| promitheus |
After doing some research about EMI filters I saw that there isn´t much else from ready made ones. The most famous are Schafner, Corcom, Belling Lee, Epcos and so on.
Does any one have links or info on discrete DIY filters or commercial ones that are suited for High End audio?
And what would be suited for audio ? I would like some opinions on this subject. |
|
|
| Sch3mat1c |
Something like:
1uF cap across the mains.
Feed lines through a 12:12V transformer (be sure the phase is correct so it appears as turns on a core, not just plain wire ;) )
Follow up with another 1uF.
Personally, I've noticed no difference.
Tim |
|
|
| peranders |
| quote: | Originally posted by Sch3mat1c
Something like:
1uF cap across the mains. |
No, no! You must only use approved and safe parts! Mains filter is usually for EMI filtering and to get good results they must be compact with short connections.
I also recommend that you only use parts approved for X and Y applications. These caps are special and also safe. |
|
|
| promitheus |
I think the best solution is to copy a ready filter and use the correct X and Y caps of course.
What would the best configuration be though for audio that is?
Any general purpose filter will do but is not the best. |
|
|
| peranders |
| I think you must ask yourself which frequencies you want to filter out. You can't really talk about "how it sounds". Normally filters for many amperes don't go so low in frequency because of the inductor size. The manufactures datasheets for component values. I suggest stronly that you buy a ready made filter becuase of the safety aspects. They aren't very expensive. |
|
|
| mrfeedback |
I use a mains filter that I salvaged from a scrapped old main frame computer system (originally quite expensive I expect).
This thing is rated at 20 amps or so (from memory - it is at my workplace) and measures 2.5" high X4.5" wide X 6" long.
I have no idea of the circuit values off the top of my head, but I have found it to be sonically very beneficial for both home hi-fi use, and for every of many instances of 5000W PA usage.
Without a lot of trouble it would be perfectly possible to brew your own, but it is notable perhaps that the commecially available examples are in soldered shut hermetically sealed cans.
Eric. |
|
|
| promitheus |
| Well from what I know the filter should at least start from 10KHz. But its not so easy to calculate the parts since you have balanced rejection and normal. Its not like a normal C-L-C filter because you have coils on both neutral and phase. |
|
|
| promitheus |
| this is a filter from LC audio and here is the attenuation |
|
|
| promitheus |
I think a filter for audio should start from 1 KHz to cut.
What I dont know is if a filter should have two or one coils in series. What effects more the dynamics. I would be in favour of one coil. |
|
|
| Bricolo |
| this is how my cheap 4€ filter looks like |
|
|
| peranders |
Most mains filters are for filtering noise coming from the (switching) power supply and going out to the mains, not the other way around. Frequency range is 1-100 MHz. Filter under 1 MHz and a few amperes are quite big.
The LC filter is effective above 200 kHz and the filter is rather complex. If you want to reduce EMI you have to have instruments, you can't just listen.
My advice here is to use a ready filter and if you don't know how a filter should look like or which values, you shouldn't think at all about this, sorry but I means this. A bought filter is safe, is approved and you know what you get. If you have problems with transients (washing machines, fridges etc.) use a VDR (approved of course) across the mains lines. |
|
|
| rbroer |
I was using a 10A filter with two coils in series for my cdp.
It improved things noticably.
Just recently I replaced it with an 80VA isolation transformer.
I tried combinations of the two, but just the isolation transformer without filter (before or after, big diff. btw) sounds best to my ears.
I will try some filter/motor run caps before and after later on. |
|
|
| peranders |
| quote: | Originally posted by rbroer
I was using a 10A filter .. |
You should allways use a filter rated for your current, not use too large. 1 A filter has much better effect because the inductors can be made with larger inductance. DIY'rs do believe that "bigger is better" but in this case "smaller is better", if we talk current rating.
Isn't btw a CD a typical device which GENERATES much RF noise and therefore needs a lot of ferrite beads, rings, filter etc. to meet CE approval? CD's do normally have mains filters inside tuned for best effect in the MHz region. |
|
|
| Bricolo |
| what about making a simple bandpass filter, tuned for 50 or 60Hz? |
|
|
| peranders |
| quote: | Originally posted by Bricolo
what about making a simple bandpass filter, tuned for 50 or 60Hz? |
Crazy really. Do you know how big the part would be? Better to analyze what's the real problem instead. Do you have something you want to filter out? |
|
|
| Bricolo |
| why would it be so big? |
|
|
| Elso Kwak |
Hi ,
Jon Risch was famous for his mains filter on Audio Asylum.
Do a search and you will find numerous posts.
http://www.audioasylum.com/audio/tweaks/bbs.html
Jon used rather large caps to ground and has been critised on that as his filter has the tendency to trip the Ground Fault Switch in the fuse box. Here in Holland this switch is obligatory in all new houses. Jon does find that a shame, but I am sure it saved lifes!:idea: |
|
|
| peranders |
| quote: | Originally posted by Bricolo
why would it be so big? |
How many mH or H do you need?
How must the inductor be designed so it really blocks high frequencies?
How big caps do you need?
How low impedance is required at 50 Hz?
How high impedance is needed outside the 50 Hz band?
Something to think about. |
|
|
| promitheus |
| 50 Hz is too low to filter. the parts needs to be giant, thats the reason most good filters start from 1-2 KHz. And these are big. |
|
|
| DIAR |
Here's a picture of my Jon Risch filter and DC blocker ("original" design by peranders). I used toroidal Amidon T200-26 iron powder cores but Jon Risch tells not to use toroidal or bobbin cores since there is a risk of saturation...
I also added an 1 megaohm resistor which discharges the caps. I will place the varistors right next to input.
The copper on the pcb is 105 um thick. Winding wire is 1.6 mm thick (~2 mm^2). >> total DC resistance of the filter is about 0.02 ohms. Inductor value is about 100 uH. I will place automatic fuses to the enclosure (automatic types won't add resistance).
DC blocker caps are 10000 uF and life expectancy is 250 000 hours (should be enough :D ) (RS No: 255-0169). It is possible to add second pair of diodes to DC blocker but for now I only need one pair as my DC offset is quite small.
http://kotiweb.kotiportti.fi/audiov...%20moduulit.jpg
I haven't tried these yet and in this photo the caps are placed + poles against each other. These caps have screw terminals. |
|
|
| commstech |
| would air core inductors used in speakers crossovers be suitable for this application? |
|
|
| Sch3mat1c |
More resistance, so even lower supply impedance. Fine if you don't draw too many amps and can tolerate 5V sag or so.
TIm |
|
|
| TwoSpoons |
Using air-core inductors you would likely pick up as much noise as you filter out. With the toroidal cores, as in the pic above, the magnetic path is closed (pretty much) so pickup is not a problem.
The use of iron powder cores gives a distributed air gap, raising the saturation flux of the core. Iron powder also becomes lossy at higher frequencies, making it even better for EMI suppression.
Layout is crucial with an EMI filter - there is little point in spending lots of $ on filter components, only to have the noise skip around the filter through parasitic capacitances. |
|
|
| DIAR |
| quote: | Originally posted by TwoSpoons
Layout is crucial with an EMI filter - there is little point in spending lots of $ on filter components, only to have the noise skip around the filter through parasitic capacitances. |
Interesting. Could you explain this in more detail :) |
|
|
| Tuur |
The transformer in an amplifier is already a very good low pass filter. A capacitor of 1uF in front and after the transformer will be enough. Use an X2 capacitor, special for mainvoltage use.
For any low power audio equipment you can use a isolation transformer, 220 volts in and 220 volts out. Also an 1uF cap at both sites.
I use it with good results. |
|
|
| Sch3mat1c |
| quote: | Originally posted by Tuur
The transformer in an amplifier is already a very good low pass filter. A capacitor of 1uF in front and after the transformer will be enough. Use an X2 capacitor, special for mainvoltage use.
For any low power audio equipment you can use a isolation transformer, 220 volts in and 220 volts out. Also an 1uF cap at both sites.
I use it with good results. |
:nod: :nod: And hence why I give engineering advice on how to make one, but won't ever bother recommending one.. :D
Tim |
|
|
| TwoSpoons |
| quote: | Originally posted by DIAR
Interesting. Could you explain this in more detail :) |
Experience from my days designing switchmode power supplies, and trying to get the equipment through EMI standards testing. Essentially if the inputs and outputs of the filter are too close, then the parasitic C will carry plenty on EMI across the air-space, bypassing the filter completely. In one case a simple rearrangement of the filter gave me something like 20dB improvement in the filter performance :bigeyes:
Another thing to remember is that bigger is not always better - sometimes a smaller capacitor works better, simply because its ESR and ESL are lower!
The whole business of EMI supression is something of a black art - fundamental engineering principles get you in the ball park, but after that its build, test,modify,test,modify,test,modify ....
and thats when you have a Rhode & Shwartz conduction clamp and a $50K HP Spectrum analyser sitting on your desk!
Realistically? Hack the filter out of a PC power supply :D
The comment about the transfomer forming a low-pass filter is quite valid for differential noise, less so for common-mode noise. |
|
|
| Mike Gergen |
Two Spoons
Have had the same issues with PWM controllers for treadmills.
a big issue is that a line filter is load dependent. Add this to all the previous posts and it becomes a challenge of the worst sort. This is why lab time for CE/FCC testing is $1500/day.
I've had good luck here in the states using a 220/110 transformer. Our 220 is 2 phase and then earthing one side of the secondary. It kills most common mode noise but it isn't cheap to do.
Mike |
|
|
| kilowattski |
| 1 megohm to discharge the caps sounds too high because the time constant will be very long. As an example 10K uF discharged by a 1 meg resistor will take 10000 seconds to loose only 33% of its charge. |
|
|
| DIAR |
| Filter caps are 0,01 and 0,1 uF. |
|
|
| Tuur |
| 0,68uF or 1uF 275V AC MKT X2 are available in many internet electronic part stores. |
|
|
| mastertech |
promitheus []
50 Hz is too low to filter. the parts needs to be giant, thats the reason most good filters start from 1-2 KHz. And these are big.
----------------------------------
Hi Promitheos excellent thread youve started my Friend
i dont include such filter in my amps, should I?, your input
would be appreciated
thanks |
|
|
| promitheus |
Hello there,
good this thread is coming alive again.
I found these new filters here.
They are very good quality and they shouldnt supress dynamics of power amplifiers because the series resistance of the coils is extra low.
http://www.diycable.com/main/produc...products_id=536 |
|
|
| daly41k |
| Has anyone designed a dc blocking circuit for incoming AC for 120 volts US operation? I would like to put this circuit inside my power amp. I already have RF and EMI filter. thanks, dave |
|
|
| peranders |
| quote: | Originally posted by daly41k
Has anyone designed a dc blocking circuit for incoming AC for 120 volts US operation? I would like to put this circuit inside my power amp. I already have RF and EMI filter. thanks, dave |
The filter looks exactly the same. The caps value is determined by the max continous current.
Notice also that if you don't have any problems with DC on the mains you don't have to fix it. DC on the mains creates only an accoustical problem, nothing else.
DIAR, are you awre of that inductors are usually winded on a single core, "common mode inductor". It's also essential to keep the filter compact in order to get the highest filtering properties. |
|
|
| daly41k |
thanks, Peranders,
The problem that I am having is that my 2kva torroids in my Aleph 1.2's buzz or mechanically hum. It could be the transformers themselves. They are Plitron brand and their spec sheet says that they can handle up to 250mv of dc on their primaries before making noise (on a 120 volt line). I just want to eliminate it as a concern. My amps are class A and draw about 5 amps AC from the wall outlet. Knowing this, how would I size the caps? dave |
|
|
|
