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

RE: NST Charging Circuit Study

Terry -

Thank you for the test data. How do you find the time to do all the work?
The PFC problem is always difficult because of the varying load conditions.
The surging currents and distorted waveforms don't help. However, it is a
challenge worth trying. The question is how should the columns be set up for
the various NSTs?

My solution follows and is similar to yours.

Type of NST                                  9/30
Line amps                                    3.351
VA     120 * 3.351 =                         402.12
Real AC watts                                164.3
% Power Factor 164.3/402.12 = .4086 * 100 =   40.86
Phase angle degrees   arc cos(.4086) =        65.88
Var   402.12 * sin (65.88) =                 367.01
PFC cap   367.01/(6.823*60*120^2) =           67.60

ammeter current      402.12/120 =              3.351
wattmeter current    164.3/120 =               1.369
Var current          367.01/120 =              3.058

However, the 67.6 uf cap would not be correct for all 9/30 NST Tesla coils
because the load would vary depending on how well the TC system is adjusted.
For example, the load may be 9000*.03 = 270 watts
This would require a 54.89 uf PFC cap. If a 67.6 uf cap is used it would
give a leading current.

For the VAR I used the trig function sine. You used the pythagorean theorem.
There are several ways to solve the PFC problem.

Are the BPS and Isec mARMS important? This would show up in the changes of
currents? As you say it is messy.

I would be interested in some data on pigs with ballasts.

John Couture


-----Original Message-----
From: Tesla list [mailto:tesla-at-pupman-dot-com]
Sent: Wednesday, September 06, 2000 1:39 PM
To: tesla-at-pupman-dot-com
Subject: RE: NST Charging Circuit Study

Original poster: Terry Fritz <twftesla-at-uswest-dot-net>

Hi John,

	Here is a "test" run I did.  9000 Volt 30mA NST with a static gap.

Cpri		AC line		BPS	Isec mARMS	Real AC line
		Amps RMS				Power In
2.5nF		2.583		764	31.896		173.89
5nF		3.043		392	37.927		185.05
7.55nF		3.287		244	41.282		179.8
10nF		3.351		120	41.8		164.3
12.5nF		3.656		120	47.57		166.1
15nF		3.85		76	48.796		142.84
17.5nF		3.762		60	46.596		126.95
20nF		3.79		48	47.53		123.6
22.5nF		3.882		2	48.54		62.23
25nF		3.77		1	47.33		51.88
27.5nF		3.656		1	45.89		50.41
30nF		3.628		0	45.43		39.46

These are the "raw" numbers from which all kinds of stuff can be calculated.

The 1 and 2 BPS numbers are just the gap firing right after the initial
turn on surge.  The gap does not fire in steady state.

The resonant at around 5nF and the LTR values at around 10-12nF are clearly

Is this data OK?  I don't want to leave out anything but I don't want to
take more numbers than needed.  That static gap case is pretty easy but the
sync rotary case will need the timing optimized for each, which will be a
chore.  The real power actually does add up to the heat loss in the system
so it does work ;-))

For power factor at say 10 nF:

Real power = 164.3 watts
VA = 120 x 3.351 = 402 VA or "apparent power"
Power factor = Real power / Apparent power = 164.3 / 402 = 0.4087
VAR = SQRT(402^2 - 164.3^2) = 367.11 VAR  (that's sort of similar to using
inverse trig functions ;-))
ACOS (164.3 / 402) = 65.876 degrees lagging for a power factor of 0.4087
(hey.. that's not good! ;-))
Or 1.37 real amps and 3.351^2 - 1.37^2 = 3.058 amps reactive.  That would
take a 67.6uF cap to fix...

One can multiply the energy stored on the cap by the BPS to get energy
going into the system.  Of course, I am just trying to get the base numbers
and leave such calculations for later.

Note that MicroSim is very good a piecewise integration and analysis.  Thus
it can "measure" things which would be difficult in real life.  Real power
and true RMS are child's play for it even when the signals contain DC to
multi-megahertz or pulsed components.  I think I will take the steady state
values and eliminate to start up portion of the signals.  The turn on
portion is interesting but not valuable for this purpose.

So from the basic numbers the study gives all the fun stuff can be derived.
 If there is anything that needs to be added, let me know.

It should be noted that the waveforms are FAR FAR from nice sine waves.
MicroSim does the piecewise integration but care should be used in
generalizing to functions and definitions that were meant for sine waves.
The phase angle has no meaning in some of these simulations but the
"equivalent" average values do.

I have or am getting data for the following NSTs:

I guess 12/60 and 9/60 are "rare" ;-))

I will do both static gap and sync rotary 120 (BPS) calculations.  This
should pretty much cover the NST Tesla coil spectrum.  I could do OBIT too
if I can find the data...  I'll wait on the pig and PT stuff since others
may beat me to the wire there ;-))  The addition of variable ballast and
reactive vs resistive makes that really messy!

Ideas or comments welcome.



At 12:19 AM 9/6/00 -0700, you wrote:
>Terry -
>Note that to get the right answers you will need to set up MicroSim to use
>the trig functions. The actual line side currents would have to be known.
>This would give you the volt amps. The secondary volts and amps would give
>you the watts. You could then easily determine the VAR.
>John Couture
>-----Original Message-----
>From: Tesla list [mailto:tesla-at-pupman-dot-com]
>Sent: Tuesday, September 05, 2000 5:03 PM
>To: tesla-at-pupman-dot-com
>Subject: NST Charging Circuit Study
>Original poster: Terry Fritz <twftesla-at-uswest-dot-net>
>Hi All,
>	With all the recent talk of cap sizes, PFC caps, and break rates it would
>not be that hard just to crunch through all the variables in MicroSim and
>just come up with a set of tables.  However, I don't have all the
>transformer data.
>I need to know three things to model an NST:
>1.	Primary DC resistance.
>2. 	Secondary DC resistance.
>3.	Open load AC current draw at rated input voltage.
>The NST transformers that I need this data for are:
>and any others that people may have a use for.
>The resistances can be found with just about any ohmmeter.  The input
>resistance will probably be very low but close is fine.
>For the AC measurement, you will probably want to turn the voltage up and
>down slowly with a variac.  Just switching an NST on suddenly may cause a
>current surge that will pop the fuse in the multimeter.  Of course, beware
>the output voltage.  There must be nothing at all connected to the output
>terminals for this test.
>If anyone would like to take this data or you already know it, please send
>the info to:
>Do be very careful of the high voltage and AC voltage on the third test!!
>Thanks a bunch!,
>	Terry