diamondears
Headphoneus Supremus
- Joined
- Oct 1, 2014
- Posts
- 1,867
- Likes
- 133
[COLOR=222222]Developing the mini Power Supply - to the Nth Degree[/COLOR]
[COLOR=222222]As the subject of power supplies has come up, it is perhaps a serendipitous coincidence that our technical director has worked very recently on the final version of the external power supply which will used for the “mini” range.[/COLOR]
[COLOR=222222]As the power supply section is just about full-crystallised, we are comfortable with sharing the results the mini power supply will deliver.[/COLOR]
[COLOR=222222]Obviously, the better a power supply you start with, the better you can make the power at the actual digital and audio circuitry inside the given piece of equipment. Insufficient power levels can impact dynamics, excessive noise may not become directly audible as noise, but it may raise jitter levels and distortion.[/COLOR]
[COLOR=222222]So what do we do? Make the best external power supply we can make, with the filtering, shielding and regulation needed and of course in compliance with all international safety standards as well as those for EMC.[/COLOR]
[COLOR=222222]Attached here is a screen shot from our AP2 showing the noise of the new iFi mini power supply delivering 15V at 1.25A (18.75W output power).[/COLOR]
[COLOR=222222]This graph covers frequencies from 10Hz to 108kHz and voltages from 10 nanoVolt (0.00000001V) to 10 Volt.[/COLOR]
[COLOR=222222]
[/COLOR]
[COLOR=222222]It is worth comparing this graph to the one shown in Walt Jungs seminal article which in 1995 introduced the “super-regulator” concept to the world, with his AD797 based design. Despite its age, this design has so far remained the gold standard against which other power-supplies are/should be measured.[/COLOR]
[COLOR=222222]
[/COLOR]
[COLOR=222222] - taken from this article:[/COLOR]
[COLOR=222222][COLOR=1155CC]http://waltjung.org/PDFs/Regs_for_High_Perf_Audio_2_A.pdf[/COLOR][/COLOR]
[COLOR=222222]Please note that graphs are not directly comparable, because the iFi mini powersupply includes line AC components caused by being asked to deliver a lot of power.[/COLOR]
[COLOR=222222]Walt Jung's measurements are only evaluated using the self-noise of the regulator with very light loading. However they do both show the self-noise / noisefloor of the circuits very well.[/COLOR]
[COLOR=222222]If measuring the total noise using the AP2 the reading is around 0.00007V (~70uV) for all noise below 80kHz. Or if we express the noise as a percentage of the 15V output voltage, it is 0.0005% noise compared to the output voltage. Again remember this is when delivering over 18 Watt of power, not at idle![/COLOR]
[COLOR=222222]If we push the iFi mini power supply to the maximum by drawing 3.75A (the closest setting on our test load - the powersupply incidentally is rated at 15V/3.5A) and thus delivering 56Watt of power, the noise levels rise a little to around 120uV (mainly cause by increased levels of 100/200Hz components). So for any power level within the design limits we get an output noise of better than 0.0008% of the DC output voltage below 80kHz.[/COLOR]
[COLOR=222222]It should be noted that this Supply does not directly power any audio circuit, it merely provides the “raw” supply which then passes through two more regulation stages and additional LC filtering inside the mini.[/COLOR]
[COLOR=222222]So we are confident that this performance is close to the very best power-supplies regarding noise, disregarding the actual implementation and technologies entirely, but instead simply looking at the performance delivered.[/COLOR]
Is it on schedule? What's the scheduled release btw, mid-January?