Re: Spark Gap Sustaining Current (fwd)

["Tesla list" <tesla@xxxxxxxxxx>]

---------- Forwarded message ----------
Date: Wed, 03 Oct 2007 15:57:21 -0500
From: Crispy <crispy@xxxxxxxxxxx>
To: Tesla list <tesla@xxxxxxxxxx>
Subject: Re: Spark Gap Sustaining Current (fwd)

On Wed, 2007-10-03 at 07:59 -0600, Tesla list wrote:
> ---------- Forwarded message ----------
> Date: Tue, 02 Oct 2007 23:23:51 -0500
> From: Bert Hickman <bert.hickman@xxxxxxxxxx>
> To: Tesla list <tesla@xxxxxxxxxx>
> Subject: Re: Spark Gap Sustaining Current (fwd)
> 
> Hi Christopher,
> 
> Tesla list wrote:
> > ---------- Forwarded message ----------
> > Date: Sat, 29 Sep 2007 16:28:45 -0500
> > From: Crispy <crispy@xxxxxxxxxxx>
> > To: Tesla list <tesla@xxxxxxxxxx>
> > Subject: Re: Spark Gap Sustaining Current (fwd)
> > 
> > Thanks a lot, that helps me greatly.  Perhaps I should elaborate
> > somewhat on my end goal with all of this.
> > I'm trying to design and make a type of disruptive discharge Tesla coil
> > that should be able to create very long sparks with minimal power input.
> > Classic TCs have a single point of disruptive discharge with the primary
> > spark gap being the switching device.  The large peak power in this
> > discharge is what causes the large sparks as compared to a CW coil.
> > However, with the same basic primary circuit, spark length can be
> > further increased by increasing the bang rate, such as with an ARSG.  My
> > understanding of this is that the ionized trails off the topload have
> > less time to dissipate.  The problem is that to have many disruptive
> > discharges in rapid succession, the overall power draw increases
> > dramatically.
> 
> And, for a given bang size, overall spark length no longer increases 
> after you reach several hundred breaks per second... the discharges may 
> get hotter and more "frantic", but they get no longer.

Ah, Ok.  Why is this?  I guess there's another factor than what I
described, but what is this factor?

> 
> > When I was considering a way to fix this, I first started thinking about
> > DC Tesla coils.  The first idea I had (which has apparently been tried
> > once before) was to use a variation of an ARSG instead of the normal
> > ARSG with a large charging inductor.  This ARSG would essentially act
> > like a SPDT relay (instead of the SPST relay that normal spark gaps
> > operate with) with an oscillator connected to the coil.  It can be
> > connected in such a way that the charging supply is physically
> > disconnected from the primary circuit when the primary circuit starts
> > resonating.  Because of this, such a large charging inductor (with a
> > value in the Henries) is unnecessary, because the charging circuit will
> > never be shorted out.  A charging inductor of a much lower value can
> > still be used to limit charging current and to double the charging
> > voltage.
> 
> This should work. However, you'll want to use a suitably robust HV 
> "dequeing" diode in series with the charging inductor. Also, the ARSG 
> (during the "charging" sequence) should have sufficiently long "dwell 
> time" so as to charge the tank cap to full voltage. This should be a 
> MINIMUM of 1/2 cycle of the effective resonant frequency of the charging 
> circuit when the gap is operated at maximum break rate. The charging 
> circuit resonant frequency will be the series combination of the tank 
> cap, DC storage cap, and charging inductor. This mode of operation 
> allows the tank cap to sinusoidally swing up to full voltage (approx. 2X 
> Vsupply). Once the tank cap is fully charged, the dequeing diode 
> prevents any further current flow. After this point, the tank capacitor 
> charging current is zero, and you really shouldn't have to break any 
> "sustaining" current. If your gap has insufficient dwell time, you'll 
> waste some energy arcing in the charging gap - the relatively large 
> charging inductor will make trying to prematurely open the charging 
> circuit an exercise in futility.

Yes, I know that I need a deQ-ing diode.  The PIV should be about the
peak voltage of the power supply, and with proper overrating, I think a
string of 30 1N4007 diodes should work (which I have already built for
the purpose).  They are rated for 1A continuous current, but much more
peak.  With a charging inductor of about 25mH, I think this should work.
Also, because my charging inductor won't be very large, dwell time won't
be an issue (I've calculated it before and there's a significant
margin).

You seem to misunderstand exactly what gap I was referring to and what I
meant by sustaining current.  The gap to which I was referring was not
the gap that was part of the charging circuit in the rotary gap.  it is
a separate static gap would be in series with the charging circuit.  I
want this gap to fire at a certain voltage (approximately the peak
voltage of the electrolytic capacitor), but not quench until the voltage
is about half that.  Since arcs are sustained by current, my worry was
that the breaks in the charging circuit by the rotary gap would cause
this separate static gap to quench very quickly, maybe even after the
first firing, which would entirely defeat the purpose.  I want to keep
it firing for a number of rotations of the rotary gap, during which the
electrolytic cap will not charge to its peak voltage again.

> 
> > After I planned out a coil of this type, I started working on the power
> > supply for the coil (which is now finished).  It consists of a 12/30
> > NST, relevant filter circuit, a hv bridge rectifier made out of a total
> > of 120 1N4007 diodes, and a smoothing capacitor made with 50 450V 22uF
> > electrolytic capacitors in series.  
> 
> Sounds OK. The higher equivalent series resistance (ESR) of the 
> electrolytics may make the overall circuit a bit lossier during the 
> charging cycle. Also, watch that you aren't significantly exceeding the 
> ripple current rating on the electrolytics. It's possible to have many 
> amperes of charging current in a resonant charging system when using a 
> lower value of charging inductance. Your dequeing diode must also be 
> capable of handling the charging current and standoff voltage.\
> 
> > Relevant equalizing resistors and
> > protection diodes were added across the components.  To test this setup
> > for durability and functionality, I connected it up with a simple spark
> > gap in parallel with the smoothing capacitor.  As expected, the gap
> > periodically fired as the capacitor charged to the gap's breakdown
> > voltage.  Interestingly, the capacitors all appear to be completely
> > unharmed after a number of repetitions.  
> 
> The acid test will be how well the electrolytics handle the ripple 
> current at the desired break rate...

What's the acid test?

> 
> > While watching these
> > discharges, I came up with an idea - what if this same "smoothing"
> > capacitor was discharged in impulses into the whole primary circuit?
> 
> Directly firing the electrolytic bank into the primary would not be a 
> good idea. Electrolytics do NOT like the voltage reversals that occur 
> when "ringing" with a primary.

No, I do not want to directly fire the electrolytic capacitor into the
primary.  I know that electrolytics will not stand voltage reversals,
and it is only ever charged in one way.  The electrolytic is similar to
a DC smoothing capacitor, with the difference that it is periodically
discharged fully by the aforementioned static gap to provide
instantaneous power boosts.

> 
> There was a very clever circuit discussed a few years ago. A 
> comparatively large DC storage cap was used with spark gap version of an 
> H-bridge circuit. The tank cap and primary winding were connected in 
> series. Pairs of spark gaps then connected the tank cap (and primary) in 
> one direction across the DC storage cap, and then with the opposite 
> polarity during the next gap firing... etc.
> 
> Each time the a pair of gaps "fired" (pair A on one cycle, pair B the 
> next), it would disruptively reverse the charge on the tank cap, through 
> the primary. As with a DC resonant charging system, the tank cap would 
> "see" a voltage swing of 2 times Vsupply, and you get a bang each time 
> the gaps fired instead of the DC resonant circuit above which will 
> charge in one position and discharge in the other. Use a 
> non-proportional font (such as Courier) for the following ASCII-art 
> schematic:
> 
> 
>     + --------------o------o------------------------
>                     |      |                       |
>                     |    A X     Cp     Lp       B X
>       HV            |      |    | |                |
>     DC In         -----    o----| |--OOOOOOOO------o
>                   -----    |    | |                |
>                     |      |                       |
>                     |    B X                     A X
>                     |      |                       |
>     - --------------o------o------------------------
> 
> 
> > This way, the ARSG could be run at very high speeds to take the most
> > advantage of remaining ionized trails in the air, and the bang size
> > would be relatively constant over the short period of time that the
> > primary spark gap was firing.  The repetition rate for sparks visible to
> > the human eye would be slower than the normal ARSG speed or the 60Hz
> > main line feed in classic TCs, but the spark length should be
> > substantially longer.  In my setup (about half done so far), the
> > repetition rate should be very roughly around 5Hz by my calculations.
> 
> I'm not sure I understand what you are proposing. However, you'll get 
> longest streamers when using break rates of 100 Hz or more. At lower 
> break rates, previous spark channels cool down too much to permit 
> bang-to-bang streamer growth. This causes each bang to become pretty 
> much an isolated event, accompanied by individual, but short, sparks.

Yeah, apparently I wasn't clear, hopefully my responses above clear it
up.  I'll try to make a schematic.

> 
> > What I am currently considering is what method I should use for
> > discharging the "smoothing" capacitor (I put it in quotes because it no
> > longer really serves that purpose).  I've built a high voltage relay
> > based on 2 solenoids and tungsten contacts, but I was wondering if a
> > simple non-quenched static gap would function just as well.  The key
> > would be to keep the arc sustained between ARSG charging cycles -
> > otherwise the entire purpose is lost, and that is what my original
> > question was about.
> 
> Hmm... I'm having difficulty understanding exactly what you are 
> proposing. Can you generate a schematic?

Yeah, I'll try to get a schematic.

> 
> > I'm working on implementing this design right now, but it's going a bit
> > slowly since this is my first year at college and couldn't bring many of
> > my materials and testing systems with me.  Also, an aggravating aspect
> > is that the rules are so strict here that I can't even use a hacksaw
> > without jumping through a number of hoops.  I will of course let
> > everyone know how it goes.
> > Are there any thoughts on this design?
> > 
> > Christopher Breneman
> > 
> 
> Bert

Thanks a lot.