>Original Poster: "Marco Denicolai" <Marco.Denicolai-at-tellabs.fi> 
>I am designing a TC with a 62 kHz oscillating frequency. As I would like to
>quench at the first notch with a rotary gap, I supposed I can set my
>coupling coeff. so that the first notch will come after about 5
>oscillations at 62 kHz. That makes about 80 us to reach the first notch.
>So the quest is how to achieve such a short quenching time.
>From available literature I found that a spark can possibly form when the
>electric field exeeds 40 kV/cm and surely forms when it exeeds 70 kV/cm.
>So, if you write down how the stationary and rotating electrode distance
>evolves with time, and calculate the time the electric field is greater
>than 40 kV/cm, you should get a rough figure of your quench time.

You have described some required conditions for an arc to start, but to
stop an arc is a different matter, much more complex, and not completely
understood as far as I have read.  Remember the Jacob's Ladder, starting
and stopping fields are very different.

>I did that and for a peak voltage of 60 kV, rotating speed of 6000 rpm,
>rotor diameter of 12" I got a quench time of over 500 us (!). Because I
>cannot reasonably increase either speed or rotor diameter, I took the
>book's solution and designed to insert a static spark gap in series with
>the rotary one.

One thing often done to reduce an RSG's effective dwell time is arranging
stationary and rotating electrodes in pairs that are wired in series and
only come into alignment 2X or 4X or 8X (actually, any integer if you're
not concerned with sync operation) per revolution, effectively dividing
the actual dwell time by the number of electrode-pairs.

>By doing that I plan to decrease the voltage at the RSG down to 15 kV: that
>way the theoretical quench time drops down to 100 us, that gives at least
>some hope to achieve the required 80 us.
>I was thinking to use a high resistance voltage divider to have 15 kV on
>the RSG and 45 kV on the static gap. The static gap would be a series of
>sections, to easily set its trigger voltage. The voltage divider should
>also remove a part of the randomness involved in air gap trigger voltage.

I doubt that putting a resistive divider across the gaps would alter the
quench time.  When the gaps are conducting, the resistors are effectively
shorted out.  But then, quenching really involves whether the gap will 
_start_ to conduct following a zero-current-crossing, so I can't say for

>Any suggestions or comments about the above story?

I'm not an expert on quenching, I believe this is one of the topics that
is in need of research and worthy of investigation.  But assuming that
you're asking this in preparation for your PhD project, wouldn't a coil
of more moderate scope be equally useful in investigating the numerous
topics suggested by this List?

Gary Lau
Waltham, MA USA