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

Re: DC tesla coil



Original poster: "S&JY" <youngsters@xxxxxxxxxxxxxxx>

Antonio,

see comments below:

----- Original Message -----
From: "Tesla list" <tesla@xxxxxxxxxx>
To: <tesla@xxxxxxxxxx>
Sent: Thursday, January 06, 2005 11:25 AM
Subject: Re: DC tesla coil


> Original poster: "Antonio Carlos M. de Queiroz" <acmdq@xxxxxxxxxx> > > Tesla list wrote: > > > > Original poster: "S&JY" <youngsters@xxxxxxxxxxxxxxx> > > > The main problem with the setup described at the scopeboy site is if you > > don't run the RSG fast enough, you end up with flaming, trailing arcs in the > > RSG because the charging reactor saturates and stops acting like a current > > limiter, and you have a direct short across your power supply - not good. > > The way to eliminate that problem is to add another set of RSG stationary > > electrodes spaced so when one set is lined up with the rotor electrodes, the > > other set is spaced half way between the rotor electrodes. The RSG then is > > wired to resonant charge the capacitors through one set of gaps, then > > discharge the capacitor into the primary through the other set of > > electrodes. Works wonderfully well, and you can run your RSG as slow as you > > want with no diode popping power arcs. As a bonus, the charging reactor > > only needs to be around 0.2 Henry, easily achieved with 4 50 mH home made > > air core coils in series to withstand the voltage across it. > > An schematic diagram, and pictures of an actual setup, would help.

The schematic is so simple that it can be described in words:  I use a RSG
rotor with 12 pass-through electrodes, with stationary electrodes on either
side of the rotor.  Filtered DC power supply connects to reactor & de-Qing
diodes, then goes to one of the stationary electrode pairs.   The other
stationary electrode goes to the tank cap.  This resonant charges the tank
cap to roughly twice the DC supply voltage.  (I don't have any pictures
yet.)

As the rotor turns, another rotor electrode presents itself between a second
pair of electrodes.  One stationary electrode is also connected to the tank
cap, and the other one goes to the primary.  This allows the tank cap to
discharge into the primary, and the DC power supply is completely
disconnected from the primary circuit.  Then the cycle repeats itself.  Very
easy on the power supply (its output is never shorted), and no ballast is
needed.

> It looks
> quite strange to charge the primary capacitor through a RSG, since the
> contacts would have to be kept for long periods (possible, with fixed
> electrodes covering a wide angle).

I calculated the dwell time, at max rotor RPM, that the charging electrodes
would be close enough to sustain conduction through the plasma, then used
that time to calculate the charging reactor inductance to charge the tank
cap in that time.  It works--rotor electrode is 1/8 inch tungsten, but I
assumed electrodes would arc through a 1/4 inch path.  Your idea of a wider
stationary electrode might be useful, but I think my tank capacitor gets
charged to at least 80% of twice the DC supply voltage.

 The transformer (NST?) would not like
> to have its output current interrupted by an opening gap.

I don't use NSTs.  Frankly, I think standard NSTs are a poor (e.g.
unreliable) choice for powering any TC architecture since the NSTs are
designed to run mostly at less than 1,000 volts most of their life while
powering neon tubes.  If they are depotted and run in oil, then they are OK.
I use a 6 MOT stack in oil followed by a bridge rectifier and filter cap.
In either case, the transformer is busy keeping the filter capacitor
charged, so it isn't bothered by gap current interruptions.

> 0.2 H is quite a lot of inductance for an air-core coil. It would be
> quite bigger than the average secondary coil...

True if a single layer coil is used.  But I use 4 air-core inductors in
series to withstand the voltage.  Each is 1 inch ID, about 2.5 inches OD and
about an inch wide, wound with many layers of #23 wire to take advantage of
mutual inductance.  Thus the reactor is orders of magnitude smaller than a
TC secondary.  Actually, I am now using a charging reactor of about 0.14
Henry to charge a 26 nF 24 KV MMC tank cap.

It all really works.  It is nice to have independent control of variable
break rates and power supply voltages.  I get 6 foot arcs with about 1.2 KVA
from a 240 volt AC supply.  As far as I know, I am the only one using this
architecture, and I invite others to give it a try and report your results.

--Steve Y.




> > Antonio Carlos M. de Queiroz > > > >