[Gordy]

The lastest form I'm aware of is to use 3 - 2X caps fore and aft of the cmc's, usually 1.0, .1 and .01Uf.  As for the cmc's, I believe that the one Occam has been using with great success is the Miller 17a unit here though I've been  using the 10a and 3a versions of the Coilcraft cmc's that I stumbled across while goggling some time ago here 

Although pricey, (in a Felixy sense) I really like these extruded boxes, especially when building two circuits in series as I'm prone to do here

I just picked up one of these jobbies to install in Natureboy's Felixes to better fit his amp/cable system... here

HTH!

[JoshK]

DOS,

Paul gave me a pair of CMCs to play with.  We haven't experimented with size (uH) to know what works best, or even what value caps work best. It looks as though the Adept uses a plethora of varying sized caps. Who knows what size cmc they used.  We need to experiment, hence the word "project". 

Here is the schematic:  Mains in > fuse > [ x2 cap > cmc > more caps > outlet ]   put as many of the [ ]'s that you want outlets in parrallel.

Hi It has been sime time now. Is it still in this "project phase?

I think you have too much faith in my ability to get things done.  I am just not that tweaky by nature, so I am not going to substitute a lot of different cap values and listen for what is best.  I'd much rather have some measure equipment that would allow me to look at the Q of the filter, among other things and play around with values until I got a few different combos that looked good and listen to the differences of the few configurations to see what looks best to garner some relationship to how sound and that measurement relate and proceed from there.  But I don't have such test equipment, so I used some cascade of cap values that seemed to make some sense, based on recommendations Paul made and through it together.  The results were pretty positive so I stuck with it.

Now ideally you will need to experiment with cap values for each piece of equipment you have and find an optimal solution for that component.  This is because the filters are dependent on current, the load, etc.  I don't think I have that much patience.  When I finish building some of the components I go going, I might be more willing to setup a bunch of parrallel Felix's in my PC box and wire a multi deck switch to switch between and compare, but I kinda doubt that will be in this year or next.

I built a Felix into my SB PSU and a Felix into my Aikido PSU but have done no optimization, nor any a/b listening.  It is just on faith from prior experience that I believe it works well.  But I have bigger fish to fry in my system for now....

Kyrill....I'd suggest you pick up appropriate caps (like a cascade of .47, .047, .01, or 1, .1, .01 for both before and after) and some of the mentioned chokes and build it.  I think you'll really like the "generic" Felix and it just may inspire you to experiment yourself.  That would really be what is needed, as we can experiment for our case, but not for yours. 

[kyrill]

hi
Why so "modest" with the cap values? If I interpret next image well enough may i assume the higher the value the lower the freq will be passed on by the capacitor? According to graph even with 2,2uF it effects "only" frequencies above 500khz. So all dirty frqs lower than that are not affected by this cap and are still on the power line. Values of.47uF only effects 1.3MHz and higher..? Or i likely misinterpret the figure..



Not possible to clean a broadband spectrum starting 1khz and up?

[JoshK]

Then you start introducing phase shifts down into a freq that may affect audio frequencies. Its all a balancing act as I understand it.

[Folsom]

Was it somewhere on AudioCircle someone posted the amount of power usage from having bypassed capacitors, because they always use some? This might be a reason why when you build your own equipment it could be wise to add a Felix inside of it so the power switch is inline before it is. Unfortunately that is not always convenient  for everyone, including myself, as the only things I have made on my own might require bigger boxes.

Also how do you think arranging different capacitors would make sense? Would you want some 1.0uf big guys on the AC in side of the CMCs? Then have say some .01uf on the sockets themselves? What kind of noise is likely to get past the CMCs, or induced by them?

I looked at some PDFs of some CMCs and it to me it would just make sense to target everything with some bigger 1.0uf or even more.

[Bill A]

Kyrill

I think what the plot shows is the self resonance of that style capacitor in different sizes.  The capacitor model being a series RLC circuit.  Below resonance the capacitor is "capacitive", at resonance where XC=XL, it is just resistive, and above, it is no longer capacitive but inductive.

Use the capacitor at frequencies below the resonant dip.

Bill A

[kyrill]

o-o Bill now you 've caught me

suppose i take the 1 uF cap  seeing the graph, where can i see or determine from its "resonance dip"?
and which frq range it is useful for ( according to the graph?)

[Folsom]

Kyrill

I think what the plot shows is the self resonance of that style capacitor in different sizes.  The capacitor model being a series RLC circuit.  Below resonance the capacitor is "capacitive", at resonance where XC=XL, it is just resistive, and above, it is no longer capacitive but inductive.

Use the capacitor at frequencies below the resonant dip.

Bill A

That would mean I would need .01 and larger if I was using the Q-4018-A right, and say .015 or larger for the Q-4007-A correct? Would it be best to go just barely under to get the most benefit?


http://www.coilcraft.com/comoco.cfm
http://www.coilcraft.com/pdfs/comoco.pdf

[Bill A]

Kyrill

It looks like the 1uf cap resonance dip occurs at 1Mhz so it should be useful below that.  How close to 1Mhz can the capacitor operate and still have useful capacitance? I don't know the answer to that. 

DoS

Power line filtering is new to me so I really don't have a recomendation for cap values to use.  This design guide http://www.coilcraft.com/pdfs/doc191_CMFiltDesign.pdf may give you some help.  In the example they used 50 ohms as an impedance value.  I don't have a clue about practical impedance values, but 50 ohms seems high to me.

One thing I did notice in this guide is the Q of the filter and the possibility of actually emphasizing a particular frequency near the cutoff when the filter had a high Q.  Not suprising, but that is something that may not have been thought about and could cause the filter to ring or oscillate. 

I would just use values recommended by others and maybe do some before and after measurements with an o'scope, if you have one, to see what effect you are getting.

[PromitheusAudio]

Just like to add some input to your project
CMC common mode choke are made from 2 opposite winding wires wound on a toroid core or something like that
Different cores would have different frequency range that it can handle

So for the most effective way to deal thru out the audio band is to get multiple CMC but make sure they are different core material

Iron core powder or something along this line made by roxbury is good from 50hz to 440hz as per their data sheet. I don't think there is much material that covers 500hz to 5Khz. I may not be aware of it but above the the 5khz there is alot of materials that can be found. Ferrites, permalloy and etc. You can check their data sheet of the material of the ferrite and check their impedance to see which zone they are filtering in

Perhaps one way to do filtering via CMC in the audio range is to wind it on a double bobbin and filled up with EI core material or C-core using thin lamination like z11 or m6 minimum. I know the Z11/m6 would be able to work well in the audio range

Best way to experiment for best filtering is to wind your own CMC choke. Its a no brainer anyway. The wires need to be Insulated to 180 Celcius or double insulated type, wind them in opposite directions and thats all. Find surpluss core material for various frequency filtering

[Occam]

...
So for the most effective way to deal thru out the audio band is to get multiple CMC but make sure they are different core material

Iron core powder or something along this line made by roxbury is good from 50hz to 440hz as per their data sheet. I don't think there is much material that covers 500hz to 5Khz. I may not be aware of it but above the the 5khz there is alot of materials that can be found. Ferrites, permalloy and etc. You can check their data sheet of the material of the ferrite and check their impedance to see which zone they are filtering in

Perhaps one way to do filtering via CMC in the audio range is to wind it on a double bobbin and filled up with EI core material or C-core using thin lamination like z11 or m6 minimum. I know the Z11/m6 would be able to work well in the audio range

Best way to experiment for best filtering is to wind your own CMC choke. Its a no brainer anyway. The wires need to be Insulated to 180 Celcius or double insulated type, wind them in opposite directions and thats all. Find surpluss core material for various frequency filtering

Ummmm.... you might actually consider reading the PDF specifications of the CMCs already mentioned.
Winding your own CMCs certainly is a no brainer, but not in the sense I assume you mean.  The chokes discussed are <$10ea. and perform quite admirably. I don't understand why you find it necessary to obfuscate a straightforward build, with a method I can only describe as throwing crap up against a wall to see what sticks. Certainly, multistage filters are an excellent method for filtering. But they are obviously more complicated that a single inductive element filter. Designing a filter requires an understanding of core characteristics, filter calculations, specifically, how the ratio of inductance (both common and differential) to capacitance, along with load, effect the poles, and actual shape of the attenuation characteristic.

A Felix can be built for less than $30, all in, and is quite competitive with many audiophile filters. It can certainly be improved upon, but typically not by the methodology you advocate.

When and if you build that c-core based, laminated stack CMC(s) based filter, I'm sure we'd all be interested in seeing some hard data. Its going to usefully filter from line frequency on up? This I gotta see.....

FWIW

[StuntHunt]

As a newbie, I offer a most gracious thank you to all of you tech meisters who are so free with your time and advice!  I've been reading posts for some time, but I finally became a member today.  

My cousin, BobM, is known to some of you, and he's been giving me quite a bit of encouragement and coaching in the fine art of DIY tweaking.  First Belden 89259 cables/interconnects, and then some speaker crossover mods to put some snap into my Kef Ref 4s (Thanks Bob).  

I'll be building one (or more) of these Felix PCs in the near future, and I'm looking forward to that and more. This is all great info. This has probably been covered elsewhere, but can this Felix design be adapted to handle higher currents (like for a 200 W/ch power amp)? Thanks again guys.  

[Gordy]

Can this Felix design be adapted to handle higher currents (like for a 200 W/ch power amp)? Thanks again guys. 

Welcome to AC ShuntHunt!  I've 10a Coilcraft Felix's built into each of my monoblock amps as my transformers pull about 5a max each.  The size of your ps will determine whether you need to go to a 17a Miller or not but, if it's a stereo 200w amp, it would probably be best.
 
About a year ago I built a rig using 3 - 4pdt NKK switches to compare two 10a Felix's verses no Felix and to compare 5 sizes of caps in different configurations from 3 manufacturers, Vishay,  PBB and Epcos.  The cap sizes were 2.2, 1.0, .47, .1 and .01Uf.  This was all set up so I could seamlessly switch caps/brands from my listening chair.  Both a bridged 3886 chip amp and a Modwright 9.0 preamp were being feed by the Felix's.  After maybe three weeks of swapping caps out I decided that no particular combo really stood out as better to my ears and in my location.  Different but, not better.  I concluded, perhaps wrongly, that the coils were doing 90-95% of the filtering and the caps were doing the supplemental polishing and I've built my filters accordingly ever since, for the most part .47, .1 and .01 chosen for practicality only.

I decided not to use the PBB's because of practical reasons as well, the .47 cap was easily twice the size of the Vishay 2.2 cap, that simply didn't fit with my plans for building Felix's into chassis's and way too small boxes   

A month later my girlfriend decided she simply had to have her forced air negative ion generator in her living room.  About 7 minutes later I got a phone call complaining about the static noise coming from her speakers.  It was bad!  I'm just guessing here but, if the music was playing at 75db, the static was only 10dB below it and in the 200-400hz range?  I installed a 10a Felix into the system (a 45w chip amp, tuner  and dvdp) and that seemed to knock down the noise by about 20-30dB or so but it was still there.  I went home and grabbed another Felix and installed it in series, the static was then barely perceptible.  Interestingly, when I moved both Felix's to feed the neg. ion generator it did an even better job.  Again, I've been building most of my Felix's as doubles ever since.  The concept was kinda proven also when I got a look at how the Isotek Nova units were built, basically a double Felix feeding a series of double Felix's each feeding an individual receptacle.

Doesn't virtually every DVDP have a SMPS in it as well?   

[randytsuch]

Can this Felix design be adapted to handle higher currents (like for a 200 W/ch power amp)? Thanks again guys. 

Welcome to AC ShuntHunt!  I've 10a Coilcraft Felix's built into each of my monoblock amps as my transformers pull about 5a max each.  The size of your ps will determine whether you need to go to a 17a Miller or not but, if it's a stereo 200w amp, it would probably be best.
 
About a year ago I built a rig using 3 - 4pdt NKK switches to compare two 10a Felix's verses no Felix and to compare 5 sizes of caps in different configurations from 3 manufacturers, Vishay,  PBB and Epcos.  The cap sizes were 2.2, 1.0, .47, .1 and .01Uf.  This was all set up so I could seamlessly switch caps/brands from my listening chair.  Both a bridged 3886 chip amp and a Modwright 9.0 preamp were being feed by the Felix's.  After maybe three weeks of swapping caps out I decided that no particular combo really stood out as better to my ears and in my location.  Different but, not better.  I concluded, perhaps wrongly, that the coils were doing 90-95% of the filtering and the caps were doing the supplemental polishing and I've built my filters accordingly ever since, for the most part .47, .1 and .01 chosen for practicality only.

I decided not to use the PBB's because of practical reasons as well, the .47 cap was easily twice the size of the Vishay 2.2 cap, that simply didn't fit with my plans for building Felix's into chassis's and way too small boxes   

A month later my girlfriend decided she simply had to have her forced air negative ion generator in her living room.  About 7 minutes later I got a phone call complaining about the static noise coming from her speakers.  It was bad!  I'm just guessing here but, if the music was playing at 75db, the static was only 10dB below it and in the 200-400hz range?  I installed a 10a Felix into the system (a 45w chip amp, tuner  and dvdp) and that seemed to knock down the noise by about 20-30dB or so but it was still there.  I went home and grabbed another Felix and installed it in series, the static was then barely perceptible.  Interestingly, when I moved both Felix's to feed the neg. ion generator it did an even better job.  Again, I've been building most of my Felix's as doubles ever since.  The concept was kinda proven also when I got a look at how the Isotek Nova units were built, basically a double Felix feeding a series of double Felix's each feeding an individual receptacle.

Doesn't virtually every DVDP have a SMPS in it as well?   

Hey Gordy
So, for these double Felix's, do you make them
C network - CMC - C network - CMC - C network

Based on the rest of your post, this might be overkill with caps, so you could also remove the middle C network, but then it would seem like you are just using a CMC that is twice as big.

BTW, I had been working on a squeezebox project, stealing much from Mgalusha projects.

I was planning to put a Felicia in it, put am now wondering if a Felix would be as good.  It would make it much lighter, and save some space inside.  I have enough room for the Felicia, but it is getting a little tight in there because I have a bunch of other stuff.

Randy

[Gordy]

So, for these double Felix's, do you make them
C network - CMC - C network - CMC - C network

Hi Randy, yes I'm using all 3 c networks, though in one case I only used the .47 in the middle position, that being Natureboy/Roger's original Felix.  Some time ago we did a comparo with his front end equipment using the double 10a Felix for his dvdp, preamp and phono amp vs. three individual Felicia's and Roger prefered the Felicia's ever slightly softer presentation to the Felix's more truthful presentation (my words!) where I felt just the opposite. 

Are you using a linear supply for the SB?  It'd be very cool if you could try both before committing!

[Folsom]

Ok so a couple of questions...

The changes after the CMC, what is more likely to occur, higher or lower frequencies? It seems to me their should be a science behind all of this...

My thoughts go like this from what I have seen, and that is introduce a 1.0-1.5uf (for 10 amp CMC) right off of your inlet, perhaps a breaker if you have one in between. It would seem overkill to have one for each CMC to me. Then run to the CMC's, I was only thinking 1 per socket at 10amp should be enough unless you know your stuff is going to take more, in which case you are going to need to route power special from your fuse box, and use special 20amp receptacles and plugs. I figured 10amp was sort of a safe way to go, as you might have per say like a couple of .5ma-3amp devices and one larger one. I would still thinking running a special dedicated line would be wisest, and if you know you are going to drain 20amps you need the special turned blade receptacle and plugs. The have after the cmc some direct capacitors across the receptacles you are using themselves. The thing with them is if I am not mistaken you have room for wires AND capacitors in separate holes (the secondary one for capacitors might require a little copper wire to be pushed over the capacitor leads, or solder the capacitor leads to them and stick them in the socket, depending on the capacitor of course). Now my debate would be what type of capacitor to use after the CMC. You have already got rid of a lot with the 1.0-1.5uf, what possible introduction or things after the CMC would arise, should tell us what size to use weather it be .47 or .1uf, whatever.

If you have too many capacitors, like plethoras of them, you will eat a lot of power even when nothing is on. I see no reason to waste any if you can help it. I think it might even be appropriate to use a switch to turn off things that do not need any power when off, unlike a Squeezebox. You could use the breaker but you might need say one bypass socket?

Another questions, breaker or fuse/s?

[randytsuch]

Hi Randy, yes I'm using all 3 c networks, though in one case I only used the .47 in the middle position, that being Natureboy/Roger's original Felix.  Some time ago we did a comparo with his front end equipment using the double 10a Felix for his dvdp, preamp and phono amp vs. three individual Felicia's and Roger prefered the Felicia's ever slightly softer presentation to the Felix's more truthful presentation (my words!) where I felt just the opposite. 

Are you using a linear supply for the SB?  It'd be very cool if you could try both before committing!

Hi Gordy,
So, as with many things audio, comes down to preferences.
Although I would like to do a comparision, I know myself better than to commit to one.  This project is taking long enough without adding this to it, think I started late fall  .
The problem is, I am putting the SB, along with a burson buffer, in this case.  I was also going to put in my Twisted pair DAC in it, as I eventually want to go balanced.

So, I will have 4 regulators and a transformer just for the bursons, then another huge, surplus transformer for everything else.  I am going to generate 5VDC to drive the display, another 5VDC to drive the SB.  Also going to generate a 5VDC and 3.3 VDC to power the DAC in the SB.  Most of these will be "teddy" regulators, expect the 3.3 VDC for the DAC, it's a shunt reg.

Problem is, if I put in the Felicia, I have to allocate enough space for it, where the Felix takes much less space.  I am also installing an internal divider, to seperate the magnetics from everything else.  Am thinking of lining the divider with some kind of mu metal, or steel/iron, something that will keep the magnetic fields out of the rest of the box.

End of thread hijacking.

DOS,
Yes, dedicated line is the way to go, but it can be expensive.  In most case, I think they will have to break into walls to add them, so then you have to drywall and paint after.

As for adding caps to outlets (which I think is what you are talking about), there are issues, like insurance problems if anything happens.  I am lucky enough to have a nice house, so I am worried about potential problems if anything bad happens.  So, I have decided not to do anything that involves outlets, other than changing them out for better ones.  To me, the price of a few caps to add to each component is cheap for the piece of mind, but everyone is under different circumstances.  But, it is something to consider, when diy'ing that involves AC.

At some point, I want to add a circuit breaker with AFCI (arc fault protection) into my audio power, I already have a GFI outlet in series with it.  Probably won't do the sound any good, but it helps my piece of mind.

End of 2nd thread jack. 

Randy

[Folsom]

Actually I am talking about capacitors inside of the box that has AC receptacles that you are using as a "Felix box". Basically it is a power filter box...

Dedicated lines can be run in cord form, instead of through wall, if need be/allowed (domestic issues).

X2 capacitors are suppose to be within law tolerances to work in AC situations like with a "Felix box" I think is the point is it not?

[samplesj]

Dedicated lines can be run in cord form, instead of through wall, if need be/allowed (domestic issues).

What in the world are you talking about?

Do you mean something on the wall in armored cable (like you see in a basement/garage)?

Or do you mean extension cords?  Do you realize the gauge you'd need for any decent length 20amp extension cord?


Neither solution is that pretty.  As far as sheetrock cutting goes, just do it.  After the first few cuts/patches you get the hang of it.  Sometimes you even get lucky enough to not need to patch/repaint anything (if you've got a crawlspace/basement).

[Occam]

......
If you have too many capacitors, like plethoras of them, you will eat a lot of power even when nothing is on. I see no reason to waste any if you can help it. I think it might even be appropriate to use a switch to turn off things that do not need any power when off, unlike a Squeezebox...

Nope. The reactive component (capacitive and inductive) of impedance doesn't work that way. It doesn't dissipate energy. The resistive component (esr in caps and inductors) and resistors themselves do dissipate energy in the form of heat. If your across the line caps, X type or motor runs, are getting really hot, you're using the wrong caps. Some power conditioners use > 60uf comprised or >100 caps across the line, and don't get warm. Google and Wikipedia are a good place to explore this. Better yet, a text on basic analog electronics from your own college's library.

Another questions, breaker or fuse/s?

Yes.   I don't mean to be obtuse, and while a magneto-hydraulic breaker, as used in some of the more bespoke conditioners, might be ideal, they're quite expensive >$60ea. A thermal breaker or fuse will both work well.