Forums

Any other details?

Input voltage

output current

Adjustable over what range

load characteristics, etc.

Sorry, it needs to run from a +24Vdc source.  Output current 1/2 milliAmp.  Adjustable from as near zero as reasonablly possible to -60kVdc.  The output voltage will ultimately be controlled by an analog voltage in the 0 to +10V range (possibly 0 - 5).  The load is an x-ray tube so it basically looks like a resistor (i.e. no negative resistance region).

Thanks for your suggestions.

Hi Max,

That is a pretty ambitious project!  My HV experience only goes to around 6KV, so I can only take you so far.

For my part, I might recommend a full bridge, fixed frequency, phase shifted, series resonant converter.  Some sort of resonant converter will be required as the secondary capacitance is reflected onto the primary by the square of the turns ratio.  That can be daunting to deal with, so charging it up and down via a resonant inductor will be much more efficient.

The UCC2895 controller would be worth considering - link is below.

http://focus.ti.com/docs/prod/folders/print/ucc2895.html

Some thoughts that might help you along the way:

• Run the controller in voltage mode for shortest minimum on time.
• Keep the operating frequency reasonably low to help deal with the reflected capacitance.
• Lower frequency also helps to keep minimum on time small as a percentage of the cycle.  That will allow operation as near zero output volts as possible.
• Transformer design, of course, will be critical.  Hopefully you already have a source, as only a limited number of vendors have this experience.  A good mil/aerospace house would be a place to start.
• To get to 60KV will likely require multiple stacked secondaries and/or multipliers.
• Place the L and C in series with the transformer primary
• Cap will need to be metallized polypropylene to handle the ripple current
• Inductor should be made with low perm, low loss powdered iron, such as Micrometals -2 material or similar.  I'm not picking vendors or favorites, just one example that I know of.
• Consider antiparallel Schottky diodes across the MOSFETs to minimize switching losses, and conduction losses during dead time.

Anyway, those are my thoughts.  Let me know if they make sense or if you had something else in mind.

Dave