regulated supply for GC

[yuval]

I saw many GC projects around here but most of them with simple power supplay - just a recitfers + cap. why not use regulated supply ?

[Nuuk]

Look closer and you will see that R!sc has used a regulated supply based on an idea by Pedja Rogic. I am currrently building the same design.

[yuval]

hi Nuuk !
it was the trigger for my question ... why don't we see more like this ? simply because it's more complicated ?

yuval

[peranders]

yuval, why would you need a regulated power supply? I think you should ask yourself this first?

Have in mind that the LM3875, 3886 have very good (above average) PSRR!

[yuval]

I really don't know but let's say I want to maximize potential of this design ?

[jackinnj]

Originally posted by yuval
I really don't know but let's say I want to maximize potential of this design ?

the first simple reason is that you are burning some power in the pass element, whether it is a MOSFET or a Bipolar transistor it is going to burn (watts in the atmosphere) the energy represented by the difference between the supply voltage and the regulated voltage, plus ... why burn watts when you can turn them into SPL's.

Secondly, any regulated supply relies on feedback from a voltage or current sensing circuit to an error amplifier which drives the base or gate of the pass transistor -- thus you have to contend with the compensation of the feedback loop, timing considerations, noise of the error amplifier --

now, if you happen to pick up a couple of Lambda 28 volt, 5 amp supplies I would not dissuade you from using them (2) to drive the LM3875 or LM3886.

[jackinnj]

a nice hefty lambda regulated supply on EB

[matjans]

why burn watts when you can turn them into SPL's.

why not build a (digital) class d amp then?

how efficient are these chip amps anyway?

[Nuuk]

it was the trigger for my question ... why don't we see more like this ? simply because it's more complicated?

If I am not misquoting Pedja here, one of the advantages of using a regulated supply is to be able to use more capacitance without the detrimental effects that has on a Gainclone with a standard PSU.

But YES, it is more complicated (remember simplicity of design is one of the big attractions of the GC) and more expensive unless you have some biggish PSU caps/spare stransformers lying around (which I did have fortunately).

If you go for a separate supply for each channel then it also becomes an exercise in logistics, with all the extra connections to the amp (I am also having to 'feed' my opamp buffers with separate supplies too).

So why am I doing it? Well like Yuval says, it is the age-old reason for doing most things in DIY hi-fi: that is to try and maximise the performance of whatever it is that you are playing your music through.

I have the greatest respect for Pedja, it was his work on buffers (as well as Joe Rassmussen) that encouraged me to build a buffered GC and that has payed large dividends. Pedja says that he prefers the sound of the regulated supply and that is a good enough reason for me to try although having the transformers and caps form my Arcam A60 amps is a big bonus so the outlay for me is not such a gamble!

[Ropie]

Originally posted by Nuuk
So why am I doing it?

Nuuk, I think you will eventually outbuild everyone in terms of the sheer amount of (high quality) stuff that you seem to get done