors were expencive, engineers worked to create 'good engineering' without unessesary dill-dall, and some issues like current reserves were probably not though of? Hence classic circuits have small (but adequite) capacitances.

Today perhaps something in the middle is a good ballpark?

For low noise in PP amps not much capacitance is nessesary. In one amp I use 30UF as first cap, feeding a PP pair of 6V6, and it is so quiet I sometimes wonder if it is turned on.

[trondareo]

Thanks for great input guys!
I was coming to the same conclusions through playing with PSU designer, but needed some confirmation that I was thinking correctly.