concept

designated power

We like Naim Audio amplifier from the 70s & 80s - foremost the "chrome bumper" series. One of their sound improving strategy were and still are designated power supplies/regulators to as many sections of a circuit as possible. 

The quality of a signal lays in the source and will then be amplified by power added to it. If that power is noisy (from its origin) or if other sections of a circuit pollute it the signal will be compromised. 

We were curious if this principle to improve musical performance would be applicable to such basic DIY machines as the raspberry pi and an I2S DAC - the whole pack normally powered by one "cellphone charger"? Even a high quality power supply will always have all different loads on its single rail.

When we build a digital music player based on a raspberry pi + a digital audio converter (don't think about using the onboard one, see 1 & 2) we have a few components which need power and which can easily be separated:

On the raspberry pi it's +5V for the main board and +5V for USB which separates the processor from it's "power hungry" peripherals (wifi dongle, usb stick or usb hard drive). 

On most DACs you can separate the +3V3 for the digital section (DVDD) from the +5V (or +3V3 after an onboard LDO) for the analog section (AVDD) – that's where you get closer to the final analog output signal.

This gives us 4 spots to upgrade the components.

So here's our main layout of bits and pieces:

* Good regulators can be expensive. One can prioritise the value they add to quality - 

being the closest to the final signal the regulator for the AVDD on the DAC is essential. Followed by the one for the DVDD (or eventual also onboard clocks) will be next....