By George White [george@palisad.com]:

There is a simple experiment that can be performed to show conclusively that the bigger effect is not an ionic wind and clearly electromagnetic in origin.  Further experiments have shown the exact nature of this effect and have led to verified predictions for how to enhance it.  Correlated analytical results have been obtained which rely on nothing more than ordinary EM physics, although an alternative curvature based hypothesis has also been developed and yields the same analytical results.

The experiment uses a candle flame to measure the effects of EM forces (just like in the original B-B experiments).  Construct an air dielectric, symmetric parallel plate capacitor with plates about 10 cm on a side and spaced about 6 cm apart (I stuck some plates into a block of Styrofoam).  Apply a 1 Hz modulated AC signal of a few kHz and a few kV (I had up to 20 kV p-p available).  Using a small votive candle, **CAREFULLY** move it around the device (including the space between the plates) and observe the flame.  The flame will symmetrically squash and relax at the 1 Hz modulation rate.  This is easily explained as the polarization in the molecules of the flame align against the applied E-fields and oppose it (look into the physics of dielectrics for more details).

Make one of the plates much smaller (a small wire loop about 3 cm in diameter works well) and perform the same measurement.  Now the candle flame will squash and relax in time with the 1 Hz modulation, but a far more pronounced effect will be the flame tip bending towards the larger plate (an indication of a net force acting on the flame).  Behind the large plate (i.e., the wire loop is shielded by the large plate), the behavior is symmetric squashing, exactly as in the parallel plate case, but both inside the capacitor and in the region of space behind the small plate, the bending effect will be very pronounced.  The net force is effectively the difference between the symmetric squashing behind the large plate and the tilted squashing ahead of the small plate.

If the candle is surrounded by a Faraday cage (i.e. an electrically floating  cylinder of brass screen), the effects will almost completely disappear in all cases.  A physical barrier around the flame (i.e. a glass cylinder) has no discernible effect.  The combination of a Faraday cage and the physical barrier completely eliminates all effects.  This indicates that there does appear to be an ionic wind component, but it's dwarfed by the EM effect.

Furthermore, using a 1 Hz modulated signal makes the effect visible at potentials far below that required for any ionization.  I first noticed the effect at potentials of well under 1 kV p-p (30 V/mm) between the plates.

Further experiments have been able to show that this is a direct effect of the warping of the mid potential surface (I'll call this the SOE for the remainder of this discussion) bisecting the plates of an asymmetric capacitor.  Using the same apparatus with differential AC applied and no modulation, use a scope probe to detect where the SOE is.  This is easily detected as the surface where the AC voltage measured by the scope probe is at its minimum (you will also see a sharp phase shift).
Be careful not to touch the probe to either plate.  You can wrap it with insulation to be safe.

For the symmetric parallel plate case, the SOE is an infinite plane, parallel to the 2 plates and exactly midway between them.  For the asymmetric case, this becomes more parabolic shaped.  The SOE moves towards the smaller plate and extends beyond the plane of the small plate.  This can also be calculated using accurate E-field/potential analysis software.

To analyze what this means, consider that the SOE defines a surface where the capacitance between it and the large plate is the same as the capacitance between it and the small plate (i.e.  if C1 == C2, VC1 ==
VC2) and each effective capacitor has twice the capacitance as the total between the plates (Ct = C1*C2/(C1+C2)).  This also means that as far as the energy stored in the E-field, the SOE defines the boundary that divides the stored energy into 2 equal parts.  The implication is that for an asymmetric capacitor, the energy is stored on both sides of the smaller plate, but only on one side of the larger plate.  Since force is energy divided by distance, the force arising from the stored energy cancels some of itself out around the small plate (but not the large plate) resulting in a net force acting on the plates.  Generally this force is quantified by q**2/2*e0*A, but is also quantified as E/d, where E is CV**2/2 and d is the space between the plates.  The fraction of this that becomes the net force depends on the geometry of the mid potential surface and how the resulting force components around the small plate cancel themselves out.

Further experiments have verified the q**2 dependence of the resulting force both qualitatively and quantitatively.  The produced force is dependent on V**2 and is always less than C*V**2/2*d.  Other experiments regarding geometry changes that affect the shape of the mid potential surface have the expected results on the net force.

At higher energies, even more interesting results appear.  The sine waves of the measured E-fields morph into triangle waves between the SOE and the large plate and into square waves between the SOE and the small plate.  This clearly indicates a voltage dependent capacitance where C between the SOE and the large plate is decreasing with increasing potential and that between the SOE and the small plate is increasing with increasing potential.

One hypothesis for this is that at higher energies, space near the small plate becomes ionized, reducing the effective plate spacing and increasing the capacitance.  However, if this was the case, you would expect to see the SOE move farther away from the small plate instead of seeing the E-field waveform on the other side morph into a triangle.

In fact, as the SOE moves away from the small plate, the capacitance between it and the large plate would increase and not decrease.

A more intriguing hypothesis is that the space between the SOE and the large plate is becoming less curved than ambient space (i.e. the effective distance between the plate and the mid potential surface is smaller in its own reference frame than it is in the observing frame) and the space between the SOE and the small plate is becoming more curved than ambient space.  This sets up a net curvature gradient around the SOE, which coincidentally, also produces a net force on the plates as the SOE is pushed towards the smaller plate (to equalize the curvature on its two sides) and drags the plates along for the ride.

You can also think of this as the less curved region falling into the more curved region.

George White
Email: Click Here