The power supply is the foundation of any system. As mentioned, in fact, all you are doing is modulating energy. That energy comes from the wall outlet. The cleaner and more stable that energy supply, the better your system performs. It’s as simple as that.
Now there are quite a few brands that offer various options for external power supplies. Think Naim, ASR, Pass Labs or MSB, for example. However, those power supplies are often suitable for only certain products and maybe even a certain series. Not for your d/a converter or streamer. So they are quite pricey upgrades because they cannot be used for other pieces of gear.
But a lot of brands have jumped into that gap. Think Sbooster, Farad, IFI, Plixir, Pura and, for example, Ferrum.
These external power supplies – often DC, pay close attention to that – can be connected to all kinds of devices. Think switches, streamers, d/a converters or other devices that don’t use too much power. The impact on playback can be significant.
Measurements
For testing power supplies, we have a few devices. We test them for stability with the Rigol Electronic Load. That device can put the data into a nice graph, so you can see what happens to the power supply under load.
Some power supplies remain nicely stable, others show erratic behavior. It is normal for the voltage to drop a bit as the load increases. This has also to do with cable losses. However: it should not drop too much. Below are three examples. You can see the differences in behavior.
We also measure “noise” on both the mains side and the DC side. For this we use the LISN with the LISN-Mate (LISN stands for Line Impedance Stabilizer Network). The LISN-Mate can separate common mode and differential mode noise so that we can measure them independently.
Below is a brief explanation of the difference in common mode and differential mode noise. Common mode is more difficult to get rid of than differential mode. Also, common mode finds its way in – and out – through the ground.
Examples
As you can see here, there is quite a difference in a good power supply and a bad, cheap power supply. The first screenshot is of a high-end linear power supply (5 volts). Common mode and differential mode noise are at extremely low levels.
The second screen shot is in a lower frequency band measured via the Prism. It shows two decent linear power supplies. The major difference is that the ‘red line model’ has better suppression of the 50 Hz harmonic. The “yellow” one does show harmonics. This will also have to do with price, among other things, since the difference is a factor of 15.
The last one is a standard switching power supply. This is simply a disaster plan. The switching noise is very high and even rises toward low frequency; on the left side of the graph.
Impulse
We also test power supplies with a pulse. This shows something about the capacity as well as the speed of a power supply. This is quite a difficult test, since we have to inject a square wave into the power supply. Then we have to capture it with a scope. The layout of this test is shown below. We have a few elements of this test already built into a box so that we can build it more easily.
So we need some additional equipment for this test. Think of an amplifier (we’re using a Bryston class AB plate-amp as the amplifier), a 25 Ohm load on the power supply (to get some current going) and then a pre-amp to boost the signal a bit for the scope.
It’s good to see that each power supply handles this differently. And that some square waves are cleaner than others. That says an extraordinary amount about the quality of the power supply.
Direct influence of power supply
Here we see a streamer – the Volumio Primo – with a standard power supply and a high-end linear power supply. We measured the jitter via our Wavecrest. You can see that the influence of the power supply is significant. The jitter almost halves! And that’s only one element we measured.