Ultra quiet DC supply +12/-12/+12?

[phrarod]

Some reason the search isn't working right now on the board so if this has in anyway been answered my apologies.

I've been playing around with separate power supplies to my Audio Note Dac board with improving sonic results. Now that its decoupled from the main supply I want to build an ultra quiet supply.

The board calls for +12/-12/+12 and their associated grounds. Any ideas or advice are welcome. I have had good results with switching supplies for other applications but don't know if it is what will work here.

 

[georgehifi]

If your lazy like me, and want something already in kit form, these are Russ White's great low noise adjustable supplies.
I have the Bipolar supply running an Analog Devices AD1955 Evaluation Board and it is sounding very very good.

Low-Current Bipolar Power Supply
http://www.twistedpearaudio.com/lcbs/lcbs.aspx

Low-Current Dual Power Supply
http://www.twistedpearaudio.com/lcps/lcps.aspx


Cheers George

 

[phrarod]

Why yes, I am lazy.

Thanks George I'll check them out.

Looks cool. I guess I would have to get the dual to give me enough outputs. I emailed them a question.

One thing about the ps is I do want it to be stout. I noticed the ps I'm using now while quieter because its separate it doesn't have as much dynamics when I was hooked up to the main ps.

 

[phrarod]

I'm about to order - turns out I would need two due to the inputs on my dac board. I showed them to a friend I really respect and his reply was "The ideal power supply would be a shunt regulator not a series regulator".

Thoughts?

 

[jwb]

Lots of people have an irrational knee-jerk affinity for shunt regulators. There's nothing that makes a shunt necessarily superior in any given application. Maybe you should just try both and see what you like?

I assume this supply is used for some op-amp power rails or something like that?

 

[ackcheng]

Can also consider these

http://www.newclassd.com/index.php?page=70

 

[phrarod]

Cool! Any pcb solution to go with it to create the entire supply?

 

[martin clark]

Lots of people have an irrational knee-jerk affinity for shunt regulators. There's nothing that makes a shunt necessarily superior in any given application.

Amen brother.

While shunt regs look like a good bet - just like series regulators - the devil is in the details. In the case of teh shunt reg, the devil is two-fold: one needs the same level of error-amp/correction performance as in a series reg and you also need excellent, wideband current-source performance.

The latter isn't quite as easy as many think! (hint: a resistor just doesn't cut it)




(edited to fix quote tag)

 

[phrarod]

Thanks Martin. Always good to understand the +/- of these. Appreciate it.

I've been doing a load of research. The new Audio Xpress has a great article on the Super Shunt supplies although they're quite expensive. Someone recommended this one because I can use more modern and better performing regs.

http://tangentsoft.net/elec/tread/

 

[martin clark]

The TREAD is a nicely-resolved use of the LM3x& regs - which are capable of quite good performance, at low cost. Here's my take on these parts, for what it's worth:

http://www.acoustica.org.uk/t/3pin_reg_notes1.html

 

[phrarod]

Thanks Martin!

 

[jwb]

Martin, that's a very insightful article about the LM317. Have you looked at the new LT3080 regulator? It's designed have the reference pin driven be either the classic resistor, or with a zener, or a voltage reference, or even driven actively. Very cool, and achieves what you're talking about with the REF resistor on the LM317.

 

[martin clark]

No, I hadn't seen that, thanks for the heads-up

 

[jackinnj]

Peranders (Sjostrom) regulators are among the quietest I have tested -- using the AD797 or LT1028.