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

Re: design questions








tesla-at-pupman-dot-com on 11.01.99 19:00:43

To:   tesla-at-pupman-dot-com
cc:    (bcc: Marco Denicolai/MARTIS)
Subject:  Re: design questions




>Original Poster: Bert Hickman <bert.hickman-at-aquila-dot-com>

>With a 19.9 kVRMS Pig running at full power, you may overvolt your 40 kV
>cap. The reason is that the voltage swing when the gap fires can be as
>much as the initial peak voltage (1.41*19.9kV) multiplied by 1.8
>(assumes that the voltage reversal is 80% of the initial cap voltage),
>or over 50,000 volts. And, the cap will see even higher voltage stresses
>if you use resonant charging and a rotary spark gap. Unfortunately, many
>coilers simply don't understand the voltage stresses that these caps
>actually see under TC use. More experienced coilers understand - usually
>by hard-won (and sometimes expensive) experience. Most pulse caps have
>faceplate ratings in volts DC. A tank cap driven off a 15 kV (RMS)
>supply should be rated at no less than 45 kVDC, and more conservatively
>at 60 kVDC to have at least some margin. Under full steam, your 19.9 kV
>system really should have a 60 - 80 kV pulse cap. In ther interim,
>you'll need to limit applied primary voltage (via the variac) and use a
>static gap across your RSG to insure that you don't overvolt and destroy
>your tank cap.
>
>-- Bert --

I am myself designing a tank supply with a 20 kVrms pig, but a 50 kVA one.
Modeling it with MicroSim I reached the same numbers.

Vpeak = 1.41 * 20000 = 28200 V

Using DC (not AC!) resonant charging actually you MORE than double that
figure because the resonant charging reactor gets already some startup
current when the tank supply is short circuited by the RSG. Let's say:

Vcpeak = 2.1 * Vpeak = 60 kV

Now, if you also suppose that the primary capacitor has left some voltage
Vr on it when the RSG opens (and this voltage is of opposite polarity to
the charging voltage), the charging reactor will sink more current into the
capacitor and will charge it to about:

Vc'peak = 2.1 * Vpeak + 2 * Vr !!!!

In reality we know that this Vr is very small so that should not represent
a problem.

Instead to oversize the capacitor to ? * 60 kV I tought to leave some
margin (probably use a 100 kV capacitor) and use a safety gap on it "just
in case..."