Review English Electric EE1 – Network Filter

Measurements network filters

The measurement setup for this little filter is relatively simple. We grabbed a very dirty PoE switch: the GS108TP from Netgear. This was fed from the LISN to keep the grounding equal in the measurement setup.

The data network is decoupled via the blue box: a CDN – Coupling / Decoupling network. (Conpower CDN T8). We measure each filter by including it in the network.

The noise is measured from the CDN which has a BNC connector. We connect this to a splitter from Tekbox that can split common mode and differential mode. From this splitter, two cables go into the PicoScope 5000. There we measure between 10 KHz and 2 MHz.

Baseline

These are measurements of the GS108TP PoE switch, directly at the measurement setup. So no filters. Above – blue – is common mode noise. Red is differential mode noise. We also did a measurement where we hooked up the GS108TP to the “input” of the CDN and measured the Volumio Primo. That way we can see the effect on the streamer.

In both cases, it is – to put it mildly – not great. But that is also the intention, because we want to see how the filters do their job.

English Electric EE1

Do these filters do anything? Well… check yourself. We see in common mode a reduction of 17dB on the peaks and about 7dB across the noise floor. That’s unquestionably very decent.

Differential mode is a similar story, although it reduces the peaks a bit less: 10dB or so. But that it has an effect is obvious.

‘Behind the Volumio’ the effect is also clear. A particularly large number of peaks disappear. On average, it’s just a much more calm picture.

Medical Grade

As you can see, the measurements are fairly similar. It attenuates a little more than the English Electric. Both common mode and differential mode. Although we see a flatter noise floor in the higher spectrum with the EE1 in differential mode.

Looking at the measurements on the Volumio, we see a slightly different result. There, the EE1 does slightly better when it comes to overall noise floor. The peaks are a bit lower on the medical grade, though.

Pink Faun

Surely the Pink Fain measures differently from the EE1. We see at common mode a big hump at the beginning. The EE1 is flat there. This does not directly affect the peaks. The average noise floor is very low: even lower than the medical grade version.

The result on the Volumio is almost the same. These are nuances.

Delock

It’s striking how similar the Delock measures up to the Pink Faun. We really had to check a few times to make sure we didn’t open the wrong file. So the story is similar to the Pink Faun.

GS108T

Then a switch. How well does a standard network switch isolate? Well: pretty decent, of course. But don’t expect miracles when a switchmode power supply is attached to it, either. Because, as we know, that too is noisy. Just like the active components in the switch itself. We now see a lot of ‘dirt’ in the same place as the “noise”. After that, it is very quiet.

Differential mode is a lot better attenuated than with a regular isolator. We honestly didn’t expect that.

Going to the result with the Volumio streamer in the chain, the picture is fairly similar but we still see a slightly less clean result than with the isolator; there are a bit more “sticks” in the spectrum. That must be the power supply having an influence. Or perhaps some active components on the circuit board of the switch itself.

Daisy Chain

Now what happens when you put a couple of insulators in series? Well … that effect varies. It can make an improvement, but it can also throw the whole thing off.

For example, the Delock with the EE1 does not work at all. It just gives more noise. The Pink Faun with EE1 works fine. That gives a little more calmness. Everything in sequence also works. But not sonically … :-).