[BradJudy]

I was thinking that the Felix is a super-simple PCB idea.  I'm not sure if the small volume would bring the costs down enough to make it worthwhile, but it would make mounting easy.  Here's a quick first pass at one using ExpressPCB:



This one is based on the Vishay 1778 X2 caps and the Coilcraft Combination CMC, but I just did a revision that also has pads for the JW Miller 8109-8121. 

Looks like costs would be ~$10 each for 20+ from ExpressPCB. 

[tanchiro58]

I was thinking that the Felix is a super-simple PCB idea.  I'm not sure if the small volume would bring the costs down enough to make it worthwhile, but it would make mounting easy.  Here's a quick first pass at one using ExpressPCB:



This one is based on the Vishay 1778 X2 caps and the Coilcraft Combination CMC, but I just did a revision that also has pads for the JW Miller 8109-8121. 

Looks like costs would be ~$10 each for 20+ from ExpressPCB. 

Hi Brad,

Why don't you start a group buy? I guess there will be a lot of AC member would like to purchase it too.

Would you count me in with eight( Felix mounted PCBs? Thanks.

Tan

[kyrill]

i would order 10
and silverplated would be nice
can those paths carry 2A?

doesn't it need 4 holes in the corner? That would help me
thx for the initiative

[BRN]

That is a great idea. I built a box with three Felixs in it and a board like this would make the project so easy to build. I just got some Coilcraft Combination CMC and could use these boards. I could use at least 6 boards.

[Gordy]

Very nice!  In a perfect world the .01Uf caps would be located as close as possible to the cmc's though, on the underside perhaps?

Gordy

[BradJudy]

i would order 10
and silverplated would be nice
can those paths carry 2A?

doesn't it need 4 holes in the corner? That would help me

The ExpressPCB guidance says a 0.150" trace will handle 6A and these are 0.250" traces, so they should handle ~10A.  Still not appropriate for a full 15-20A implementation, but it's the largest trace they will do.

Yes, I need to add mounting holes to the design. 

[BradJudy]

Very nice!  In a perfect world the .01Uf caps would be located as close as possible to the cmc's though, on the underside perhaps?

I can certainly start by reversing the order of the caps so the 0.01uF's are closest to the CMC. 

[BradJudy]

It also occurred to me that if implementing a multi-outlet box is a common approach, a second input can be added to the H and N lines for daisy-chaining.  Are many expected to implement a multi-outlet box and want daisy-chaining?  Of course, this might be too much current for the bridge trace on the first board (sum of the current of the rest of the boards). 

[kyrill]

i dont use daisy chaining
i think dedicated seperated lines are better sounding

[mgalusha]

Nice Brad. You can actually get both the .01 and .1 caps under the edge of the CMC, at least with the small 3A model. I built a second one yesterday and took a photo of this. I'll post it when I get home and download it from the camera.

I'd likely take some of these as well.

[JoshK]

Just an observation...

Can you spot the two other Felix's?



Different configuration for different purposes, but same topology and same idea.  Just more inspiration for how to effectively use the Felix concept in DIY for filtering. 

Here we are filtering DC and the big guy is meant to filter much lower in frequency (120hz). 

Credit:  This is Thoersten's DRD amp supply schematic. 

[BradJudy]

FYI: I realized this morning that my spacing was all messed up for the CMC and the 1.0uF cap, so I'll have a new one posted tonight.  The lead spacing was correct, but I didn't leave enough room for the part itself - classic rookie PCB design mistake. 

Some questions for the group:

To what degree should other sizes of caps be accommodated?  The current layout handles Vishay 1.0, 0.1 and 0.01 uF caps.  To a certain degree other values in the Vishay line will fit, but some will not.  For example, I don't think the 0.47uF one fits in the current largest cap position.  Is there another major brand of X2 that needs to be accommodated? (I looked at Wima, but no one seems to stock their X2's). 

It looks like I can fit pads for both the CoilCraft combo and JW Miller 8100 series, which seem to be the two popular options.  I'm willing to say that's sufficiently flexible - agreed?  CoilCraft also makes an equivalent to the JW 8100, but I haven't seen anyone use it here.

Ground line: is it needed or does it make the board wider for no reason?  Having a straight ground trace makes installs look cleaner, but doesn't serve a functional purpose.  Should it stay?

Optional fuse mount - Is it worth squeezing in a mount point for an option fuse on the board?  If so, what are your preferences on type?  I haven't tried this yet, so I don't know if it will be a pain to fit in.

Particular board size - should the design be made to a particular board size to allow using a particular case with rails (e.g. extruded aluminum options like the Hammond 1455)? 

My prior question: Is an additional step of input side points for H and N worth doing to accommodate daisy-chaining?  Or does this just set up the design to have current overload issues.

[kyrill]

FYI: I realized this morning that my spacing was all messed up for the CMC and the 1.0uF cap, so I'll have a new one posted tonight.  The lead spacing was correct, but I didn't leave enough room for the part itself - classic rookie PCB design mistake. 

Some questions for the group:

To what degree should other sizes of caps be accommodated?  The current layout handles Vishay 1.0, 0.1 and 0.01 uF caps.  To a certain degree other values in the Vishay line will fit, but some will not.  For example, I don't think the 0.47uF one fits in the current largest cap position.  Is there another major brand of X2 that needs to be accommodated? (I looked at Wima, but no one seems to stock their X2's). 

It looks like I can fit pads for both the CoilCraft combo and JW Miller 8100 series, which seem to be the two popular options.  I'm willing to say that's sufficiently flexible - agreed?  CoilCraft also makes an equivalent to the JW 8100, but I haven't seen anyone use it here.

Ground line: is it needed or does it make the board wider for no reason?  Having a straight ground trace makes installs look cleaner, but doesn't serve a functional purpose.  Should it stay?

Optional fuse mount - Is it worth squeezing in a mount point for an option fuse on the board?  If so, what are your preferences on type?  I haven't tried this yet, so I don't know if it will be a pain to fit in. I would be very happy with this

Particular board size - should the design be made to a particular board size to allow using a particular case with rails (e.g. extruded aluminum options like the Hammond 1455)?  yes this would be nice (for me) if it will perfectly fit a standard shielded case

My prior question: Is an additional step of input side points for H and N worth doing to accommodate daisy-chaining?  Or does this just set up the design to have current overload issues.

[tanchiro58]

Hi Brad,

I built the first Felix with soldering the Coilcraft and Vishay 2x (1.0, 0.1 and 0.01) film caps on a small pcb then I screwed them down to the aluminum box. The fuse is attached to the box. It is a hard work in the beginning but it was fun and happy when I got an excellent result.  It works better than the Equitech PC I own before.

Tan

[JoshK]

My $.02...fwiw and all that...

To what degree should other sizes of caps be accommodated?  The current layout handles Vishay 1.0, 0.1 and 0.01 uF caps.  To a certain degree other values in the Vishay line will fit, but some will not.  For example, I don't think the 0.47uF one fits in the current largest cap position.  Is there another major brand of X2 that needs to be accommodated? (I looked at Wima, but no one seems to stock their X2's). 

Depends on how easy it is.  Otherwise stick with the Vishay, or you could provide one extra set of pads to accomodate the .47 & 1.0 options.  I wouldn't go crazy unless it is easy.  Broskie on his Aikido boards provides a whole series of cap lead pads to accomodate a huge range of sizes.  I don't know if this is hard/easy and whether it changes the price so I'll leave it at that.

It looks like I can fit pads for both the CoilCraft combo and JW Miller 8100 series, which seem to be the two popular options.  I'm willing to say that's sufficiently flexible - agreed? 

Agreed.

Ground line: is it needed or does it make the board wider for no reason?  Having a straight ground trace makes installs look cleaner, but doesn't serve a functional purpose.  Should it stay?

I don't really see the need for it.  Plus, I have to imagine it is a tiny bit safer without, since you aren't likely to be as current limited with wiring.

Optional fuse mount - Is it worth squeezing in a mount point for an option fuse on the board?  If so, what are your preferences on type?  I haven't tried this yet, so I don't know if it will be a pain to fit in.

That would be a really neat thing to have on the output side of the board.  Would people need the input side as well?  I tend to use appropriate size breakers instead of fuses on the input side.

Particular board size - should the design be made to a particular board size to allow using a particular case with rails (e.g. extruded aluminum options like the Hammond 1455)?

 
I tend to fit mine into my gear directly, but I may add into a PC case I have instead. So I am not the target.  However, it would seem to be nice for those using the LMB or Hammond extrudeds.

My prior question: Is an additional step of input side points for H and N worth doing to accommodate daisy-chaining?  Or does this just set up the design to have current overload issues.

I probably wouldn't for current overload issues.  Makes people have to run seperate wiring for each which is likely safer.

[BobM]

Optional fuse mount - Is it worth squeezing in a mount point for an option fuse on the board?  If so, what are your preferences on type?  I haven't tried this yet, so I don't know if it will be a pain to fit in.

That would be a really neat thing to have on the output side of the board.  Would people need the input side as well?  I tend to use appropriate size breakers instead of fuses on the input side.

There are several ways to locate the fuses (1)  in line with each gang of caps (protecting the circuit if they fail), thereby keeping the fuse out of the primary line, or (2) on the output side of the board on each of the + and - lines, thereby having the fuses directly in the primary line. Of course using option (1) means a board change - ganging one end of the caps together into the fuse then out of the fuse to the primary line.

Not sure if my feeble attempt at describing this is clear.

Bob

[Gordy]

CoilCraft also makes an equivalent to the JW 8100, but I haven't seen anyone use it here

Paul and I both looked at the standard Coilcraft cmc's 6-8 months back and didn't care for the quality of the windings, the Miller's are definitely of a higher build quality.  I never actualy tried one though, not sure if Paul did or not. 

[tanchiro58]

...the Miller's are definitely of a higher build quality.

Gordy,

Yes it is high quality since it is toroidal transformer with much less noise, is that right?

I just take a look on Newark website. Which models of JW Miller 8100 did you and Paul choose? If you chose the Miller 8120 17A would that be accepted to the application for a tube preamp and dac and/or tube amp (SET 2A3 amp) or I have to get different 8100 model? Thanks for help.

http://www.newark.com/jsp/search/productdetail.jsp?SKU=63K4088&CMP=AFC-TL10000001

Tan

[Gordy]

Hi Tan,

I was comparing build quality of the Miller vs. the standard CC only, I've never compared them subjectively.  Once I tried the 3717/4818's I stayed with them. 

The Miller 8120's should be very good for filtering anything your 15a wall receptacle is handling, though you'll do even better using them for the amp(s) and using the 3717a's for your dac, preamp etc.  The 8120 is what Paul used in the Felix's he showed/tested with such good results at the various N.Y. Raves last year I believe.  I like using a 10a 4018 Coilcraft in each of my ~160w monoblock amps and 3717's for everything else.  Next up is a Felix for that nasty computer smps. HTH!

[BradJudy]

Here's a second pass.  It has the pads for the JW 8100 series (two sizes) and an additional pad for the 0.47uF Vishay.  The 0.047uF is the same size as the 0.1uF, so it already fits. 

I added a temporary silkscreen that shows the footprint of the Coilcraft combo CMC as well as the footprint of the 1.0uF Vishay.  I also resized the image to be about lifesize (on my monitor).  The current board design is 5" long.

I looked at the Hammond 1455 extruded cases and this doesn't work well for sizing to fit.  In order to get one with the height clearance for the CMC (1.3" for the CoilCraft, 1.65" for the larger JW 8100), you'd have to go to a size that's at least 4" wide.  Even with the ground trace the board is only 2.4" wide.