Felix for Tube Amps, What to do?

Syrah · « **on:** 19 Aug 2015, 07:53 pm »

Hi All,

Using my Noisesniffer I recently discovered that my Dodd 120w tube amps fed quite a bit of noise back to the line. I have four of these plugged into a PS Audio P10, and it seems to do very little to block the noise from so called zone to zone.

I then remembered that I have quite a few of the the attached Felix boards kicking around in my closet. So I'm thinking of making four of these for my amps.

I have searched these forums, but it is not entirely clear to me what I need to order.

One of the more recent posts suggests that the Coilcraft CMT4-10-15L is the way to go for amps. Do most people agree with this? Will these fit in the printed circuits?

In terms of the caps, what's the best kind to use and from where should/can I get them?

And also, it's not clear what some of the extra holes in the circuit board are supposed to do.

Thanks!

Folsom · « **Reply #1 on:** 19 Aug 2015, 08:43 pm »

All coilcraft chokes fit. Also JW Miller (Bourne) 8xxx series I believe.

What's not clear? Just fill the caps with what fits, X2 rated. They'll probably range in size from 1uf to .01 with .1/.47 in the middle. The specific caps aren't overly important, just X2 rated.

H = hot , N = neutral .

Syrah · « **Reply #2 on:** 19 Aug 2015, 10:02 pm »

Thanks. I got it. So the extra holes are just to accomodate different sized caps then.

Is my best bet for a high powered tube amp the Coilcraft CMT4-10-15L?

Since it's 15a rated, do I need a fuse if I'm using a line that's 15a fused anyway?

Is the best combo with this using 1uf, 0.1uf, and 0.01uf? And if so, which brand works best and where's the best place to find them?

Occam · « **Reply #3 on:** 19 Aug 2015, 11:28 pm »

A couple of points -

1. I don't believe that Felix board can accommodate the Coilcraft CMT4/3 CMCs with 1.5" x 0.9" & 1.2" x 0.7" mounting pin spacing or the larger Bounes 81xx CMCs like the 8121 & 8120 with the same spacing as the CMT3. Its quite easy to mount these CMCs on perforated prototyping boards, and wire everything together as appropriate underneath.

2. If you look at the specifications of the Coilcraft CMT4 CMCs ( or the CMT3s for that matter) their differential (leakage) inductance, 100uH+ for the CMT4s, is quite high and might not be appropriate for this amp, limiting dynamics. A better choice might be the Bournes / JW Miller 8120 or 8121, rated at 17 and 20 amps, with substantially less leakage/differential inductance than the Coilcraft CMTs.

3. Some components simply don't work well with many mains filters such as my CAT-SL1 preamp or Ncore amps using the Hypex SMPSs as they already incorporate a Felix like input mains filter. An alternative might possibly put Feli feeding the components which your power amp might contaminate.

Regards,
Paul

Folsom · « **Reply #4 on:** 20 Aug 2015, 12:44 am »

Oops. I'm pretty certain the 8119-8121 JW/Bourne fit. Those are the slender more spread out holes. Coincidently identical to the Coilcraft CMT3.

The combo diff and comm Coilcrafts fits the boxier shaped holes.

Not sure why I was thinking Coilcrafts all fit... Probably since there's so much smoke in the air here I can barely see more than two blocks....

Syrah · « **Reply #5 on:** 20 Aug 2015, 02:22 pm »

Thanks both. That sounds like sage advice. In a prior filter that I built, I've used the two types of coils in the attached photos. The first one was recommended to me on here I believe, the second was used in Gary Dodd's power filter and he recommended it specifically for his amps. I have no idea what it was, he just sent me a few for free (the kind of guy that he was). I found that both restricted dynamics, but Gary's restricted it less. I still prefered them without.

DigiKey only seems to have the 8020RC in stock. I'm not sure what the RC signifies.

So would this one fit in the printed circuit and be appropriate for the 120w tube amps? What's the best cap pairing with this? In the past I've used Auricaps across the line (I know, I know...) in a metal box and with a switch so it's not on all the time. I might go with X caps this time though, for some peace of mind.

My other (non amp) components are already isolated, since my four 120w tube amps are the only things running off my PS Audio P10. The problem is the amount of noise each amp seems to be feeding back to the line to infect each other amp. It seems to be cumulative as my NoiseSniffer gets louder and louder as I turn each one on.

Occam · « **Reply #6 on:** 20 Aug 2015, 10:14 pm »

Syrah -

Which of the 5 zones in the back of your P10 do you plug each of your 4 Dodd monoblocs into? Where are you plugging the noise sniffer into? In a back outlet on the P10, and if so, which zone? Or into a wall outlet? If a wall outlet, is it on the same mains circuit (fed by the same circuit breaker) that feeds the P10?

Syrah · « **Reply #7 on:** 21 Aug 2015, 01:49 pm »

Thanks Paul. I plug each amp into a different zone. If I plug the sniffer into any zone without an amp, it spits out more and more distortion with each amp that I turn on. So the zones really don't seem to be very isolated.

When I plug all amps into a single outlet with caps accross the line (0.01, 0.1, and 0.47) the noise levels drop.

That's why I got thinking that a high powered amp friendly Felix for each amp might be the ticket.

All of my other components (pre, DAC, computer, TT, phono stage, etc) are plugged into a PS Audio PPP. The NoiseSniffer is dead quiet when plugged into the PPP, so I'm not worried about the tube amps infecting the other components, I'm only worried about them infecting each other.

Do you think the 8120-RC with caps accross the line is the way to go? Will it fit the Felix circuit board?

If not, I was thinking of using terminal strips housed in a metal box. It certainly makes it easier to change out CMCs and caps.

Occam · « **Reply #8 on:** 21 Aug 2015, 05:16 pm »

Syrah,

Neither the Bournes 8120 or 8121 will fit on the 'Official Felix board/kit' -
http://www.audiocircle.com/index.php?topic=51941.msg464578#msg464578
That RC suffix simply means its RoHS compliant, with no lead on its 'tinned' leads, etc...

The first differential inductor which you posted a picture of is 50uH and rated for 15 amps. It appears to be far too much. As to which of the Bournes CMCs would be preferable, dunno. You want as much common and normal mode inductance as feasible to increase the filtering, but not so much (specifically differential/leakage) as to compromise dynamics. The 8120 has 9.5uH leakage and the 8121 4.0uH.
To put additional series inductive filtering in addition to that already there for each zone on the P10 concerns me, in that cascaded series filters can interact in strange ways. Proper damping can address this, but is a time consuming effort to optimize. Assuming you are bi-amping??? with your monoblocks, I'd assume the 2 bass amps are fed from the 2 'high current' zones/receptacles, and the treble amps are fed from 2 of the non 'high current' receptacles, in an effort to minimize potential dynamic constraints.

Ideally in dealing with noise, one should address it at their origin, which appear to be the Dodd monoblocks themselves. The source might be the solid state rectifiers in the Dodds. [This is a WAG, and in no way a criticism of Gary Dodd's (RIP) efforts] If so, this might be directly addressed by a mod recently popularized by John Swenson, and AFIK suggested by Jim Hagerman. This involves putting snubbers [typically a 330 Ohm and 0.022uF cap in series across the secondary(s)] of the power transformer, to minimize switching noise of the rectifiers turning on and off and possible resonances of those transients interacting with the transformer.
https://www.google.com/search?q=John+Swenson+snubber&ie=utf-8&oe=utf-8

Ideally you could find someone locally with the proper skills and inclinations who might also have access to the amps schematics to pursue this route, or find somebody who has already tried those mods on the Dodds.

Regards,
Paul

Syrah · « **Reply #9 on:** 21 Aug 2015, 07:12 pm »

Thanks Paul. Fortunately or not, I live on a Caribbean island, so finding a local tech that I trust or sending four heavy mono blocks away to be modded is not an option for me.

Do you know that the P10 outlets are filtered somehow? They describe them as seperate zones, but my NoiseSniffer would beg to differ.

Maybe my solution is to find some kind of terminal strip or block to which something like the 8120 could plug into, maybe a block with a wire insert on the side and a connector for 8120 pins on the top. That would make it easy to try different ones.

If I end up not going with a CMC, is there still and advantage to caps accross the line for each seperate amp?

Folsom · « **Reply #10 on:** 21 Aug 2015, 08:54 pm »

Capacitors will provide differential mode filtration (such as rectifiers). It's likely this will help a fair bit. Common mode noise is most common from small transformers; probably not the like of your big monoblocks.

I'd use two instead of one cap. It's a relatively easy thing to do, and the benefits are nice.

Can you spot the rectifiers? If they're large chassis mount bridge style they might be easy to change to a HEXFRED to reduce noise more. (If they're SS)

BTW the 17a CMC from JW/Bourne I don't think has a reported loss of dynamics to date. However as Paul (Occam) brought up, resonances are real with these when things get complex (it might have more to do with why some amps don't like power conditioning than anything else).

Syrah · « **Reply #11 on:** 28 Oct 2015, 07:37 pm »

OK. So I bought the below parts to try a Felix with these amps. I might first try them without the coil, then with the coil, to see how they sound. I can probably get them all to fit on the Felix board, with some twisting and turning.

Do these parts look OK?

Occam · « **Reply #12 on:** 28 Oct 2015, 07:49 pm »

Syrah,

The parts seem fine, assuming the larger blue caps are X2 or Y2 rated.
I'd think using the 8120 CMC on that group buy board would be awkward, at best.
Its quite easy to assemble a Felix on a perf board and run and solder the connecting wiring underneath.

Regards,
Paul

Syrah · « **Reply #13 on:** 28 Oct 2015, 07:57 pm »

Thanks Paul. I live on a Caribbean island though, so it's tough getting things I don't already have, even something as simple as a breadboard. I have to be like the Professor on Gilligan's Island.

Just as a purely theoretical question, if I end up note using the CMC, what effect does using 2* caps of each value give - as opposed to a single 1, 0.1, 0.01?

Folsom · « **Reply #14 on:** 28 Oct 2015, 08:29 pm »

Every time you double the amount of capacitors, you half the ESL/ESR. They're adding filtration. At these small numbers it doesn't raise any concerns.

However this will be a fairly weak filter without the CMC. The capacitors will be less effective, and overall attenuation won't be great.

Syrah · « **Reply #15 on:** 8 Nov 2015, 02:45 pm »

Hey Guys,

The bad news. I just discovered that I need reading glasses. Nothing like soldering a circuit to reach that conclusion. I looked pretty funny soldering a circuit with my wife's reading glasses on.

Anyhow, I made one prototype after countless cold solders, redo, resolder, repeat. It definitely killed the noise - my noise sniffer was dead silent. But the amps sounded a bit lifeless.

Then I noticed something. The mere fact that the Felix was plugged into the P10 made every line on it quiet. So I then tried a simple filter with just the three caps plugged into one socket of the P10. Sure enough, almost no noise coming from the noisesniffer. So it seems that whatever noise the Dodd 120s are spiting out is indeed differential mode noise (mostly) and the caps seem to knock it out. So for now, I'm going to finish up a filter plugged into the P10 with just three caps across the line.

I might just finish a couple of Felix's for my B system, TV, etc.

Thanks for all of your help with this!

Occam · « **Reply #16 on:** 8 Nov 2015, 05:21 pm »

Syrah,

Congrats on both your results and your perseverance.The felix should work quite well on your TV
And quite like you, one of the most important things I've learned from my efforts in building and testing power filters is that prescription reading glasses can be a wonderful 'life' accessory...

-Paul

Syrah · « **Reply #17 on:** 9 Nov 2015, 05:49 pm »

I seem to have only solved one problem - the noise generated by the amps themselves. There is still on again/off again noise on the line. The good news is I've identified the culprit. It's a split AC unit with a high efficiently inverter. Luckily the unit's power is easily accessible since it's all wired outdoors.

So my thinking now is, instead of trying to further filter the amps, I should filter the AC unit instead. Is this the way to go?

It's a 240v unit.

What kind of filter should I get, if this is the way to go? I will have my sparky install it. A plug and play from Digikey would be ideal - but they have way too many (and I know way too little about my needs).

Thanks,
Fraser

Syrah · « **Reply #18 on:** 9 Nov 2015, 06:06 pm »

This looks like a pretty extensive supply as well.

http://www.murata.com/~/media/webrenewal/support/library/catalog/products/emc/acline/c09e.ashx

Folsom · « **Reply #19 on:** 9 Nov 2015, 07:15 pm »

You can put a Felix on it, just choose X capacitors that have a higher voltage value. Unless it's 3-phase, then it's a bit more complicated.

The Felix will reduce noise coming out of it, common and differential.

But there's other options like an isolation transformer; it depends what you'd like to spend. It sound like it's pretty awful, and needs all it can get. You could use a much larger capacitor bank on it, but with them I'd want to implement some provisions such as discharge resistors and such. This is a little more experienced territory, the type that Occam will tell you should be uncomfortable if you're not somewhat experienced.