[tanchiro58]

tanchiro58,

Also questioned the others about whether this configuration changes the sound since I built the my first Felix based on the original schematic too...

The first Felix I built used the Miller CMC and .47/.1/.01 caps; I built it exactly as the original schematic, placing the output .47uf cap next to the CMC, and the .01 bypass cap farthest away.
I put this one in the AC line to my preamp and was very happy with the results.

After seeing the other builds which placed the smallest value caps next to the CMC, I built another Felix this way using the same parts, and hooked this one up to my preamp.

My results are subjective; but I seem to notice a little more detail and speed to the sound with the .01 bypass caps close to the CMC; at least in my audio system.

I did not notice much difference with the different Felix config when they are used for my small TV monitor.

Cheers.

Hi Hap,

To be clear I assume that you said the latest Felix model  you built with 0.01mF closer to CMC (Main in 1.0, 0.1, 0.01--CMC--0.01, 0.1, 1.0 main out) has more details and faster sound then the previous Felix (based on Gary-GBB's schematic) with 0.01mF closer to CMC on main input and 0.01mF farther away from CMC (Main in: 1.0, 0.1, 0.01--CMC--1.0, 0.1, 0.01 main out). Correct me if I am wrong between the above Felix's configuration. Me too I am very happy with Gary's schematic in my system for now but I will try your way with two 0.01mF closer to CMC (mine is Coilcraft 4018/10A). Thanks.

Tan

[Hap]

tanchiro58,

It is just my subjective opinion that the Felix configured with the smallest bypass caps closest to the CMC sounds better in my audio system.
As for Gary's Felix as well as the original for that matter, this configuration I prefer:
Main in: 1.0, 0.1, 0.01--Coilcraft Q4018-A CMC--0.01, 0.1, 1.0 main out;
Main in: 0.47, 0.1, 0.01--JW Miller 8109 CMC--0.01, 0.1, 0.47 main out.

For the record, it is not my idea to configure the caps like this; there are many posts from Occam and GBB that suggests this configuration, and also give logical reasoning as to why the caps should be installed this way.

Cheers.

[Occam]

Guys - There is no need to acknowledge things that are standard procedures in the realm of power electronics, bypassing, etc.... Important things yes, like Gary as the quintessential hipster, my people skills, Gordy's chick magnetism, JoshK's stunning good looks.... 

[tanchiro58]

JoshK's stunning good looks....  Cool

Really? 

[BradJudy]

Randy has pointed out to me that my design is based on a different set of caps than Gary recommended.  I designed it based on the Vishay/Roederstein F1778 X2 caps and his was based on Vishay/BC MKP 338 X1 caps.  Accommodating the lead spacing on the MKP338 X1's wouldn't be a problem, but they are notably thicker than the F1778's, so the board would have to be expanded.  The MKP 338 series also has an X2 rated set, but the lead spacing on the 0.01uF in that series is too close to safely fit this design. 

What are everyone's feelings on cap models? 

BTW: I corrected a spacing issue with the JW Miller CMC, and everything else appears correct (pending cap decision).

[tanchiro58]

This is what Gary has suggested:

Coilcraft P3717-A CMC, available on Coilcraft's website
--- I used X1 rated polypro caps from Digikey in values of 1uf, 0.1uf, and 0.01uf on each side of the CMC
--- Part numbers are BC1576-ND, BC1587-ND, and BC1588-ND - Vishay/BC 440VAC X1 rated caps

I built one which works fine. 

[Gordy]

BradJudy

Rather than expand the board further, would it be possible to add solder pads to the top of the board at the .1Uf locations?  At 16.5mm the 338's will fit within a 3/4" standoff.  Or will the width/thickness of the .01 338's be problematic as well?

[BradJudy]

Bottom mounting is an option for the 0.1uF.  The thickness of the 0.01uF shouldn't be an issue.  Adding a 15mm spacing to C3/4 is easy enough - I just made that change.

The 1.0uF would be right on the input and output pads.  Here's a quick image with an overlay of the MKP 338 1.0uF cap in the C1 position.



Any reason X1's are being used?  It seems like overkill.

[TomS]

Bottom mounting is an option for the 0.1uF.  The thickness of the 0.01uF shouldn't be an issue.  Adding a 15mm spacing to C3/4 is easy enough - I just made that change.

The 1.0uF would be right on the input and output pads.  Here's a quick image with an overlay of the MKP 338 1.0uF cap in the C1 position.



Any reason X1's are being used?  It seems like overkill.

These look really great and I'd definitely be in for a few PCB's on a group buy.

Any consideration to space the line/neutral/gnd pads to accommodate readily available Eurostyle screw terminal blocks as an option?

Tom

[randytsuch]

FYI people, the X1 caps that Brad orginally used are rated for 275VAC, and the ones Gary used for his project (that I bought) are rated for 440VAC.

Either one would be fine for the 115VAC we have here in the US.  275VAC is a little close to European voltages, but I am not sure how much derating is recommended for AC caps.  I remember seeing somewhere that England runs a little higher (maybe up to 240?), and that would leave 35V derating, which would be a little close for me.  I would try for at least 20%.  We derate 30% at work, but different caps and different application.

Brad
Did not get a chance yet to look at the revised layout, with all the different caps, will try to do it today sometime.

Randy

[kyrill]

ok ppl

dont have to make it perfect, we need an operational decision and carry it out
we are all waiting

[BradJudy]

FYI people, the X1 caps that Brad orginally used are rated for 275VAC, and the ones Gary used for his project (that I bought) are rated for 440VAC.

Just to clarify, X1 caps are all rated to 440VAC and X2 caps are all rated to 275VAC.  My original layout used X2 caps. 

[Occam]

Randy,

While your conservatism regarding these safety cap's voltage ratings is admirable, in this case, any cap rated X1, X2, Y1, Y2, is fine for this application, whether one is in a nominal 120 or 240vac country. When these caps are rated they go through a rigorous certification process (unlike most other caps), and tested at voltages substantially above nominal.

All thing being equal, type '1' caps are going to significantly larger than type '2', and 'Y' caps, larger than 'X'.  When space is a constraint, go with X2 rated caps for the Felix. As to which type (and brand or specific model) is preferable from an objective or subjective perspective, I ain't going there......

Regards,
Paul

[BradJudy]

ok ppl

dont have to make it perfect, we need an operational decision and carry it out
we are all wating

Don't push me, I'm weary enough about liability issues on this project, I don't need people bugging me about getting it done fast. 

[kyrill]

ok

[BradJudy]

I don't mean to be a jerk, but the reality is that even in the most ideal situation, final boards will take weeks to be in my hands.  Here's the process:

Finish design and order prototype boards.
Order test-fit parts from at least three places (three different CMC's, multiple sizes of two different lines of caps)
<wait a week to get everything>
Test fit everything - if there are problems, go back to step 1 (each iteration is another $200)
Start a group buy thread that allows sufficient time for people to get in orders and send payments (one week?)
Order production boards
<wait two weeks>
receive boards, verify fit with test components
package and mail boards to everyone who ordered
<everyone waits a few days to receive them>

Even once a board is complete, these types of group-buys take several weeks. 

Naturally, all of this assumes I don't run into problems or have other time sinks come up. 

[TomS]

I don't mean to be a jerk, but the reality is that even in the most ideal situation, final boards will take weeks to be in my hands.  Here's the process:

Finish design and order prototype boards.
Order test-fit parts from at least three places (three different CMC's, multiple sizes of two different lines of caps)
<wait a week to get everything>
Test fit everything - if there are problems, go back to step 1 (each iteration is another $200)
Start a group buy thread that allows sufficient time for people to get in orders and send payments (one week?)
Order production boards
<wait two weeks>
receive boards, verify fit with test components
package and mail boards to everyone who ordered
<everyone waits a few days to receive them>

Even once a board is complete, these types of group-buys take several weeks. 

Naturally, all of this assumes I don't run into problems or have other time sinks come up. 

You're not being a jerk, just using common sense.  I, for one, very much appreciate your initiative and effort to go through this.  It's not easy.

If somebody really NEEDS a Felix right now, it's about a half hour job to wire it PTP on perf board you can get at any RatShack, so just go do it (it's only 7 parts max).  I plan to order some of the PCB's for the long term just to have around to build when I need a few more.

Tom

[mgalusha]

Order test-fit parts from at least three places (three different CMC's, multiple sizes of two different lines of caps)

Brad,

Let me know and I'll send you one of the Coilcraft CMC's and the Vishay X2 caps. I might be able to get them mailed today but if not, shouldn't be a problem for tomorrow.

Mike

[BradJudy]

Thanks for the offer Mike.  The CoilCraft and Vishay 1778's were the ones I was planning on using for my own builds, so I'll be buying some of those anyway.  Plus, I'll send you one of the proto boards to test-fit anyway.  I'll just end up with a bit of frankenstein Felix builds with a combination of CMCs and caps. 

[BradJudy]

Any consideration to space the line/neutral/gnd pads to accommodate readily available Eurostyle screw terminal blocks as an option?

I'll take a quick look at that option as I do like screwblocks in general. 

Note that there is not a ground pad on this design - the feedback was to eliminate it as it did not serve a function on-board and should be handled off-board.