Clean Mains Power

One of the most important issues with my hifi was the dirty mains in the building. There were obvious problems such as turning on a bathroom light, which created a brief “crackle”, and there were less obvious problems such as the high frequency noise running over the 50hz mains that polluted the gear and spoiled the definition and stereo image. I didn’t realise how significant this was until I decided to build a mains filter – and then I was truly astonished at the change in the sound quality. The stereo image was so much better and the depth and kick to bass was jaw-dropping. Really great! But then my mains was truly ghastly so ymmv. One thing I found, even changing the capacitors in the filter itself could strongly influence the sound so this is a trial and error kinda thing and a scope really helps.

The first thing I tried was disappointing – it was a basic filter (inductor-capacitor or LC) power bar. I honestly heard no difference and the bathroom light problem was still there. Next I tied a Belkin Surgemaster and was equally disappointed. I checked on my old scope and neither of them made much difference to the broadband hash of noise, so I decided to build my own – a filter and isolator – based on lots of different advice such as Jon Risch.

p1080075

The circuit is very simple, based on LC filters. The mains live connects through a 5A fuse to a 12AWG (2mm) air core coil, about 4mH, that was designed for heavy duty car subwoofer, good for over 20amps. The output from the coil is shunted to neutral through three X-rated caps in parallel, with the values in decades : 1uF, 0.1uF and 10nF. This forms an LC filter to remove symmetrical interference on the mains. The coil is large and this could affect dynamics but if a HUGE transformer is placed after it, the equipment sees the output impedance of the transformer so the large input coil isn’t an issue.

So, the output of the filter is connected to a 1KVA transformer, 220V in and out, to create an isolated output with low output impedance. The laminated core of the transformer also limits its ability to induce HF noise on its output so the transformer is a second stage filter. The output of the transformer goes into more LC filters. The first LC is a 200uH 10A toroidal coil in series with each transformer output, followed by 2 sets of 3 Y-rated caps (paralled, as above, values in decades) shunted to a mid-point from each line. The mid-point is connected to earth so this creates an LC filter on each line shunted to Earth. This is effectively the third filter and it removes asymmetric noise from the output of the transformer. Following this there is a fourth filter: 2 100uH 10A toroids, one on each line, the outputs of which are shunted to each other via 3 X-rated caps paralleled (100nF, 10nF, 1nF) to remove the last of the symmetrical noise. The end result is a very clean 50Hz (ish) wave – super clean power and MUCH better sound quality. It’s really one of those “can’t believe it” moments. It’s very simple to bypass to do back-to-back listening tests and the results are not subtle. Plus the bathroom light has no effect anymore either 🙂

p1080072

I fitted a voltmeter and switch so I could monitor the input and output voltage and turn off the voltmeter too. Thankfully, the output is slightly higher than the input, which has dropped down to 210V at times and gone up to 228V occasionally. The front also has a “bypass” front mains socket and the output of the filter is a six socket power bar.

p1080076

So if you have a scope and haven’t checked out your mains, it’s a must – if your building is as bad as mine there is a lot to be gained from a power filter.

p1080074p1080073

Jon’s page moved so I’m concerned it may disappear, so I’ve cut and paste the essentials here, just in case:
Schematic Key
______ = wire connection
o = junction where wires or leads are connected/soldered together
) = wire connection skipping over another wire, no junction
H = Hot AC line, the smaller of the two rectangular outlet holes
G = Ground, the “U” shaped hole on the AC outlet
N = Neutral, the larger of the two rectangular outlet holes
***************************************************************
AC Line Filter Schematic
jon-risch-filter
C1,2, 3 = 0.1 uF 250 VAC metallized polyester AC line rated for
continuous use across the line, which is designated as “X” type caps.
C4 = 0.01 uF 250 VAC metallized polyester AC line rated cap as above.
L1, L2 = 50 to 100 uH powdered iron rod core inductor with DCR less
than 0.03 ohms.
For AC Surge Suppressor and Filter Schematic
B1 = Circuit breaker, 5, 10 or 15 amps as appropriate, thermal type.
V1,2 = Harris Metal Oxide Varistor 150 VAC rated
C1,2,3 = 0.047 uF 250 VAC metallized polyester AC line rated for
continuous use across the line, which is designated as “X” type caps.
C4 = 0.01 uF 250 VAC metallized polyester AC line rated cap as above.
L1, L2 = 50 to 100 uH powdered iron rod core inductor with DCR less
than 0.03 ohms.

Advertisements