# Re: Sparklength inquiry

```to: Greg

Spark length is very dependent on pps (pulses per sec) or break rate
because the plasma has less time to cool off.  Initial phase of plasma
creation pumps a considerable amount of energy into the creation of light
so the luminous factor is important from a purely aesthetic valfue.  Power
is required to form the plasma (requires potential for ionization).  It's
the power in the secondary inductor not the potential that really
determines sparklength.  Up to now I'm sure you're with me.

Now enters the question that has been the subject on this List for some
time -- how to determine the optimum break rate.  It will vary with each
system, however, one can optimize the break rate or pps for each system.
Up till now everyone has been discussing the sec spark length, however, to
determine optimum break for any system another important factor must be
considered -- how many watt-sec of energy is lost in the primary spark gap.
In any large system, especially the one like yours, there is a tremendous
amount of energy wasted formation initial ionization and producing UV light
and lots of heat.  Attemping to measure the lost energy directly with
electrical instrumentation is very difficult, however, there is a much
easier way which we have used in the past in analysis of our systems to
determine optimum pps.

Remove the secondary inductor for any system.  We must assume you have a
rough idea of the load impedance your secondary inductor represents due to
o-scoping the waveform, looking for the 10 hz decay, and then determining
loaded dynamic Q factor as John Couture has suggested -- which is of course
directly related to pi.  In our experiments we made an assumption that
loaded operating impedance as seen by the primary inductor was around 30
Ohms (if one scopes the current in the primary and compares this value with
the sec in place to a parallel resistor of equal value you can set the two
impedance.  We then place the 30 Ohm "Glo-bar" type resistor in parallel
with the primary inductor so that it simulates the operating sec load.

Now comes the fun part -- instead of using electrical instruments to
measure the losses, one employs a bit of chemistry -- in particular
calorimetry -- to measure the losses.  This works very well because
watt-sec is directly equatable to calories of heat.  In this test, of
course, certain constants are required -- constant cap value, constant
freq, constant primary inductance, and constant average input power to the
system.  Resistance load in primary simulates the Q factor the sec coil
would be if it were in place. At 2 pi radians per cycle, # cycles x 2 pi =
dynamic loaded Q of system.  The idea here is to have all constants except
one -- the number of pps for your system.  This becomes the variable for
each test.

A thermistor (or thermocouple) is epoxied to the body of the resistor.  Use
a relay to drop the HV off quickly and within a 1/2 sec switch on the leads
from an analog meter to the thermistor or thermocouple to read the heat
value in the resistor.  Using a special sparkgap with "extra" stationary
electrodes that can be paced in series as necessary or using a speed
control on your rotor will allow you to take different heat values for each
pps combination.  I always feel the best way is to use a constant speed
motor and just vary the number of series breaks with jumpers as you can not
always be sure all terminals are firing -- one or two may skip which would
give you false assumptions.  The differential between these powers
represents the losses in the primary sparkgap.

Once the losses are determined for different configurations using
calorimetry one can make a few graphs which will illustrate the best
possible combination for each particular system tested.   You will note if
the freq is a constant there is a particular value formed which is the
product of the cap size x the pps (break rate).  The product of these two
form a constant which can be used to make optimum system determinations.

For smaller or medium size systems you can run the system in air and just
run the oscillator for 5-10 sec (be sure to accurately measure this time
value which must be equal for each test run).  For large systems we usually
drop the big resistor into a small plastic tank filled with xmfr oil and
run it for 10 sec.  The oil prevents overheating of the resistor.

The idea here is to carefully achieve a good balance between the break rate
which at higher power levels does produce longer sparks and the total
combined losses each individual spark gap represents.  The other variable
again becomes the product constant of the cap size and pulse rate.

I trust this information is of assistance to the discussion regarding break
rates and cap sizes.

DR.RESONANCE-at-next-wave-dot-net

----------
> From: Tesla List <tesla-at-pupman-dot-com>
> To: tesla-at-pupman-dot-com
> Subject: Re: Sparklength inquiry
> Date: Saturday, September 26, 1998 7:53 PM
>
> Original Poster: "L. Robertson" <LWRobertson-at-email.msn-dot-com>
>
>
> -----Original Message-----
> From: Tesla List <tesla-at-pupman-dot-com>
> To: tesla-at-pupman-dot-com <tesla-at-pupman-dot-com>
> Date: Thursday, September 24, 1998 5:37 PM
> Subject: Re: Sparklength inquiry
>
>
> Hi Greg
>
> The only time I see Vpri falling off with increasing BPS is after
> the charging system is maxed out. I guess any current limiting
> device, such as a resistor used for filtering, would produce
> similar results.
>
> LR
>
>
> >I am quite convinced that Vs is *not* an inverse function of BPS.
> >(unless the charging circuit is broken)
> >
> >But you might not believe my measurements.  So I should ask if
> >anyone else on the List has data from *actual operating coils*
> >which might indicate whether Vs, and therefore Vpri, varies as
> >an inverse function of BPS.
> >
> >--
> >
> >
> >-GL
> >www.lod-dot-org
> >
> >
> >
>
>
>

```