[Prev][Next][Index][Thread]

Re: Unexplained arcing current

>Original Poster: "B**2" <bensonbd-at-erols-dot-com> 

>    1.  Placing a resistor between the bypass capacitor and the
>transformer secondary winding decreases the Q of the RLC circuit
>formed by the secondary winding inductance, resistance, and the
>capacitance added by the bypass capacitor.  It also absorbs the energy
>from the parasitic capacitance of the HV winding as it attempts to
>discharge through the spark gap

In either an RCR network, or a (2R)C network (where the total resistance
is the same or 2*R), the discharge path for the parasitic capacitance is
through a resistance of 2*R, as is the series damping resistance, for both
cases.  No difference that I can see.


>>Through simulation, with reasonable component and parasitic values,
>>I see no evidence of NST secondary oscillations at the gap closures.
>>Can you suggest a circuit that may show this?"

>    2.  Use a current probe and an analog scope.

I value my scope too much to hook it up to an operating coil.  The
simulation models for NST's have so far been shown to be accurate.  If
there is indeed something going on that is not being properly modeled, I
would very much like to correct that.  Having a model that accurately
predicts operation is the best way to truely understand things and design
the best possible protection network.

>>the amount of energy dissipated by the NST secondary capacitance (small
>>compared to the bypass cap values) is insignificant, not enough to
>>carbonize the tar.

>3.  My previous measurements indicated around 2 nF for a 15Kv, 60 mA
>transformer from one bushing to ground.  I just now measured 700 pF
>from a 12 kV, 30 mA France transformer from one bushing to ground
>(This transformer has been unpotted, ultrasonically cleaned, coated
>with red glyptal, and reassembled for 12 kV, 50 mA service).
>That is a lot of energy wasted and a good argument for a DCTC!
>    I have dissected several neons.  In one case (Jefferson 15 kV, 60
>mA) the short was between the fine wire connecting the secondary to
>the bushing and the metal
>transformer casing.  In the other (France 12 kV, 30 mA) the short was
>between the same wire and the transformer laminations.  Neither
>transformer had a field grading bushing at its output.  In time, with
>any kind of abuse,  it will streamer through the tar and short due to
>the high field enhancement from the wire.

2 nF per bushing?  My gut tells me that's high, but then my gut's never
been calibrated to read capacitance, and I've not actually ever tried to
measure this.  Is it possible to measure this without disconnecting the
core connection on each secondary winding?

Assuming the parasitic capacitance really is 2nF per bushing or 1 nF
bushing-bushing, a static gap setting of 25KV, and a break rate of
240BPS, the power lost due to this capacitance is:
240 * .5 * 1E-9 * 25E3**2 = 75 Watts.  This power would be 
proportionately dissipated in the resistance of the secondary windings
and external resistors.  With my 15/60 NST windings each measuring about
2.3K and my external R's being 1.6K on each side, about 44 Watts would be
dissipated in the resistance of the two secondary windings.  Even so,
being dissipated over such a large mass, I wouldn't expect this to
promote a carbon-tracking breakdown.

>>There is no doubt that a resonant voltage rise occurs when caps are
>>added in parallel to the NST secondary.  This is only evidant however if
>>the circuit is allowed to ring over multiple half-cycles, something that
>>cannot happen with a reasonable static gap, and has no bearing on
>>parasitic oscillations or power dissipated in the resistors."

>4.  A static gap fires several times during each half of the sine
>wave.  When it does you get the energy stored in the bypass and
>secondary parasitic capacitances dumped into the gap through the
>resistor.  The two capacitances are charged and discharged many times
>during the 4 ms period of the 60 Hz half wave, depending upon how wide
>your gap setting is.   I measured ringing current spikes almost twice
>as great as the 60 Hz current waveform off the 12 kV, 50 mA France
>this morning.

Peak pulse currents are high, agreed.  But I maintain that with
reasonably small bypass and parasitic capacitance values and break rates,
the RMS current from the caps is not that significant. 


>>I'm not sure I see the difference, as far as what current pulses the
>>NST secondary sees,  between an R-C-R network, and an 2*R-C network.
>>With the RCR network, the parasitic capacitance of the secondary is
>>discharged by a resistance of R+R through the spark gap.  With an RC
>>network, assuming the total resistance is the same, the discharge current
>>is the same as well.  With reasonable break rates,

>5.  The R between the NST and bypass capacitor slows down its
>charging.  This helps to reduce the high frequency ringing. See (5)
>above.

Again, I don't see how operation of an RCR network differs from an RC
network if the total resistance is the same.  The charging rate of the
bypass cap is not affected by the R's, it is limited by the tank cap
charging rate.  The DIScharge rate is limited by the R's, but it wouldn't
matter if it is from an RC or RCR network.

Regards, Gary Lau
Waltham, MA USA