Parallel mains filter or purifier info needed !

Status
Not open for further replies.
Jump to Latest
#4
Have you heard about Auricap tweak:

http://didnt.doit.wisc.edu/audio/CDP/cdp.html

And here's mine:

http://kotiweb.kotiportti.fi/audiovideo/DIY/auricap/


The idea is to install 2 Auricap capacitors (0.47 uF) between neutral and hot. Auricap has been the best cap for the job. X2 or caps like that won't do...

I couldn't find any Auricaps in Finland so I had to settle with Arcotronics high quality (and high priced) Metallized Polypropylene capacitors.

You will find more info at http://www.audioasylum.com/audio/tweaks/bbs.html

I don't believe that a well designed filter with chokes in it will decrease dynamics. As long as components are chosen carefully and the L kept low, there won't be any problems.

I'm going to build a kick *** filter but there's still a lot of designing to do. Here's an unfinished scheme :


An externally hosted image should be here but it was not working when we last tested it.
 
#5
promitheus said:
Yeah but L´s in series with current can effect the devices connected after.
for instance dynamic loss by amplifiers.
Click to expand...
The thing with L's IS to smooth the currents! I think you are a bit wrong about the "dynamics" because this is dependant of the whole power supply.

The filters you are thinking about are made for > 1MHz but special filters can go from 100 kHz, or even lower.
 
#7
For audio equipment the filters should start cutting from 200 Hz-1000 Hz. That means the inductors should be bigger the shoe boxes. 1 mH is nothing you need something like 100mH at 10-20 Ampreres if you don´t want the filters to effect power amps.

For small devices you can use a normal filter but the problem is to design a filter for the whole system.
 
#8
Don't take me as an experienced guru, I'm totally newbie with these devices. I have been designing this filter with a few finnish high end enthusiasts.

The whole idea is to keep the inductance very low. Here's a few PSpice simulation pics:

An externally hosted image should be here but it was not working when we last tested it.


An externally hosted image should be here but it was not working when we last tested it.


An externally hosted image should be here but it was not working when we last tested it.


Can anybody give some advice how to proceed with this filter?
 
#9
#13
well actually my question was if somebody knows any filters that are connected on the mains in parallel.

These filter don´t use any parts in series with the current but in parallel with the mains. More than one LCR networks work in parallel to filter all the frequencies above 100Hz.
 
#14
Filter in parallel aren't very effective at low frequencies and also _can't_ be made effective. You must have L's also.

I think you once again should investigate your need for a filter and where the cut-off frequency should be and also have many dB's it should filter.

Don't forget that a normal PS creates harmonics and "polute" the mains, not the other way around. If you do happen to have real problems with the mains quality you maybe shouild consider a transformer with shield?

Maybe you should improve the decoupling in the amp instead?
 
#15
Konnichiwa,

promitheus said:
I have read a lot of info about normal LC mains filters and their problems. I think a better solution is to use a parallel filter across the live and neutral.
Click to expand...

You should note that the primary objective of parallel filtering is not so much the reduction of RF noise (though these are involved as well). You may still require LC Lowpass filters in addition to parallel filtering.

promitheus said:
How can that be calculated?
Click to expand...

Up to the point where the blind current makes your electricity meter spin way to fast for your comfort, as much as possible.... ;-)

promitheus said:
I have here one solution at this link:
http://www.tnt-audio.com/clinica/mains_e.html

Are there any other circuits out there?
Click to expand...

I have seen a similar principle promoted by David Magnan

http://www.magnan.com/columnprint.shtml

peranders said:
I think you are unaware of WHAT you are going to remove. L's are required when you need more effective filtering than only C's.
Click to expand...

Hmm. Are YOU aware of the actual aim of parallel/shunt "filters" on the mains? You need inductors for lowpass functionality, this is true. Have you ever calculated the leakage inductance of the local power stepdown transformers and of the entire powerline up to the Flat/House distribution Panel?

Sayonara
 
#18
Konnichiwa,

Just some more general PLC remarks.

First, the "normal" power line in the UK has an inductance of > 1mH on average, much more can be found.

Second, most powersupplies use a very narrow conduction angle due to rectification and large value input capacitors, meaning as a result very narrow, steep slope current pulses that re-charge the PSU Cap's. In transformers with a high leakage inductance and with a high primary inductance (in other words LL or EI Transformers) the local inductance reduces the slope steepness. With transformers having a low primary inductance and low leakage inductance (in other words torroidal transformers) the power line inductance dominates the feed impedance to the rectifier. (BTW, this to a large degree explains the observable sonic differences between mains transformers).

Third, if you connect a suitably large value capacitor (or RC Zobel ideally) you can cance the inductive rise of the power line impedance. Ideally rather large value capacitors in parallel to the transformer secondary winding can also reduce the impedance at high frequencies for the power source. Secondly, such capacitor(s) will also (if they have a sufficiently low ESL & ESR) operate in conjunction with the line inductance (really line and local distribution transformer leakage inductance) as lowpass, with quite good effectiveness.

Just for example, my original "simple shunt" (aka: "Thunderstone Audio Superstealth Megamanga Mains Line Ultracleaners ®") forms a lowpass with a -3db point in the 1KHz range, with a fairly low Q and the ability to have a quite low local PSU line impedance at higher frequencies, limiting the line impedance to < 6 Ohm with the higests impedace occouring around 2KHz. Now plug several of these shunt filters into free outletts and your impedances plumets, using 4 pcs as the higherst impedance < 3 Ohm at 900Hz....

By comparison, a untreated line with only a few 10nF (as found in normal LC Power Line filters which may be build into the IEC sockets) will have a very high line impedance at higher freqencies, peaking at around 90 Ohm @ 30KHz. Thus any noise pushed back by the high slope current peaks is not effectively sunk and the local disturbances will be fairly high, as there is no local reservoir that can supply these peak currents.

Now, from the above it should be also clear why filters with big L small C Combo's need to be applied with due care and consideration. Most filters expect a resistive load, but are exposed to a dynamic load with an extreme impedance range. One option is to limit at least the impedance rise by terminating such filters with a suitably value RC Zobel circuit, alternatively just getting the capacitance after the filter quite high will solve the issues the "brute force" way.

So, as with all things in audio, power line conditioning is better considered with great care and a deep enquitry into the matter.

Sayonara
 
Status
Not open for further replies.
Top Bottom