[Occam]

Its relatively easy to find isolation transformers on eBay that are configured as follows-

Some like the SignalTransformer DU series have multiple taps for adjustment of input/output voltages, but the majority simply have dual 120v primaries and secondaries, like toroidal medical isolation transformers, and 'extreme' EI cored transformers from Topaz, MGE, Xentec and Elgar.  These all provide grounded shields between primary and secondary and are generally very good units.

Typically, I've configured these transformers with the primaries in series (wired for 240vac) and the secondaries similarly wired in series for 240vac, but with the center tap of that secondary connected to ground. When you feed the primaries 120vac (after downrating to 1/2 the Kva rating to reflect the current rating constraint), you get 120vac, balanced, split phase 60-0-60vac.
For US code compliance, balancing transformers require a gfci output, as the 'neutral' which is a grounded conductor (NOTE! this is not the same as the 'ground conductor', the mains safety ground), now carries voltage and components do not typically switch and fuse that expected neutral line.

In N.A. (and Japan) a residence typically already has a source of balanced, technical power, our 240vac outlets (in Japan 200vac?), which are used for large appliances like airconditioners, stoves, dryers, etc... I'm unsure as to the availability of similar split phase, doubled voltage outlets in single phase 240vac countries.

So what happens if we turn a balancing transformer around, feeding it 240vac balanced, producing single ended 120vac at the output? Like so -

not shown is the needed input switching and fusing

Consider what C1, a properly specified 'X', across the line cap does. What does it do in the presence of common mode noise? What does it do with normal/differential mode noise?

FWIW
 

[JoshK]

Pretty clever!  I like it. 

[mgalusha]

Nice Paul. I thought about something along those lines one time but ended up with the BPT boxes and never did anything about it.

[Occam]

Well, the alternative methodology for parallel caps on a balanced input or output is Y caps from each phase to ground...... In either case, common mode noise rejection is effected by shunting that noise to ground, with Y caps, capacitively, and with a X cap, via a capacitive shunt to the antiphase winding and ultimately to ground. In both cases the net of the common mode noise is additive which will show up on the ground line. Assuming the ground line has minimal impedance, this should? not be of concern.
The case of Y caps is different for normal/transverse mode noise. On the output (and in the best of all possible worlds on the input) normal mode noise is balanced differential, and the net of the Y shunt currents should sum to 0. But the X cap methodology is different. What we are trying to accomplish is to dissipate that noise energy via the creation of atiphase magnetic flux, but the attenuated noise has the same electrical phase and would not 'cancel' at ground, as in the case of Y caps.

As to which is more effective (or less constrained by the balance of the windings and caps), dunno. My only 240vac outlet is an upstairs airconditioner outlet, so I've only compared this in the 'betwixt and between' location of a Felicia conditioner. The back to back series windings of the secondaries allows one to experiment with leaving both mating secondaries, floating, or either/both centertap grounded, making those connections balanced. For my Felicias, I've found maximal benefit from grounding the centertap of the input of the output transformer, similar to the above schematic. So in the Felicia, I ground both centertaps of the series wired output transformer windings, and use only X caps across the phases. [If you're not familiar with the Felicia balancing conditioner, this will make no sense  ]

Although Y caps are on the balanced outputs (aided by series differential inductors) of Glasband's balanced conditioner patents, I rarely see the Y caps used in actuality. Could folks with access to commercial balanced conditioners please comment on how the caps are configured in their conditioners?

On another note, I believe the schematic in my first post is similar to how a Torus conditioner is configured when fed from split phase 240vac circuit.

FWIW,
Paul

[Occam]

For those who've got a conveniently located 240vac outlet (In N.A./Japan where they're split phase / balanced) Newark caries a 115/115:115 50va isolation transformer that one could try on a lower powered source component -
http://www.newark.com/jsp/search/productdetail.jsp?sku=18M9252

One could also use it to run a 120vac component outside of NA off a 240vac single phase mains line, or in N.A. as a 120vac:120vac isolation transformer by simply wiring the 2 primary windings in parallel.