RE: Cap-driven x-former?

["Tesla list" <tesla-at-pupman-dot-com>]

Original poster: Jim Lux <jimlux-at-earthlink-dot-net> 

At 08:37 AM 6/8/2004 -0600, you wrote:
>Original poster: "Godfrey Loudner" <ggreen-at-gwtc-dot-net>
>Hello All
>
>Tom McGahee's summary of these H&R transformers can be located at the
>Tesla-2 list in the Old Archives, Mar 14 1997 to Apr 10 1997 Archives. I
>think its at Mar 21 1997. The third winding is referred to as a resonant
>winding that causes the open circuit voltage to jump from 2.7KV RMS to
>5.4KV RMS.
>
>Godfrey Loudner

I doubt it's resonant... I suspect that it's more the cancellation of the 
leakage inductance by the capacitor (well.. it sort of is resonance, in 
that resonance is where the Xl = Xc) but it's not "ferro-resonance" in the 
sense of voltage regulating transformers.

If you do a quick calculation...


4500V limited to 400 mA implies an inductive reactance of about 11.3 kOhms 
on the secondary side.

Say you wanted to entirely cancel that with a series or parallel 
capacitor.  Let's start with trying on the secondary side... It works out 
to about 0.24 uF (at 5000 VAC), which is a pretty bulky PFC  (1.8kVAR)

Maybe it's easier to do it on the primary side...

Start trying on the primary side at 120V.  The turns ratio is 38:1 
(approximately), so the impedance ratio is 1444:1.  That makes the 
reactance about 7.8 ohms.  At 60Hz, that would take C = 1/(7.8 *377) farads 
(340 uF), which is a fairly large "motor run" capacitor, given that they'd 
probably need to rate it at 300V and 8-10 amps or so.

Hmm.. howabout at some intermediate voltage, like 640 V (7:1 turns ratio 
from the output 4500V): impedance ratio is 49:1, so we're looking for a 
reactance of 226 ohms.  Oddly, that's about 11.7 uF (at 60 Hz)... 
remarkably close to the 10 uF capacitor typically supplied with the 
transformers.

Capacitors like these aren't particularly high tolerance, and most 
designers really try to avoid having a leading power factor on loads, so 
you'd typically pick the next smaller capacitor below the calculated value: 
10 uF.

Something else to think about on these units is that the capacitor on the 
tertiary winding is essentially in parallel with the primary.  However, the 
magnetic shunts which create the leakage inductance are between two of the 
windings.  I can't remember whether it's between the primary/tertiary pair 
and secondary, or between primay and tertiary/secondary pair. this would 
make a difference in the exact behavior, because with three windings, you 
can't necessarily move impedances from one winding to another willy-nilly.

I've been meaning to drag one of these out and do some real mutual L 
measurements, because they are a sort of interesting device.

Note that the failures that people have had with these transformers were 
using them with primaries in parallel and secondaries in series, putting 
the secondary winding at twice the rated voltage from the core, NOT in the 
cascade configuration using the tertiary to pass the power up the stack, 
where the voltages are all at the manufacturer's spec.


I also note that there is a version of these with a center tapped winding 
(mine only have 3 straight windings and 6 terminals), perhaps these are the 
H&R version?.  Fr. Tom's writeup also indicates that the tertiary winding 
is coaxial with the HV winding, which is NOT the case for the C&H 
transformer.  In the transformers from C&H (which I have), the windings are 
quite conventional, and stacked on the center leg of the E/I core. The 
windings are held in place by wooden wedges (and a pile of varnish, of course).

The comments about unstable operation with the capacitor connected are 
probably related to interactions between the primary tank cap and the 
residual reactance of the transformer.