Hello,
The leakage inductance of the NST is high, as it is designed to limit the
current once the neon tube has ignited. This inductance will store and feed
back energy into the primary cap, so that the voltage over the cap could
theoretically reach twice the unloaded peak voltage of the NST. This will be
limited to a lower value by the spark gap ignition volage, the safety gap
and be depending of the resonance frequency of the primary cap / NST leakage
inductance combo. But it could very well be over the NST peak voltage, so I
would recommend designing the filter cap for a higher voltage.
If there is no spark gap at all limiting the voltage, it could reach a much
higher value because of chaotic ferroresonance, formed by the leakage
inductance / primary cap circuit. This is more likely if the resonance
frequency is close to the mains frequency, but once an oscillation has
started, the NST core will be saturated, which changes the inductance, so it
can resonate at several frequencies at once, in a chaotic state. If allowed
to go on, it will surely destroy the insulation in an instant. For that
reason it is prudent to design the resonant frequency mentioned away from
the mains frequency (the "LTR design" - lower than resonant), and to always
have a safety gap, which should not be to widely set.
The NST is rather fragile when it comes to voltage, as the unloaded max
voltage is only reached for an instant when the neon tube ignites, and then
settles for a much lower value, and the insulation is designed with this in
mind. That you blew the filter caps indicates that you have some degree of
resonant voltage rise in your charging design. Put in caps with more margin,
but also be a bit careful with the main spark gap and the safety gap with
respect to max voltage.
I would think that the resistors in the filter would also contribute to rob
a potential resonance of energy, perhaps that is a more important function
than the low pass filtering itself. Wire wound resistors will introduce yet
another inductance into the circuit, wchich theoretically could cause RF
ringing. I don´t know if that is of any importance, but I used a string of
metal film resistors when I built a NST coil long time ago.
Regards,
Jan
-----Ursprungligt meddelande-----
Från: Joshua Thomas <joshuafthomas@xxxxxxxxx>
Skickat: den 21 november 2021 01:43
Till: tcml@xxxxxxxxxxxxxxx
Ã?mne: [TCML] Terry/lowpass filter design, OBIT
Hello all,
I've been struggling to build a lowpass ("Terry") filter to protect my NST
from RF feedback. The first one suffered arc-overs because the capacitor
leads were too close together, and the second one had the capacitors fail -
apparently only 9kV per rail wasn't enough. It's a 12kV NST so I assumed
18kV > 12kVx(sqrt 2) and therefore would be sufficient.
Apparently not.
What suggestions are available for those who have made a NST filter that was
successful? Particular capacitor types/brands would be welcome, as well as
how the physical layout was accomplished. I use two 500ohm/100W wire-wound
resistors for the R half of the RC filter.
On a different topic I got a used OBIT for $35, but it appears to be a
solid-state one running pulsed DC. The model is Allanson 2275-628G. 17.5kV
"peak", 45mA, at 20khz. I mainly bought this for curiosity to see how it
might perform on a coil. I'm a little suspicious of the voltage and
amperage, as the total VA rating is only 87.5VA - which is a fraction of the
560VA that would be expected from 12.5kV RMS (17.5kV peak) at a full 45mA!
What experiences has anyone had with one of these?
Cheers,
Joshua Thomas
--
Joshua Thomas
My new email address is: joshuafthomas@xxxxxxxxx Please update your
information if you have not already done so.
_______________________________________________
Tesla mailing list -- tcml@xxxxxxxxxxxxxxx To unsubscribe send an email to
tcml-leave@xxxxxxxxxxxxxxx