In modern society with countless laptops, phones, flat TVs and, of course, LED lighting and solar panels, we see more and more pollution on the mains grid. For many people, this is no problem at all. But it is for us – enthusiasts of hifi gear and music – an issue. Why? It’s pretty simple: your system may perform less well. Fortunately, there are solutions in the form of separate power supplies and power conditioners.
Modulating energy
If we have to explain very simply what a hi-fi system does, it is modulate energy. An amplifier takes energy ‘out of the wall’ and converts that energy to voltages and frequencies corresponding to the music. This energy goes to an amplifier that pumps it to a speaker that converts it into pressure differences in the room. That’s it.
This applies to every device to a greater or lesser extent. Also for digital devices. After all: they too use energy ‘from the wall’ to eventually send out energy that, for instance, a pre-amplifier can do something with.
So in short: a hi-fi system modulates energy. And the cleaner that energy is, the better it can do its job.
dBs? What does that say?
Before we start, we need to say something about measurements. You will see dBs in many graphs. This can be confusing. Very simply put, the number of dBs says something about signal strength.
So in the case of loudspeakers, that could be vibrations (1000 Hz with 86dB sound pressure, for example), but in a measurement of noise, it could be 10 kHz at -75dBm compared to 0dB. It is important to realise that in measurements, 0dB is often the maximum value!
Do you see the ‘m’ after dB? That stands for milli, which in turn refers to milliwatts. Similarly, there is dBu, which is used to measure voltage, separate from impedance. In the measurements of the Prism, you sometimes see these. Finally, you sometimes see dBuV (microvolt).
Measuring equipment settings
You should know that the result of a measurement depends very much on the settings. Regardless of the device. With a spectrum analyser, measurements are highly dependent on the so-called BW – BandWidth – setting. That analysis bandwidth says something about the size of the ‘sample’ an analyser makes.
You must understand that a spectrum analyser does not see the whole spectrum at once. It runs through the spectrum; we call that a sweep. During that sweep, it grabs pieces of the spectrum to analyse. You set the size – the sample – with the bandwidth. The smaller: the more precise.
We can set the bandwidth to 10,000 Hz, but also to 1 Hz or 10 Hz. So this bandwidth makes the measurement more precise and also apparently – indirectly – reduces the noise floor, because you are measuring deeper into the signal. Of course, the noise floor of the device does not go down, but it appears that way on the analyser. Be aware of that.
So why don’t we always set the analyser to 1 Hz? Very simple: that takes too much time. We often measure between 9 kHz and 1 MHz at power supplies and filters. That’s quite a spectrum to analyse. With a ‘precision’ of 1 Hz, one sweep takes an insane amount of time. Maybe 20 minutes or more. That’s why we set it to 100 Hz (on the above example, the BW is 1000 Hz and you can see the sweep time at SWT, bottom left. So about 10 seconds). 100 Hz BW at 1 MHz spectrum is accurate enough ánd reduces the sweep time to about 10 seconds. By averaging a few sweeps, we get a nice result.
Indirectly, the above also applies to FFTs – Fast Fourier Transform; a method to convert signals from, say, the time domain to frequency domain – at the Prism dScope. By increasing the FFT value – e.g. from 32,000 to 128,000 – we go deeper into the signal. This can make the noise floor appear lower, although it is not really so. So pay very close attention to the settings!
We always take whole screenshots of the screen so you can see the settings.
In the following paragraphs we have described how we measure power conditioners and power supplies. And how you can read and interpret the measurements.