# 30 kV 10 mA power supply

Discussion in 'Electronic Design' started by booth, Nov 6, 2006.

1. ### boothGuest

Hi All,
I need to design a 30 kV 10 mA capacitor Power Supply. Any Idea which
method is the best. How will it be isolated and what would it size be.
Many thanks for any hints.

Why?

3. ### Guest

Peter Baxandall devised his class-D oscillator in the first instance to

drive a nasty capacitative load. It looks quite like a Royer inverter,
but there is a capacitor across one of the transformer windings, which
-
with the inductance of the winding - forms an LC oscillator whose
resonant frequency determines the frequency of oscillation. The output
is a sine wave - more or less.

Jim Williams of Linear Technology rediscovered this circuit some years
ago, but - with a fine disregard for historical reality, describes it
as
Royer inverter. It is written up in Linear application notes AN49, AN55

and AN65

http://www.linear.com/pdf/an65f.pdf

His target application is back-lighting panels for the LCD screens of
portable computers, but his circuit generates the sort of voltage that
you want.

10mA at 30kV is 300 watt, so you may have to scale everything up a bit.

We've posted quite a lot on this circuit from time to time - search

4. ### Ancient_HackerGuest

Those voltage and current levels are kinda tricky to reach. AT 30KV
you're going to get lots of corona, unless you immerse everything in
insulating oil. The corona quickly eats away at any insulation, and
makes lots of nasty ozone.

If weight is no object, go buy a hefty neon-sign transformer, a bunch
of 5KV diodes, and 15 2Kv 0.1 uF capacitors.

If you want to minimize size and cost and weight use a flyback
transformer from a projection TV and a hefty oscillator. Many of these
run at 27KV and can probably be pushed to 30KV.

5. ### Mike HarrisonGuest

And parallel them up for the required current - MUCH easier to let someone else do the tricky HV
engineering for you.

6. ### meGuest

All the methods seem to be equally fine for your application.

I have to advise against any isolation to avoid any safety compliance.

It should fit comfortably in any bread box sized enclosure.

7. ### Guest

That's actually rather similar to a video monitor/television HV power
supply; you can get lots of standard components and design
examples if you use a flyback switching PS for this.

Cheapest in production quantities might be to use a TV type
tripler rectifier with a simple flyback inductor (the ones for TV have
lots of addons for sweep linearization and focus). The old
way to do this was with a Cockroft-Walton capacitor/rectifier
array, but output impedance was an issue.

8. ### Frithiof Andreas JensenGuest

The TV transformer/rectifier kits will not deliver 300 W on the EHT side; they
are rated for maybe 30 W.

Ignition coils for cars are *much* better. I once designed a relatively simple
resonant power supply for a small RADAR kit using a standard four-cylinder
ignition coil to get 10 kV, 60 W - bigger coils exist. Frequency has to be low -
i.e. 20-30 kHz - anyway because capacitance at those voltages absolutely kills
performance; hence the resonant design - an EHT transformer *is going to ring*
so we might as well use that to ease the di/dt requirements on everything.

Maybe one could "simply" build a six-phase unit with a six-phase Cockroft &
Walton tripler for the 30 kV, 300 W?

Another way - if O.P. wants his own transformer/rectifier stack - is to lay out
secondary windings as discs or single layer with intermediate rectifiers wound
in the same direction so that the voltage swing between windings is limited and
the voltage to the core is mainly DC.

For 30 kV - I think one needs Oil (or lots & lots of room). SF6 would do also
;-)

9. ### Winfield HillGuest

Homer J Simpson wrote...
Yes, good question.

15kV or 30kV supplies aren't trivial. Is this a one off?
Then we ask, can you get by with +15kV and -15kV to make
a nice balanced 30kV, with ground at the midpoint? If so,
you can use two of the elegantly-designed Spellman 15kV
MP-series 10W modules (i.e., 660uA max), one each of the
positive and negative types. You can use these to create
a programmable power supply with sub-uA current monitoring,
so you can evaluate corona discharge, etc., and fix it.

The 15kV MP modules are available on eBay for about \$46
each from w_gohar, e.g., these positive and negative ones,
http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=300045907267
http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=300045907144

10. ### neon

1,325
0
Oct 21, 2006
Let me see where can you find a power source like that cars have it TV have it. why don't you look at those items to get ideas.TV can only gives you 5 ma and cars only 180 mw per pulse. I hope that you know that this is a killing type of power you can only make one mistake there is no reprive or second chance.

Last edited: Nov 7, 2006
11. ### Guest

Oops, sorry, you'd need 30 of these modules for 10mA.

12. ### Guest

We buy our 30Kv 1 mA lab packaged supplies from Gamma High
Voltage,around 1500-2400\$ depending on options. do you really need 10
mA?

Steve Roberts

13. ### boothGuest

Thanks for all the answers. If the TV flyback transformer is not enough
for 30 kV 10 mA where can I find one. Or where can I find the suitable
ferrite cores.Any Idea?

Homer J Simpson yazdi:

14. ### Homer J SimpsonGuest

This is a dangerous supply. Why do you need it?

15. ### boothGuest

The application is a Capacitor Charging circuit and we are going to
need 20 devices in a year. Instead of buying one 30 kV 10 mA unit for
approx. 3K we prefer to built them. We also need to modify the standart
units. So because of price,quantity and modifications we decided to
built one. Thanks..

16. ### John GGuest

You did not really explain why you need it.
Charging a capacitor does not answer what you are doing with the charge.
Do you really understand how dangerous 300 watts at 30kv is and if you
charge a capacitor of any size (if you can find a 30kv capacitor) then the
potential (Pun Intended) for disaster is quite high.

John G

17. ### boothGuest

John G yazdi:
The capacitors are discharged using spark gaps. We have them all and
have completed several devices. That is no problem, It is simply a
medical application. We are aware of risks and have done all the
isolations in the system (using HV silicone rubber isolating compound).
Our problem right now is the generator design..

18. ### Homer J SimpsonGuest

What is the duty cycle of charge and discharge?

E.G. Charge once every 30 seconds, 10 seconds to get to full charge or .. .
.. .?

19. ### boothGuest

Homer J Simpson yazdi:
Discharge happens at a rate of 2Hz. we have 500msec for charge. 2
different capacitors are used, 40nF and 1.2uF.

20. ### Tom BruhnsGuest

Since you want to charge a capacitive load, it seems like you should
consider ways that do it quickly and efficiently. I believe it's
Linear Technology that's been advertising an IC lately that does a good
job charging photoflash caps. There are enough external components
that it can be tailored to a pretty wide range of energy (and voltage)
levels. There may well be other manufacturers that make similar
products.

Though a standard TV flyback transformer is unlikely to do the job for
you, the people who make them do know how to make high voltage
transformers. I would certainly look at getting transformers from a
company like that, that already understands the art, the technology.
The power level you're talking about is a bit more than, but not out of
line with, the power handled in the horizontal output transformer of a
large CRT display, though in a CRT display, much of the power goes into
horizontal deflection.

Also, have you considered resonant charging? Is that a possibility?
That's a lot more "gentle" on the supply than dumping an uncharged
1.2uF suddenly across a fixed-voltage supply. It sounds like the goal
is not to make 30kV, 10mA power supplies, but to charge capacitors.
10mA would charge the 1.2uF cap at about 8000V/second, so it would take
a bit under 4 seconds to charge it, assuming the supply is
current-limited at 10mA. Resonant charging could recharge it much
faster, though it takes a rather large choke, capable of storing about
135 joules in this case, to put 540 joules into the 1.2uF cap.

I'm curious: what's the peak current in the capacitor when it's
discharged? Is it a special capacitor that's rated for pulse discharge
service?

Sounds like you do have an understanding about working with high
voltages, but as others have done, I must add: the energy level
represented by 30kV on a 1.2uF capacitor is easily lethal, and that
voltage can jump long distances. Please be very careful working around
stuff like that! For the inevitable lurkers who are getting ideas