Whats wrong with large filter caps for Gainclone?

[Mesh]

1 - originally posted by janneman: The data sheet of some of these chips shows small supply caps;

The problem I have with that is that the data sheets are aimed at low end applications - not esoteric HIFI. I dont think National would ever recommend 30000 uF for a chip amp. I also dont think they would be suggesting 300 VA or 400VA per channel! for a 30 watt amp running class AB.

[Mesh]

JO DIRT wrote: I for one use 15 000 uf per rail in my p/s for the lm3886 amplifier that I built. The p/s is remotely located so I follwed basic design principles and placed 1000 uf/ .01 uf caps directly by the chip.

This is how I think I would build this amp !

[Mesh]

I always thought that the ultimate power supply for an amplifier would be a battery ( apart from the obvious problem of producing 30V-0-30V at several amps) . Is this not true for a gainclone?

[HDTVman]

Hi mesh

Go a head and build your GC with the big power supply. In would go for regulated supplies myself.

I am convinced that a large number of folks out there in audioland like the sound of a soft power supply. Thats what they have listerned to all of there lives. Thats how they think electronic sound reproduction should sound. Hey, thats OK. If thats what you like more power to ya. Oh, you need the big power supply for that LOL.

later BZ

[millwood]

Originally posted by janneman


1 - The data sheet of some of these chips shows small supply caps;Jan Didden

on the same token, I don't recall national suggesting dual / mono transformers, fast recovery diodes, diamond heatpads and Egyptian maple casing either, .

The chip was probably designed for low-end applications and it wouldn't help its cause had National suggested all those "hi-end" stuff.

Plus, it just may not benefit from lots of high-end stuff.

my gainclone is running off about 9900uf per rail at the PS. Putting 3300uf per rail on the chip doesn't improve nor deteriate sound either - to my half deaf ears anyway.

[millwood]

Originally posted by Joe Rasmussen
Keep this in mind, the Gainclone chip has an awesome 120dB PSRR (typically).
Joe R.

yeah. And it goes downhill from there: 60db at 10k, and 20db at 1mhz for -psrr.

[Richard C]

Exactly right. Even amps with high PSRR benefit from well regulated PSUs. High PSRR is does not mean that the PSU doesn't matter.

A DC voltage power supply does just that. Perhaps those who run there amps form batteries could improve the sound by putting resistors in the supply rails and injecting a bit of load depended 50Hz... Nice

[Pedja]

Originally posted by millwood
The chip was probably designed for low-end applications and it wouldn't help its cause had National suggested all those "hi-end" stuff.

Did you miss something here?

Pedja

[planet10]

Some possible technical reasons:

1/ i'd expect the biggest would be the HF spurii (sp?) that any cap input power supply produces... the bigger the caps the higher the frequency of the grunge (and with the same energy packed into a smaller time-frame, a taller spike). As Millwood points out the PSRR of he chips falls as frequency goes up, so lower frequency/smaller caps is good.

2/ the impedance of a smaller cap is typically smaller making the supply lower impedance (this could be negated by using lots of small caps)

dave

[Nuuk]

Here we go again! Another drawn out debate over something that is surely subjective.

It is so easy (and reasonably cheap to build one of these Gainclones that it could be done in less time than it takes to read threads like this.

So the answer is build one - with a standard supply, a regulated supply, with two or even four transformers, one, two or four bridges, with small caps, huge caps or with batteries.

When somebody has built the lot, they can post on a thread like this with some authority. Otherwise we are just going over the same ground yet again; and I'm trying to avoid ruts!