[Occam]

Cantskienuf,

Your transformer is working as it should, but sadly its not producing balanced (technical) power, but rather isolated power. 'Isolated' in this case has a specific meaning that it is not balanced, having a center tap tied to ground, nor is it configured as a typical isolation transformer, with one of the ends of the secondary winding tied to neutral. Rather, it is configured as what is typically used in a hospital environment, with a floating output, with no reference to ground whatsoever. This 'isolated' output is used in hospital environments as it greatly minimizes the leathality of potential shocks. I find this puzzling as the GIS-500 is not listed as a medical device, and by code (FWIW), non medical isolation transformers are supposed to re-bond the nominal neutral to ground at the transformer's output. It is not unusual that under minimal load you'd get those voltage readings. And if your taking those readings at the output of JonRisch powerconditioner, the Y capacitors in it (from 'hot' to ground, and from 'neutral' to ground) are functioning as a resistive divider with the centerpoint tied to ground. But while this will produce a balanced reading under your measurement conditions, it does not establish true balanced, technical, power, in that is does not provide a canceliation of chassis leakage currents.
The above was simply a verbose way of saying, yes, your measurements are perfectly reasonable, but no, this is not true balanced power providing the specific benefits of technical power mentioned above.

[Occam]

A possible topological enhancement to the Felica
aka
Going all differential upside Felicia's head

Felicia has never had 'Y' [from the power wires 'to-ground', usually from 'line-to-ground' and 'neutral-to-ground'] in her innards. These caps are to shunt differential noise to ground. I've eschewed them, as they inevitably leak noise onto ground wire. Also, the current of the shunted signal could trigger a downstream leakage detector, or worse, present a shock hazard.
Heretofore, Felicia has relyed on the 'accross the line' caps [aka the 'betwixt and between' or 'X' caps] to convert differential noise to common mode noise, to be dealt with by the flanking transformers -
http://www.audiocircle.com/index.php?topic=18441.msg161831#161831&highlight=magic#161831

Given my 'do no harm' attitude, Felicia has relied upon those 'X' caps for fear (either real or imagined) of contaminating the mains ground. But looking at the schematic (top post of the previous url) there is a potential opportunity, if we ground the center tap of the input transformer's secondary and leave everything else alone. A substantial benefit of balanced power is the minimization, via cancellation, of reactive leakage currents -
http://www.equitech.com/articles/widescreen.html
This grounding of the input transformer secondary's center tap, will make those reactive leakage currents from the secondary windings of both transformers balanced with respect to ground and hence cancel to a large extent. The output transformer's output is already grounded at its centertap, which leaves only input transformer's input (primary) unbalanced with respect to ground. If that Felicia is housed in a conductive enclosure (and by definition, that enclosure is grounded to the mains ground) the result will be minimized ground leakage and corruption.

While transformer action can minimize common mode noise, it does not do the same with differential noise. But ideal transformer action does something very useful with differential noise - it converts single ended noise to differential BALANCED noise.  And now, the noise shunt provided by 'Y' caps to ground is also balanced, and ideally much of that differential noise cancels, minimizing whathever crap is dumped onto the mains ground.

Long story short -
1. Accross the 2 lines of the secondary-to-secondary connected transformers [which already have the 'betwixt and between' cap(s) accross], connect 2 .1uf Jantzen from PartsExpress caps in series with their writing oriented symmetrically; meaning that the writing for both caps starts with both on the outside or both on the inside.

2. Connect the center of that cap string to the center-tap of input transformer's secondary.

3. Connect the mains 'star ground' to that same center-tap of the input transfomer's secondary.

I plan on testing this with the connections in 2 & 3 gang switched via a dpst switch, which would change between just having the 'Y' caps function as a .05uf accross the line, 'X' bypass, or additionally provide (whatever)benefits of the differential filtering of the 'Y' caps with minimized reactive leakage.

If you don't follow the instructions, wait until some feedback on initial results is posted. My schematic drawing program is behaving strangely, so I can't currently draw a picture.

FWIW

[JoshK]

very interesting, and easy enough to try.

[Occam]

Here is a kludged together schematic


Note that the only difference from the current standard -


is the addition of grounding to the input transformer secondary's center tap, and the grounded Y caps. These are shown with switches for testing the new topology's efficacy.

[seppstefano]

Kudis Occam, JoshK and ciao all,
I've recently got into the mains filtering world (I live in Italy, just for voltage issues).

My first build was a simple parallel cap-based filter (a variation of Thorsten Loesch), already using motor cap (400VDC).

Recently I build a pairs of isolation filters (not balanced), each unit using a pair of 220->85 parallel RC zobel at 200KHz parallel 85->220 trafos.

These trafos (0,160 KVA) have no center tap. Would you think is it possible to use them at half the voltage input and parallelizing them (both input and output)? I should use 4 trafo for a Felicia, but theortically I can't see drawbacks, do you?

Thanks in advance,

Stefano

[Occam]

Stefano,

Some schematics (required fusing is not shown) -


But first things first. Optimize the powerconditioning characteristics before you complicate the circuit with extra transformers. While balancing is a substantial improvement, it would be far more efficient, cost/complexity wise to optimize the conditioning part first.
From your post on the AudioAsylum, I've no idea as to why you've put a Zoebel with a .01uf cap and 70ohm resistor in place of the Cx capacitor. The Cx capacitor(s) are critical for the optimal performance of the conditioner. In my version with a 175va transformer, I use an 18uf cap, bypassed with a .47, .1 & .01uf capacitior. I don't know the optimal value for your specific transformer, but would guess its between 4 and 20uf for the 'bulk' cap. You'd need to experiment. Your choices of bypass capacitors depends on your own cost benefit analysis.

For the Co caps, I use a 0.47, 0.1 & 0.01uf cap. I don't presently use a Ci cap.....

[seppstefano]

Ciao Occam, ciao all,
thanks a lot for kind and fast reply I thought of something like in your second drawing

To be a bit more informative, I must say that my system already benefits a simple unit based on caps. This includese a 5uF motor cap, paralleled by a 0.1uF and a 220KOhm between L/N and a 0.1uF paralleled with a 220KOhm between N and E.

This unit lightly filter all the mains (arranged in a star - rather tree - geometry).

At the leaves I thought of adding a filter for digital nasties (a-la Jon Risch). Then I discovered Felicia. Now, I'm evaluating to modify my roadmap and include Felicia (this for the history).

The R in the zobel R/C between my trafos is meant to eat the energy (i.e. noise) at frequencies >= 200KHz.

A thing I cannot well understand is how can each parallel cap by itself eat the spurious energy? Is it correct to say that without a R, the energy is dissipated by the trafo winding (so the trafo gets hot)? Isn't it better to add a R eating the energy? Now, the R of my trafo is around 10Ohm. Rather than 18uF, 10Ohm wouldn't is be possible use a R=180Ohm, in series with the 1uF cap. What would you think?

Many thanks for sharing,

Stefano

[nature boy]

Anyone find another inexpensive source of transformers for Felicia?  BG Micro is still showing no availability.

NB

[Occam]

Stefano,

Its well and good that you've put a capacitive noise filter ahead of your 'Felicia'. Your Thorsten Loesch / VanEvers type filter (a more sophisticated version of Ci) is in parallel with all components connected to that mains line and therefore offers noise filtering to components not fed by your 'Felicia'.
But what you have implemented is not a Felicia, it is an isolation transformer, offering minimal powerconditioning. Similarly, if you add those extra transformers as shown in the schematics above, you'll simply have a balancing transformer. Do not assume that your capacitive filter ahead of Felicia eliminates the need for a properly sized Cx.

Please read the Felicia 'sticky' -
http://www.audiocircle.com/index.php?topic=18441

and Jon Risch's reply on AA -
http://www.audioasylum.com/scripts/t.pl?f=tweaks&m=124472

Capacitors do not serve to dissapate noise, they are frequency dependant resistors, which simplistically function to divert noise away from your powered component. While your Zoebel might minimize a resonance, its noise filtering contribution is minimal. Put in a Cx cap. Experiment to optimize for your specific transformers.

[seppstefano]

Ciao Occam,

> Stefano,
> Its well and good that you've put a capacitive noise filter ahead of your
> 'Felicia'. Your Thorsten Loesch / VanEvers type filter (a more
> sophisticated version of Ci) is in parallel with all components connected
> to that mains line and therefore offers noise filtering to components not > fed by your 'Felicia'.

Indeed, the TL type filter filters ALL components.

> But what you have implemented is not a Felicia, it is an isolation
> transformer, offering minimal powerconditioning.

I'm aware of this. I have not implemented any Felicia. Only yesterday I have been told of the Felicia project.

> Similarly, if you add
> those extra transformers as shown in the schematics above, you'll
> simply have a balancing transformer. Do not assume that your
> capacitive filter ahead of Felicia eliminates the need for a properly sized
> Cx.

> Please read the Felicia 'sticky' -
> http://www.audiocircle.com/index.php?topic=18441

Already read it, thank you.

> and Jon Risch's reply on AA -
> http://www.audioasylum.com/scripts/t.pl?f=tweaks&m=124472

Got it, thanks to your indication

> Capacitors do not serve to dissapate noise, they are frequency
> dependant resistors, which simplistically function to divert noise away
> from your powered component. While your Zoebel might minimize a
> resonance, its noise filtering contribution is minimal. Put in a Cx cap.
> Experiment to optimize for your specific transformers.

...hhhmmm, the real part of their Z acts effectively as a resistor, indeed. So it dissipates power = Re(Zc)*Ic^2. That's why I thought of the power issue... I will try anyway

Have a nice evening (at least from my point of view and meridian...),

Stefano

[Occam]

.... hhhmmm, the real part of their Z acts effectively as a resistor, indeed. So it dissipates power = Re(Zc)*Ic^2. That's why I thought of the power issue... I will try anyway

Stefano,

This misunderstanding is my fault. I should not have describled capacitors in this function as 'frequency dependant resistors' but rather as frequency dependant reactances. Pure capacitors and inductors do not have resitance (actual components have the DCR of inductors and ESR of capacitors), they have reactance (X). Unlike resistance (R), reactance does not dissapate current, rather, it offers inertia to the flow of current. This is why if you hook up a properly rated large value capacitor accoss your mains line, it does not get hot, rather it is shifting the phase between the voltage and current waveform.
It is the combination (in the complex plane) of resistance and reactance which comprise impedance (I), and it is the frequency dependant impedance which determines the effectiveness of a filter, whether for a power filter or a loudspeaker crossover.
http://www.google.com/search?hl=en&q=Kirchhoff%27s+Voltage+Law+reactance

All I can hope for is that you'll be pleasently surprised at the benefits of adding the Cx capacitor(s).

I do realize that the 27 pages of this thread is quite difficult to wade through, and I hadn't realized you were new to the thread.

I never cease to be amazed and impressed by the eloquence of non-native English speaking posters. All the moreso when I look of the language skills within my own family and country.

Regards,
Paul

[BobM]

Hey Paul,

Stik it yous snothead. You aksing for troubel talkin' bad bout us Newawkers dat way.
 
Bob

[Occam]

Hey Paul,

Stik it yous snothead. You aksing for troubel talkin' bad bout us Newawkers dat way.
 
Bob

Well, the one advantage of living in cosmopolitan New Yawk is leaning the meaning of curses in a variety of languages. Sadly, most of those curses were directed at me   .

Kidding aside, in Europe many folks know at least 2 other languages quite well. Thankfully, my high school French is so bad, that in defense, others I'm trying to converse with will switch to English. My eldest son  chose to attend RPI specifically because it didn't have a language requirement. My youngest is attending a liberal arts school, and is thankfully forcing him to gain fluency in another language.

[ctviggen]

I hear this argument all the time -- Americans don't know enough foreign languages.  But where do you practice such speech?  Even my dad, whose mom came from Italy, can't speak a lick of Italian.  I know as much Italian as he does.  In Europe, you're just a short trip away from somewhere they speak a different language.  Here (other than going to Brooklyn or the deep South), we're no where near anywhere a foreign language is spoken.  Although I wouldn't mind learning Spanish, as their are a lot of people of Spanish lineage here (and in AZ, where I used to live).

[seppstefano]

Stefano,

This misunderstanding is my fault. I should not have describled capacitors in this function as 'frequency dependant resistors' but rather as frequency dependant reactances. Pure capacitors and inductors do not have resitance (actual components have the DCR of inductors and ESR of capacitors), they have reactance (X). Unlike resistance (R), reactance does not dissapate current, rather, it offers inertia to the flow of current. This is why if you hook up a properly rated large value capacitor acco ...

Ciao Paul,
thanks for feedback and let me know if you want me to start writing some italian (oder, also weniger, Deutsch).

Being absolutely non contradictory, I remind some theory from my EE courses. Pehaps we could agree on:

http://www.mpdigest.com/Articles/Sept2001/atc/Default.htm

Aside from an academics point of view, ESR is typically a low value so perhaps we have no problems on power and my worries are exaggerated

I'll give the 10uF cap between a try

One (of many) thing I don't understand is how's that we have no leak current when between windings (at sufficiently high frequency the cap acting as a short)?

Ciao,

Stefano

[Christopher Witmer]

nature boy, on page 26 of this thread, Paul Occam posted this link to transformers that seem like they can substitute for the ones that BGMicro was carrying earlier.

http://www.mpja.com/directview.asp?product=7846+TR

Other than that, I recommend checking ebay. If you know what to look for, you can usually find something appropriate without too much trouble. If you don't know what to look for, it is probably a good idea to ask on this thread before purchasing a particular transformer, because sometimes it is easy to make careless mistakes, like getting a pair of machine control transformers that have dual secondaries but only single primaries. (Of course, if the secondaries were of just the right voltage for balancing to get the requisite output power, then all would be fine, but usually that is not the case and thus the sort of transformers I just described would be a no-go.)

[Gordy]

Hi NB,

A magnamonious benzefactor has supplied some candidate transformers for your magnum opius!  PM me as to when we can try to char that cement slab of yours  

[audioferret]

Have you compared the Felicia to the following websites?  This guy does an excellent job describing how to wire the conditioner.  But, after reviewing your documentation, I am wondering if his setup is really balanced power.  What is the difference between using one transformer and two (in series) like you have?


http://home.comcast.net/~thomasw_2/CheapskateBP1.html

or his second site, showing the innards of a trancendant sound power conditioner...and his rebuild of the same.  The TC page is claiming that their item is true balanced power...so what's different?

http://home.comcast.net/~thomasw_2/TranscendentSoundKit.html
http://www.transcendentsound.com/power_supply.htm

[JoshK]

Thomas' project is definitely balanced power but it is not a Felicia.  Using two transformers with caps in between helps filter common mode and differential noise, so it is more than just balanced power.

[Occam]

Audioferret,

Thanks for the references to those interesting URLs. As Josh said, while both of those projects, like the Felicia, provide the benefits of balanced power, the 'nulling' of reactive leakage currents on the fed components (to the extent that those center tapped output windings are balanced), those projects take a different approach (and I believe, efficay) towards the power conditioning function.

As described by ThomasW, both the Transcendent balancing conditioner and his own project use Avel Lindberg Y23 series toroidal transformers for balancing. Their charateristics are described in the PDF downloadable at the PartsExpress link for the 800va, dual 60v secondary toroid Y236905, and in the case of the Transcendent project, the 1000va version, Y236959. IMO, neither of these transformers are particularly appropriate for their role in balancing condintioners. Specifically, they lack the interwinding grounded shield between primary and secondary windings that you'd find in commercial balacing conditioners like those from Equitech and BPT. Due to the lack of that interwinding shield, the capacitive coupling between the primaries and secondaries are very high bandwidth and will do little to stop the passage of high frequecy noise. This is especially true in the case of differential mode noise, where the intrinsic rejection of common mode noise by transformer action does not come into play. In the case of Felicia, while its transformers also lack an interwinding shield, winding capacitive coupling is minimized by its physical construction that physically separates the primary and secondary windings, split bobbin construction.

I realize that both projects take steps external to the toroid for power conditioning. While I've no doubt that commercially available powerfilters, Corcom, Shafner,Schurter, etc... might be effective in that role, personally, I've yet to find any that I've found to my subjective liking. Additionally, I'd suggest that both of these toroidal based products would benefit substantially from a big honking X cap accross the primaries to facilitate turning differential noise into common mode noise, that those toroids might deal with more efficiently. I'll also point out that the choice of 115v primary/ 120v secondary transformers, could be problematic for those residing in high mains voltage areas. (Felicia has the opposite problem, in that it drops voltage). For an explanation as to how Felicia works -
http://www.audiocircle.com/index.php?topic=18441

And yes, I read ThomasW's follow ons 'improving the....", but the use of a highly capacitively coupled transformer, will forever constain the efficacy of those projects, or as my Mama allways told me -
'You can't polish a turd.'
This is not a criticism, simply an acknowledgement of the constraints conferred by the economic choices made.

While all this technical gobbledeguuk, might (or might not) be convincing, the proof is in the listening. I have heard the Felica compared to the Transendent Power Conditioner, as have others, right here in the Nasty Apple, home of truly appalling mains power.
http://www.audiocircle.com/index.php?topic=18319.msg167642#167642&highlight=transcendent+felicia#167642

I'm certainly not stating that Felica is the be all and end all of balanced powerconditioners, but I will state that the Felicia, within the constraints of what it can power, is substantially [put your own adjective here, metrics being quite subjective   ] better than balancing conditioners using transformers, with overwound, non shielded primaries and secondaries, whether they be EI or toroid cored. YMMV

FWIW,
Paul