Simple question... I'm going to use a fused mains inlet from Schaffner and I just assumed the version with filter was the smart choice to reduce noise conducted into the amp.
It has since been suggested that it would be unwise to filter the mains.
What say you, learned folk? Filter or no filter?
Pops
It has since been suggested that it would be unwise to filter the mains.
What say you, learned folk? Filter or no filter?
Pops
I would vote for filter, so I would be really curious to know what the reasons are why filtering would be unwise.
My daily work is in instruments of which an amplifier is an important part of its workings and a mains filter is incorporated into the inlet. Of course the frequency range it must work in is quite a bit higher than what would ever be needed for audio. Noise is measured down into the pA ranges and it's unforgiving of noise as there's no masking like our hearing does, it all adds up.
My daily work is in instruments of which an amplifier is an important part of its workings and a mains filter is incorporated into the inlet. Of course the frequency range it must work in is quite a bit higher than what would ever be needed for audio. Noise is measured down into the pA ranges and it's unforgiving of noise as there's no masking like our hearing does, it all adds up.
Last edited:
Definitely filter for digital and line level analogue like preamps, tuners etc. Always overrate the current rating - 6A should be adequate.
Sometimes mains filters are reported to reduce dynamics with power amps. This may just be due to ferrite cored inductors in the filter saturating, so go bigger. - 10 or 15A.
In both cases it's easy to put the filter in an external box with a short lead to the item you want to test and try it.
Sometimes mains filters are reported to reduce dynamics with power amps. This may just be due to ferrite cored inductors in the filter saturating, so go bigger. - 10 or 15A.
In both cases it's easy to put the filter in an external box with a short lead to the item you want to test and try it.
Last edited:
Cause most of the time it sounds bad?
It is probably unreasonable to expect that cheap caps and ferrites may do anything good for the sound in such a responsible position, yet this is what most AC filters contain.
It is probably also a question of sound priorities, Some listeners enjoy "black backgrounds" and lack of glare, while others are annoyed by the constriction and lack of sparkle.
I find it easier to obtain pleasant sound by careful choice of power cords. More substantial forms of filtering have so far not worked for me.
The main reason for using RF filters in mains IEC sockets is to stop noise from switch mode power supplies contaminating the mains, not the other way around.
Any loss of "Dynamics" is purely in the listeners mind. Voltage drop on inductors with a very low inline resistance would give catastrophic results including smoke from overheated inductors within the IEC filter!
Any loss of "Dynamics" is purely in the listeners mind. Voltage drop on inductors with a very low inline resistance would give catastrophic results including smoke from overheated inductors within the IEC filter!
I make the 4th vote for using a mains filter.Only Analog seems to be saying no filter.
Performance of a power supply is affected by the source impedance looking back into the mains.
A good filter will not affect that source impedance. Good in this context is a standard small can IEC socket style mains filter.
The impedance that is rasied is the HF impedance and particularly the differential impedance. The common mode impedance is reduced by the filter. The caps are extra impedance routes to the enclosure, lowering the apparent source impedance.
The common mode inductor has virtually no effect to the 50Hz impedance.
BTW,
a properly assembled "in can" mains filter actually filters both ways. It filters emi coming in via the mains cable and it also filters the emi generated internally and trying to escape out via that same mains cable.
Performance of a power supply is affected by the source impedance looking back into the mains.
A good filter will not affect that source impedance. Good in this context is a standard small can IEC socket style mains filter.
The impedance that is rasied is the HF impedance and particularly the differential impedance. The common mode impedance is reduced by the filter. The caps are extra impedance routes to the enclosure, lowering the apparent source impedance.
The common mode inductor has virtually no effect to the 50Hz impedance.
BTW,
a properly assembled "in can" mains filter actually filters both ways. It filters emi coming in via the mains cable and it also filters the emi generated internally and trying to escape out via that same mains cable.
I plug everything into an external block which has graduated filters and surge protection, no problems so far.
The cable between your protected block and the input to your amplifier is susceptible to picking up airborne emi and converting that to cabled emi.
The filter needs to be attached to the enclosure of the amplifier so that emi is directed to the enclosure.
It would be OK to use two stages of emi attenuation, first at your distribution block and then at the equipment enclosure.
Last edited:
Thanks - I think I will stick with my filter
Schaffner 9290 6A double-pole version.
http://www.mouser.co.uk/pdfdocs/Schaffner_DS_FN9280_90_low_02.pdf
Schaffner 9290 6A double-pole version.
http://www.mouser.co.uk/pdfdocs/Schaffner_DS_FN9280_90_low_02.pdf
when adding components (or a filter ) its best to determine why and if you need it , rather than gathering arguments against the thing. IDK we'd never get any thing airborne (aka feature creep).
I have analog test instruments w/line power still measuring 100 uV with out those things.
I have analog test instruments w/line power still measuring 100 uV with out those things.
Last edited:
Best approach I have found so far is to filter everything that goes in or out of a box, including the mains. There is so much noise on the mains and in the air these days from WIFI, SMPS, CFLs, LEDs that it is always worth it. You could always do a full EMC test to see if you need it or not, but that gets expensive for a one off. Just my 2p
Brian
Brian
If you think about amplifier design for a moment, original designs rolled off with a 3dB point around 22kHZ. The reason is that our ears stop at around 18kHZ and not only is it a waste of power but can and does amplify HF interference. Why would anyone want to amplify 44kHZ CD noise?
I read some posts saying the author has achieved 250kHZ response. I often wonder ...
If you use appropriate frequency roll off control, interference is gone by design.
I read some posts saying the author has achieved 250kHZ response. I often wonder ...
If you use appropriate frequency roll off control, interference is gone by design.
Not sure about this John, as most modern transistors will operate in the MHz range even if the final output is limited to 20kHz. Stopping MHz getting inside the box to these transistors has to be a good idea, doesn't it?
Brian
Brian
Two objections I can made:
1) Most LC filters are designed to have some defined input and output impedances (or admitances), and if not satisfied, the behaviour of the filter will or may be undesired.
2) The common mode inductors have two winding in opposition, so the field inside the core is always zero unless a heavy loss of equilibrium is established between lines, as a return via the earth wire.
1) Most LC filters are designed to have some defined input and output impedances (or admitances), and if not satisfied, the behaviour of the filter will or may be undesired.
2) The common mode inductors have two winding in opposition, so the field inside the core is always zero unless a heavy loss of equilibrium is established between lines, as a return via the earth wire.
Indeed every little helps but if you control with NFB, the frequency range of your amplifier, even if the semiconductors are good for 50MHZ they will be capped to a gain of 1 at that sort of frequency, even less in a good design.
^this
It's always better to stop EMI /RFI at the source, rather than filtering every sensitive thingy. Conducted noise can be filtered, but unless you have proper shielding it can always be bypassed by RF.
BTW most commercial audio is designed w/out earth grounds, often adding one after the fact can cause more hum than yer looking to 'fix". (countless chip amp ground loop threads.)
+1 for filtering. I design and own a company that sells high end audio equipment. It is all analog. Filters help (key word is help) remove crud on the AC mains. We have been unable to verify any loss of quality in the sound output of any device. Then again we over build the secondary side of the power supply. I don't wish to start any arguements or disparage others thoughts, but I would not build an AC mains power gear with out a line filter.
so this makes you an expert on what exactly. what high end gear do you make and sell? IMO most here would consider 'pro audio' on design competency over any 'high end' product."We have been unable to verify any loss of quality in the sound output of any device" Anybody can do this, infact the harder of hearing can pass more tests.🙂 Id trust a measurement any day over hearing tests> before and after spectral plots.
Schaffner impedance assumptions
Schaffner specify the frequency response of the filter (<3 dB attenuation below 10 kHz rising 35 dB at frequencies above 1 MHz) for '50 Ohm/50 Ohm Sym' and '50 Ohm/50 Ohm Asym'.
Ermm... what does this mean? Is this the impedance assumed on the line side and amp side? What does sym(metrical) mean in this context?
Thanks Osvaldo.
Schaffner specify the frequency response of the filter (<3 dB attenuation below 10 kHz rising 35 dB at frequencies above 1 MHz) for '50 Ohm/50 Ohm Sym' and '50 Ohm/50 Ohm Asym'.
Ermm... what does this mean? Is this the impedance assumed on the line side and amp side? What does sym(metrical) mean in this context?
Thanks Osvaldo.