Dan's 5MHz 60kV High Voltage Divider

50MHz Bandwidth 60kV High Voltage Divider

60kV High Voltage Divider with compensation box.

Introduction

Featured here is a 1000:1 60kV high voltage divider with a bandwidth of 50MHz. This high voltage divider has a high enough

bandwidth to cover almost any transient voltage waveform present in tesla coiling.

Theory of Operation

The function of a high voltage divider is simple. Connect a high voltage at the input, and read a corresponding scaled version of the input

voltage at the output of the divider which can then be safely measured via a multimeter or oscilloscope. The scaling factor of this particular

voltage divider is 1000:1. So for a 20kV input, you would get a 20V output signal. Seems simple enough. Just connect a bunch of large

value resistors in series and form a basic resistor divider. Does it work? Well, yes if you want to measure DC only! The problem is that

due to the parasitic lead-to-lead capacitances of the resistors, and distributed parallel capacitances of the resistors to ground ultimately kill

any type of AC response from the hv divider. For example, if you have a simple 1000Meg divider which has a 5pF parallel capacitance,

you can already see you will have trouble measuring waveforms even at 60Hz !

To correct this problem, we must also employ an effective capacitive divider scheme in parallel with our resistor divider network. At high

frequency, the capacitive divider works almost identically to its resistive counterpart. Cool. We now have a resistive divider to take care

of the low frequencies while we also have a capacitor divider incorporated into our hv divider to take care of the high frequencies. Not

quite so. The problem now is that if you looked at the frequency response of the divider, you will see that the frequency response is

flat and nice near the DC end and nice and flat at the high frequency end, but in the middle, the gain of the hv divider dips considerably.

To fix this, a compensation network needs to be designed so that the frequency response of the hv divider remains flat over the entire

bandwidth of measurement. And this is indeed the most difficult part of the hv divider design. One needs to deal with the problem of

how to go smoothly from a resistive divider at low frequencies to a capacitive divider at high frequencies, while keeping a flat gain across

the mid frequencies. The compensation network also needs to take into account the connections to and from the output of the divider

as well as the coaxial cable. Therefore, a dedicated coaxial cable needs to be used with the hv divider at all times.

Performance Testing

To verify operation of this high voltage divider, I decided to test it first on my relatively low voltage MiniKatz tesla coil as shown above.

4kV/38mA MiniKatz tesla coil in operation.

To verify operation of this high voltage divider, I decided to test it first on my relatively low voltage MiniKatz tesla coil as

shown above. I connected the top of the divider directly to the primary tank circuit of the tesla coil. The output of the

hv divider was then connected to my 2430A Digital Storage Oscilloscope where it was captured as shown below:

As one can see, I've captured the ring-down immediately following the discharge (via spark gap) of the MMC capacitor into

the primary tank circuit. The frequency of this ring-down oscillation is the resonant frequency of the primary tank circuit which

is approximately 471kHz.

Schematics and Electrical Design

The following schematics are the design for the 60kV high voltage divider. (Requires Adobe Acrobat 5.0)

PSPICE Simulations

More to come soon . . .

Last modified August 16, 2005 08:05:20 PM