Re: High Freqency Voltage Stress on NSTs

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

Original poster: "Bert Hickman by way of Terry Fritz <twftesla-at-qwest-dot-net>" <bert.hickman-at-aquila-dot-net>

Terry and all,

Thanks for providing the very interesting tests! At this point, I'm not
sure I share your conclusions. A couple of comments follow. 

The test measurements seem to show that there's no evidence of voltage
concentration towards the outer portion of the winding for the frequency
range of 10 - 510 kHz. However, while these measurements provide voltage
information, phase information is not apparent. It may be more meaningful
to measure differential voltages between the input (outer layer) and
internal taps, and run the swept frequency measurements to a higher
frequency. However, even this approach will likely not reveal the complex
behavior seen under transient conditions. 

There has been much speculation that NST damage is not so much induced by
backfed tank RF at the coil's operating frequency, but is perhaps caused by
high frequency transients or parasitic oscillations generated when the main
gap fires. A more revealing measurement might be to make differential
voltage measurements between the outer winding and internal taps while
driving the secondary from a fast, low impedance, pulse generator. I
suspect that you'd see effects that are not easily seen when using steady
state excitation.

BTW, there IS a significant body of work that's been done on the transient
response of distribution transformer windings. This has been driven by
necessity, since transformer manufacturers and utilities have had to learn
how to safely deal with voltage transients from switching, lightning, or
arcing grounds. Some excellent discussions can be seen in the 10th edition
(1973) of "The J & P Transformer Book, a Practical Technology of the Power
Transformer", by S. A. Stigant and A. C Franklin, and in "Power
Transformers and Special Transformers" by S. Rao, Khanna Publishers, 1996.

Summarizing from these sources, the transient behavior of a transformer
winding is a function of the frequency makeup of the incoming transient,
and the geometry of the winding/insulation. At low frequencies, winding
inductance will dominate. Since winding inductance tends to be
linearly distributed throughout the winding layers, the result is a
relatively even voltage distribution across layers. However, if driven by a
fast voltage transient, the voltage distribution across winding layers will
be mostly governed by the distribution of winding-to-winding capacitances
and winding-to-ground capacitances. Initially, the inductance of each
winding layer "looks" like an open circuit. 

As the transient event continues, the network of inductances and
capacitances begin to "ring" at various frequencies, altering the initial
voltage distribution. In many transformers (NST's included), each winding
layer tends to be similarly constructed as the layers above and below, with
layer-to-layer insulation also being identical. With this configuration,
the winding layer-to-layer capacitance should be about the same as we go
from the outer layers towards the core. However, (unless special
electrostatic shielding techniques are used) layers closest to the core
will tend to have greater winding-to-core (ground) capacitance than those
further removed. This causes non-uniform voltage distribution under
transient conditions.

Numerical analysis of this complex system can be quite challenging.
However, there appears to be significant theoretical AND empirical evidence
that transformers driven by a step voltage transient can develop voltage
stresses, with 60-90% of applied stress appearing across the outermost 20%
of the winding. As a result, layer-layer voltage stress may be 3X-5X normal
levels. As the transient event continues, internal oscillations are
established within the network of inductances and capacitances in the
winding layers, and the system begins to behave as a type of transmission
line. As the region of excessive voltage stress migrates inward, voltage
stresses hit the middle winding layers. The disturbance continues to
propagate inward, being "reflected" from the innermost (grounded) layer,
and the reflected disturbance creates a region of _even higher_ stress
between the innermost 20% of the windings and the core! Depending on the
insulation system's "weakest link", actual breakdown may occur ANYWHERE in
the winding. However, in looking at the transient curves from the J&P book,
breakdown would appear to be more likely between the innermost 20-30% of
the winding and the core. I can scan some of this in if there's significant
interest.

In any event, RC filters SHOULD be used to prevent rapid dv/dt transients
and parasitic oscillations from getting back into the NST windings. While
we can debate the destructive mechanism that's occurring inside the NST's,
there's little dispute regarding the effectiveness of these filters... :^) 

Interesting Stuff!! 

Best regards,

-- Bert -- 
-- 
Bert Hickman
Stoneridge Engineering
Email:    bert.hickman-at-aquila-dot-net
Web Site: http://www.teslamania-dot-com



Tesla list wrote:
> 
> Original poster: "Terry Fritz" <twftesla-at-qwest-dot-net>
> 
> Hi All,
> 
> I tested an NST for high frequency winding stress.  The data is at:
> 
> http://hot-streamer-dot-com/TeslaCoils/MyPapers/NSTWindingStress/NSTWindingStres
> s.html
> 
> That is supposed to be one line.
> 
> It is also the first paper at:
> 
> http://hot-streamer-dot-com/TeslaCoils/MyPapers/MyPapers.htm
> 
> Perhaps we need to rethink the use of NST filters...
> 
> Many thanks to Billy for the NST too!!
> 
> Cheers,
> 
>         Terry

-- 
Bert Hickman
Stoneridge Engineering
Email:    bert.hickman-at-aquila-dot-net
Web Site: http://www.teslamania-dot-com