mains RF filter

[habana]

I'm thinking of making a small filter to plug into a wall socket to remove some RF noise. The sort of thing I'm looking for is a device that I can plug in and out. I plan to use a plastic power supply unit with an earth pin.

If I connect a 100nf (X2) capacitor across the LIVE/NEUTRAL and a 4700nf (Y) across the LIVE/EARTH and another across the NEUTRAL/EARTH. Would that achieve the desired effect?

I want to avoid "ringing" which I understand can occur in these types of circuits.

Thanks.

[tere]

Hi!
Don't put any capacitors between the Live & Earth! You can kill your neighbours or people you live with.
Go to the shop and buy a UPS or a simple noise reducer. This late is much more cheaper and prevents any lightning damages too.
If you want to build a correct NR, get a "fast" ferrite ring and turn few rounds of wire on it. You have to turn Live and the Neutral, don't messing with the Earth!

[Roberto Amato]

Originally posted by habana
I'm thinking of making a small filter to plug into a wall socket to remove some RF noise. The sort of thing I'm looking for is a device that I can plug in and out. I plan to use a plastic power supply unit with an earth pin.

If I connect a 100nf (X2) capacitor across the LIVE/NEUTRAL and a 4700nf (Y) across the LIVE/EARTH and another across the NEUTRAL/EARTH. Would that achieve the desired effect?

I want to avoid "ringing" which I understand can occur in these types of circuits.

Thanks.

It will. I attach a graph of the attenuation assuming an imput impedance of 50ohm and an output impedance of another 50ohm, standard values for comparing mains filters. DON'T use a 4700nf capacitor between earth and neutral/live wires, the value is MUCH TOO HIGH and potentially dangerous. You should use 4700pF not nF... which is 4,7nF. The filter won't filter very much (even if the graph say so, the 50 ohm imput impedance is normally purely theoretical) in real life conditions but won't hurt, either. Ringing, in a form or another is always present. As you see, the graph is not simply a curve, but at those frequencies does not do any harm. This filter also is almost completely transparent to common mode interference (you would need inductors to achive that), the graph is concerned only about differential mode.

ciao,
Roberto Amato

[Roberto Amato]

Sorry, I forgot to include the whole graph. In blue is depicted differential mode attenuation and in red common mode. I stated wrongly that the filter is almost transparent to common modes, which is not... just the check box I didn't check stayed transparent...

ciao,
Roberto Amato