Re: High-Bandwidth Primary Circuit Behavior (fwd)
---------- Forwarded message ----------
Date: Wed, 29 Apr 1998 18:37:07 -0600
To: Tesla List <tesla-at-pupman-dot-com>
Subject: Re: High-Bandwidth Primary Circuit Behavior
After carefully studying the data and running some numbers in excel,
it is obvious that the energy loss responsible for the linear decrement
waveforms in gapped LC circuits is due to the instability that occurs at the
zero current points in the primary system. The scenerio is as follows:
1. The gap is a good conductor until the current reaches the zero
level. At this point the gap shuts off into a high impedance state.
Although there is suffieceint voltage present to sustain the arc, there is
not enough current. The fact that the current through the arc must reverse
direction at this point is also a significant factor.
2. During the next 300ns (in my system) the impedances of the primary
wiring, self capacitance of the primary coil, etc. ring and have various
instabilites that consume energy from the system. There is gap heating,
certainly, but modeling suggests that the skin effect in the primary system
may also be burning up much energy. Power waveform studies suggest that
there are frequencies well beyond 40MHz in these bursts of energy.
Apparently, all the electical activity during this time is going towards the
waste of primary energy. Only when the gap again reaches a suffiecint level
can the arc be re-ignited and normal current flow restored the the system.
3. During the time of the gap's shut down, the current in the primary
system is not allowed to carry out the normal charging of the primary
capacitor. This accounts for the relatively constant reduction in the
primary peak voltage through the firing cycle. Careful inspection of the
voltage waveforms show that the normal decaying sine waves have their peaks
chopped at the times the gap is shut down indicating that the sinusoidal
charging has been interrupted (this actually occurs just after the voltage
peak). You can actually see the waves being reduced. Near the end of the
firing cycle, the waveform looks rather like a sawtooth wave do to this
effect. Apparently, the current level needed to re-ignite the gap is fairly
constant which accounts for the linear decrement. No doubt this current
level can be modified increasing effiecency with different designs.
This implies that the losses in the gap could be substantially
improved if the current needed to sustain the arc could be reduced. I
thought pumping RF energy into the gap may help sustain the arc but it seems
to have a tremenious amount of RF already! Early work with muti-gap spark
gaps suggest that they are very lossy but many people swear by them.
Perhaps there are factors in a full system with the secondary in place which
are yet to be found. I am building an oil-cooled gap that will allow more
study of all this.
Many questions and possibilities remain. But, of course, that is way we are
> I have been spending most of my time looking at primary circuits with
>my probe system. As a result, I have written another paper that tells about
>what I have seen so far.
>The HTML and the original Word97 documents are at:
>Sorry about the large files but I wanted to get all the fine details in
>the scope pictures and such.
>Hopefully, this paper will explain what is going on. There are a few loose
>ends but I think the basics are there.
>Many thanks to all of you who have helped out with suggestions and ideas.
>Comments are very welcome.