Best regulation for 5.5kV power supply / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

Best regulation for 5.5kV power supply

1 2 3 next

Move Thread

LAN_403

Turkey9

Mon Aug 29 2011, 12:06AM

Registered Member #1451 Joined: Wed Apr 23 2008, 03:48AM
Location: Boulder, Co
Posts: 661

Hello everyone,
I need to charge up a .4uF 6kV capacitor for use in a pulsed application. What would be the best way to build a well regulated supply that outputs 5.5kV? The supply needs to be as small as possible with preferably the smallest amount of noise generated to interfere with other circuits. I was thinking a small transformer that outputs 550V fed into a ten stage multiplier. Can multipliers be regulated with a feedback loop?

The capacitor will need to be charged around 100 times per second, so the supply will need to be around 150W.

I'm open to all suggestions. Size, power, and fairly good regulation are the only important factors.

Thanks!
Jesse

Dr. ISOTOP

Mon Aug 29 2011, 01:33AM

Registered Member #2919 Joined: Fri Jun 11 2010, 06:30PM
Location: Cambridge, MA
Posts: 652

Capacitor charging should be constant current. What you really want to do is have a current setpoint, and have the controller watch the capacitor voltage until it reaches the desired voltage, then shut it off.
Application?

Sulaiman

Mon Aug 29 2011, 04:17AM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3141

0.4uF @ 6kV = 7.2 Joules

x100 pps = 720 Watts at 100% efficiency......

About 150W is inexcusible!

Turkey9

Mon Aug 29 2011, 05:17AM

Registered Member #1451 Joined: Wed Apr 23 2008, 03:48AM
Location: Boulder, Co
Posts: 661

Sulaiman wrote ...

0.4uF @ 6kV = 7.2 Joules

x100 pps = 720 Watts at 100% efficiency......

About 150W is inexcusible!

Hahaha ooops! I have four caps that I will parallel. The power calculation was for one of the caps! Thanks for correcting me!

720W at 100% efficiency seems like a little too much for me. Lets decrease the discharge rate to around 20 pps to better match the 150W power I'm looking at.

What kind of topolgies can provide constant current? I know of SLR but that is it; can multipliers do this?

Dr. ISOTOP

Mon Aug 29 2011, 05:52AM

Registered Member #2919 Joined: Fri Jun 11 2010, 06:30PM
Location: Cambridge, MA
Posts: 652

Turkey9 wrote ...

Sulaiman wrote ...

0.4uF @ 6kV = 7.2 Joules

x100 pps = 720 Watts at 100% efficiency......

About 150W is inexcusible!

You can't PWM an SLR to vary current as it will throw off the soft-switching.
With that being said, 200W@400V is only half an amp, which is little enough so that you can use a hard-switched driver.

Shrad

Mon Aug 29 2011, 08:13AM

Registered Member #3215 Joined: Sun Sept 19 2010, 08:42PM
Location:
Posts: 780

what you could do to regulate voltage/current would be to use thyristors to phase-angle switch your cascade multiplier

this is an idea I had for a variable voltage HV PSU a couple weeks ago, as I have plenty of thyristor bricks and trigger transformers, only lack a big insulation transformer

Dr. Slack

Mon Aug 29 2011, 08:21AM

Registered Member #72 Joined: Thu Feb 09 2006, 08:29AM
Location: UK St. Albans
Posts: 1659

If you have a pulsed application (so no need for steady current) and rapid recharge operation (10s of pulses per second) then you might want to consider how the DC resonant TC charging system works.

For a 5.5kV on the final capacitor, you'd need a constant voltage supply of 2.8kV, which you could get just from a single recitified MOT.

Assuming that the capacitor starts each cycle with 0v on it (you need a constant starting voltage to get a constant final voltage) then, when you connect a charging inductor and series diode between your supply and the capacitor, current begins to flow. The entire charging process takes half a cycle of LC resonance between the charging inductor and the output capacitor. For the first quarter cycle, current builds up reaching a maximum when the output capacitor is equal to the power supply voltage. At this point, both the output cap and the inductor have the same stored energy, 25% of your final target energy. For the next quarter cycle, the inductor pumps its energy into the output cap while the current falls back to zero, ceasing when the output cap has reached double the supply voltage. At this point, the diode prevents the current reversing. Without it, the other half cycle would take place fully discharging the output cap and putting all the energy back into the power supply again.

This topolgy has a lot to recommend it for pulsed applications. The current starts and stops gently, you get inherent voltage doubling, you can use a very simple low impedance voltage source power supply and get a well-defined current into the load. Iron-ware for the charging inductor is readily available, and just suits the 20-100 shots per second speed range you have in mind. Use a few series flourescent lighting ballast chokes, or MOT secondaries on cut-open cores.

The major problem I can see with this particular application is the technology for the switch. The voltage is a bit low for efficient operation with a spark gap, though a very closely space RSG will work, and it's a bit high for silicon, you'd need to stack several IGBTs and take care of voltage sharing.

Pulse by pulse regulation is not really possible, though with a consistent starting voltage, and controlling your primary power supply, you could get consistent voltage on the output capacitor

But it's worth a thought.

Proud Mary

Mon Aug 29 2011, 11:02AM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

The resonant voltage doubler outlined by Dr Slack is widely known in thyratron modulator circuits.

1314613792 543 FT0 5c22 Thyratron Modulator

The charging choke 8 in series with D7 and D8 forms the voltage doubler. D7 and D8 are in series simply because silicon diodes of sufficient voltage rating were not available when this circuit was designed. The same applies to D9 and D10 in series with the resistor across the Rayleigh Line PFN, where they stop reverse voltage transients - swing - driving the thyratron anode negative, and mucking everything up.

An almost identical circuit will work well enough - subject to electrode wear - if a triggered gap or enclosed trigatron is substituted for the 5C22 thyratron.

To bring the circuit up to date, a series string of avalanche transistors can easily be used as the switch at 5.5kV - and el cheapo varieties like 2N5551 are perfectly suitable and very fast. There are plenty of suitable circuits about on the web. This Zetex application note isn't about 2N5551, but will give you some good ideas:

Here is a design for a 4 kV solid state Marx impulse generator using 2N5551 which you can easily modify. The pulses have a 1 ns or faster rise-time:

There is a super Rayleigh Line PFN java calculator at the bottom of this page here:

Pinky's Brain

Mon Aug 29 2011, 04:20PM

Registered Member #2901 Joined: Thu Jun 03 2010, 01:25PM
Location:
Posts: 837

Personally I'd go for a dual resonant charging circuit.

Their example on page 17 is pretty close to what you need. Except they used an aircore transformer to get the 0.6 coupling factor transformer they needed, air core transformers are rather complex, and with the power and voltage you need you're probably better off simply taking a high coupling factor transformer and putting an inductor in series to lower the coupling factor.

You only need regulation of the "low" voltage supply to get a known high voltage on the output capacitor.

Turkey9

Mon Aug 29 2011, 10:53PM

Registered Member #1451 Joined: Wed Apr 23 2008, 03:48AM
Location: Boulder, Co
Posts: 661

Thanks for all the replies!

A few questions:

With the resonant charger, could I connect multiple ones in series? The output capacitor from one would feed into the other. I imagine that the capacitance would have to decrease as more sections were added to preserve the total energy stored in each stage. I really like the resonant charger.

What size of inductor would I need? Will it need to be iron core?

It seems now the issue is generating the 2.5kV for the resonant charger. This could be a smps type of design, correct? As long as it can provide the current needed. I really would like to stay away from using a MOT as a power source as it is very big and bulky. However, it can be used as a step as I learn more of how to miniaturize everything.

1 2 3 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob