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

Re: Optimal Quenching Tests



Tesla List wrote:
> 
> > Subject: Optimal Quenching Tests
> 
> Subscriber: hullr-at-whitlock-dot-com Mon Jan  6 22:19:21 1997
> Date: Mon, 06 Jan 1997 16:55:17 -0800
> From: Richard Hull <hullr-at-whitlock-dot-com>
> To: tesla-at-pupman-dot-com
> Subject: Re: Optimal Quenching Tests
> 
> Tesla List wrote:
> >
> > Subscriber: bert.hickman-at-aquila-dot-com Sat Jan  4 21:54:49 1997
> > Date: Sat, 04 Jan 1997 19:34:26 -0800
> > From: Bert Hickman <bert.hickman-at-aquila-dot-com>
> > To: tesla-at-pupman-dot-com
> > Subject: Optimal Quenching Tests
> >
> > Hi all!
> >
> > I had a few more thoughts on quenching after answering John Freau's post
> > today.
> >
> > The bottom line:
> > Existing methods for estimating proper quenching time predict
> > excessively long quench times.
> >
> > Why:
> > We can define "ideal quench" as the point at which we have transferred
> > all the energy we can from the primary to the secondary. We let the
> > one-way primary->secondary transfer go to completion, but prevent the
> > reverse from happening. However, quenching too early leaves some energy
> > stranded in the primary.
> 
> Actually, by quenching too early there is no energy actually in the
> primary circuit at all, it is open.  The magnetic energy collapsing back
> in on the primary is just lost in space.  It never goes back to ES energy
> in the resonant capacitor.  

I should have said the primary tank circuit. If we quench too early,
we'll "strand" electrostatic (ES) energy in the tank cap, although the
primary elecrtomagnetic (EM) energy may indeed be 0. I suspect its
virtually impossible to quench if the primary current is significantly >
0. But it IS possible to prematurely quench at a primary current
zero-crossing. Any primary EM energy that hasn't yet been transferred to
the secondary or yet dissipated will wind up reverse-charging the tank
cap _as long as the gap conducts_ to the next zero-crossing. Unless we
quench near the first notch, some orginal "bang" energy will remain in
the tank cap. 

> The energy collapsing back into the resonator
> does indeed find use and rings up the secondary quite normally, but with
> less energy than if the collapsing magnetic energy had been regenerated
> in the primary system. 

I think we differ on the mechanism. Ring-up of the secondary reflects
increasing secondary energy levels - the lower the "k", the longer the
secondary "fill" time. Any increase in secondary energy can _only_ come
from one place: the primary. While an oscillating secondary
current/voltage is essential for reactive energy storage, it cannot
account for further increases in secondary energy during ring-up. This
additional secondary energy only comes through a series of "pushes" from
the primary. Each half-cycle "push" transfers additional energy to the
secondary until the primary has no more left to give (i.e., the first
"notch"), as long as we don't quench too early!

> As a matter of fact, the energy for this single
> first arc is the most magnetic energy that the system will ever see in a
> single sweep even if allowed to ring for the proper quench period. All
> other primary magnetic sweeps would be of lesser energy.  It is the
> secondary which can't fully utilize this inital huge magnetic energy
> pulse, requiring a "fill time" to achieve maximum output. A sort of
> inductive inertia, if you will.
> 
I agree. However, if only a portion of the primary energy can transfer
to the secondary during each half-cycle, the remainder self-couples back
into the primary, where the induced voltage then reverse-charges the
tank cap, less losses. "Bang" energy is conserved, not irretrievably
lost. I've examined this process in theory, with an analog storage
scope, and have modelled it via PSPICE - there are no inconsistencies
between theory and experiment. 

I looked at the total secondary ES+EM energy (1/2LsIs^2 + 1/2CsVs^2) in
a PSPICE model. The secondary energy climbs ("fills") during each "push"
from the primary. However, these pushes are NOT of equal amplitude!
Surprisingly, the 1st push (the biggest primary current swing) actually
transfers relatively _little_ energy into the secondary! However, the
next one(s) pick up the pace, and the final one(s) level off as we
approach minimal primary energy. The number of pushes required is
directly related to the coupling coefficient ("k") and the rate of
system losses (primarily the gap, and secondary streamers/arcs). If you
simultaneously look at the
the total primary (EM+ES) tank energy, there's a direct one-to-one
correspondance in declining rimary energy as it gets "pushed" into the
secondary. Total system energy (Primary + Secondary) is conserved and
constant, less gap and resistive losses.

Using a PSPICE model of my 10" coil, and plugging in various "magic" k
values, I simulated quenching at various primary current zero crossings
before and after the first primary notch. A simple 8 Ohm resistance was
used to model the series gap (based upon earlier primary high-power Q
measurements). The following chart reflects the relative portion of EM +
ES energy transferred to the secondary at each primary current
zero-crossing versus the best-case quench (100%). Note, in particular,
the relatively low portion of energy transferred at the 1st zero
crossing:

 Quench at                 % of Max Secondary
Primary Current          Energy at time of quench:
Zero Crossing    k=0.28  k=0.22  k=0.18  k=0.153  k=0.133
=============    ======  ======  ======  =======  =======
    T= 0            0%      0%     0%      0%       0%
    1st            27%     16%    15%     12%      10%
    2nd            74%     58%    44%     38%      33%
    3rd           100%     89%    77%     67%      60%
    4th            99%    100%    97%     90%      82%
    5th            87%     87%   100%    100%      97%
    6th                    59%    88%     99%     100%
    7th                           66%     88%      98%
    8th                                            86%

<SNIP> 

> As I get the H2 Thyratron on line for variable dwell,  I should be able
> to tell whether attempts to calculate the dwell time are close or not for
> the production of longest spark.  I am assuming you are probably in the
> ball park, though.
> 
> Richard Hull, TCBOR

Although the thyratron may be limited to quenching only at primary
current
zero crossings (barring external current-shunting/snubbing), this is
probably all we'll ever really want to do! And the control potential of
precise quenching is still awesome, particulary for making repeatable
measurements!! Looking forward to your results!

Safe coilin' to you!

-- Bert --