[Prev][Next][Index][Thread]
Re: Question about ballasting and VA
At 05:40 PM 3/3/00 -0700, you wrote:
>Original Poster: Terry Fritz <twftesla-at-uswest-dot-net>
>
>At 03:42 PM 03/03/2000 -0800, you wrote:
snip...
>
>Hi Jeff and All,
>
>In a reasonably designed (not all screwed up ;-)) coil, the various parts
>have been more or less impedance matched to each other (whether you know
>you did it or not ;-)). The output arc impedance is matched to the
>secondary system impedance. The secondary is matched to the primary, the
>primary is matched to the charging system....
>
>When you vary the current limiting, you are either raising or lowering the
>charging system's effective source impedance. Since the secondary is
>basically a fixed effective impedance at a given voltage and output arc
>level, there is indeed a "best" current to supply. Too little, and the
>system starves for power and tries to suck more current. Too much and the
>gap just starts to cook off heat. The impedance matching (almost SWR
>matching) and such really does apply here. Problem is, the system is
>nonlinear and very complex so massive computer modeling is needed to figure
>out exactly where to set the dial. That's why all of us in the "double E
>bunch" use MicroSim hour after hour and get big multiprocessor computers
>with so much RAM an eight digit calculator can't add it all up...
>
>However, you "other guys" just turn the dial to the "sweet spot" to do the
>some thing. Hmmmmmm ;-)))
>
>Cheers,
>
> Terry
> Genuine double E engineer
Terry
I'm not sure that the behaviour of the tank charging circuit is neccesarily
non-linear.
Something I've always wondered about is if one can consider an inductively
ballasted tank charging system to be a simple low-pass filter? e.g. if you
reflect the primary side L to the secondary side of the HV transformer and
consider it with the tank C, don't you just get a L-section type filter?
If this idea is correct, then the charging circuit will have a linear
response over the region we are interested in. When the inductance is lower
than a critical value, the current will only flow in short pulses and the
total power entering the tank will be small. Clearly the behaviour will be
non-linear in this region. But when the critical inductance is exceeded,
the flywheel effect of the inductance will cause the current to flow
continuously and the output voltage will only change slightly with changes
in the load resistance. Assuming the impedance matching point is greater
than the critical inductance, then predicting the filtering action under
this condition is straightforward.
Alternatively, since we are feeding linear components from a linear source,
shouldn't we expect to get a linear response?
Thinking out loud,
Gavin Hubbard