[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: NST's shorting out.
Original poster: Vardan <vardan01@xxxxxxxxxxxxxxxxxxxxxxx>
Hi Martin,
At 03:47 PM 3/20/2006, you wrote:
These are follow up questions are for Terry regarding your email:
If you doubled the number of NSTs without
changing the primary capacitor size, it is very
possible that the primary will go resonant and
over voltage the NSTs. 12kV at 120mA has a
resonant capacitor size of 27.3nF. If you are
running 30nF (from below), then you are right on
top of the resonant value. If nothing stops it,
the voltage might skyrocket to say 80kV!!!
Two follow up questions here. Is there a
difference in the resonant cap size between
running a "modified" single NST (removing some
shunts and running only 15-20 sec to prevent
overheating) at 120 ma versus running 2 NST's in
parallel where you have two cores
operating? Does the Java program take this into account?
When you remove the shunts, the current rating
goes up. So you may have say a 15kV 80mA
NST. Page four of this formula blurb has the equations for various cases:
http://hot-streamer.com/temp/FormulasForTeslaCoils.pdf
This chart has the standard ones and you can sort
of interpolate the values too:
http://hot-streamer.com/temp/MMCcapSales.gif
As you basically increase the NST's power, you
should also increase the primary capacitance. It
does not have to be really "exact" or anything,
but it should be sort of kinda close.
I think you can input any current of an NST in
the program. You can measure the current of the
modified NST with an AC ammeter but be very very
careful!). Don't go near the meter during the
test and be double sure to set it up
right. Else, the 15kV will cook the meter real good!
Second, what if I run at 180 ma ?
The primary resonant capacitor size would now
be 0.0398 and my smaller .03 cap would be
considerably below that value. Would this be
less likely to cause the resonance voltage spikes?
I don't know right off. It "sounds" a little
risky. It might be small enough to limit the voltage but I am not at all sure.
A rotary gap that is not a synchronous type is
not recommended at all. It will randomly fire at
any spot on the AC cycle. A sync gap is twice as effiecient:
Besides the efficiency difference, is there any
other reason to get rid of the async gap? Do
you think it is contributing to the kickback?
If it fires a lot, it "should" keep the voltage
down. But there are a lot of details than can
affect things so it is hard to give a general
answer. If you don't want to modify the motor
and all, I would make a simple static gap rather
than use an async gap. Others who have used
async gaps with NSTs might know more here.
It was a significant improvement to my static
gap ( SG wasn't air quenched, though) so I
really hate to get rid of it just yet. A
synchronous gap is not on the horizon right now.
Maybe have a fairly good static gap in parrallel
with the async. Or, just have the safety gaps
properly set. But the safety gaps might overheat if they fire a lot.
As far as another approach to eliminating the
primary resonance, what if I add an extra turn
to the primary coil? Then the system becomes
detuned even further, with the reference cap
size going down from .026 to .021. Would this
make the system less prone to these resonant spikes? (Java run below):
Only the NST and the primary cap determine max
resonate voltage. So changing the primary coil
will not affect anything there. The 60Hz
resonance is between the NST and the
capacitor. You need to either load the NST with
so much capacitance that it will not go over
voltage or be sure some spark gap will fire
anytime and everytime the voltage starts to get to high.
Primary Outputs:
96.81 kHz = Primary Resonant Frequency
15.87 % high = Percent Detuned
28 deg° = Angle of Primary
68.43 ft = Length of Wire
0.499 inch = Average spacing between turns (edge to edge)
2.5 inch = Primary to Secondary Clearance
90.086 µH = Ldc-Low Frequency Inductance
0.02123 µF = Cap size needed with Primary L (reference)
373.534 µH = Lm-Mutual Inductance
0.2 k = Coupling Coefficient
5 = Number of half cycles for energy transfer at K
25.17 µs = Time for total energy transfer (ideal quench time)
Or alternatively, if I go away from the (nice)
Maxwell .03 mfd pulse capacitor I am using, I
have three other caps, each 10 kv: 0.5, 0.3 and
0.3 mfd which I could run hook up in series
giving me a total of .115 mfd (but only 30 kv
total- I worry about that). Reducing the
primary turns to 5 would make this system:
That is a lot of capacitance. I would think the
load would be far too great then and the shunts
would just saturate (another fascinating subject)
causing a loud "buzz" from the NST. Best to try
to really get the caps matched up right.
Primary Outputs:
106.37 kHz = Primary Resonant Frequency
7.72 % high = Percent Detuned
48 deg° = Angle of Primary
27.48 ft = Length of Wire
0.753 inch = Average spacing between turns (edge to edge)
2.5 inch = Primary to Secondary Clearance
19.468 µH = Ldc-Low Frequency Inductance
0.09792 µF = Cap size needed with Primary L (reference)
184.788 µH = Lm-Mutual Inductance
0.212 k = Coupling Coefficient
4.72 = Number of half cycles for energy transfer at K
21.55 µs = Time for total energy transfer (ideal quench time)
I am concerned that these Chicago Condenser caps
are not as well suited to Tesla Coil
application, but is this the better approach?
Get MMCs like everyone else ;-)) You might be
able to just get a smaller value to add to your
30nF cap and save some money. It all sort of
depends on how you can mix and match it all together.
Cheers,
Terry
Many thanks for your help, Terry.
Martin Fryml