Very High Voltage Piezoelectric Power Supply - Digitaly Variable - Lightweight SMT / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

Very High Voltage Piezoelectric Power Supply - Digitaly Variable - Lightweight SMT

1 2 3 4 next

Move Thread

LAN_403

Uriah

Sun Jul 15 2007, 08:44AM

Registered Member #898 Joined: Sun Jul 15 2007, 07:50AM
Location:
Posts: 20

I am in the process of designing a lightweight power supply using all surface mount components. A driving circuit powers a very expensive Piezoelectric Transformer having a driving circuit pulsing halfwave DC at 65.5kHz. The driving voltage is 1.25V to 41V, while the current is adjusted with the voltage to be the only constant. The AC output (75V-2460V) is rectified and stored in a capacitor. A special IGBT is used to discharge the capacitive discharge circuit at a variable frequency of 1-100Hz.

My problem is stepping the voltage up from the Piezoelectric Transformer output. I need a final output of around 50kV and at a relatively high current from the capacitive discharge circuit. Because of numerous considerations a cascade multiplier type circuit would be WAY TOO BIG AND HEAVY. It is basically impossible to make a multiplier of this type operate at such a low frequency.

...unless I can find some 50kV capacitors with high values...

I need suggestions on how to solve my problem please. This is for aerospace (air-born) application so weight is one of the biggest considerations. The second of course is a reliable system with a long life span.

This does mean that a wire wound transformer will not fit the application in any way, as they are bulky and heavy and often un-reliable systems during constant operation.

Attached here is the schematic for the circuit without the secondary circuit to which this threadâ€™s question addresses.

PARTS LIST:
C1 - 0.1Î¼F
C2 - 1Î¼F
C3 - 10pF to 1nF 5kV protective isolation capacitor
C4 - (presently undefined value) 5kV discharge capacitor
D1 - 100V rectifier diode (regulator protection)
D2, D3 - SM3F 3kV rectifier diode
D4 - Freewheeling diode
R1 - 240Î©
R2 - 200Î©
S1 - NPN
S2 - MOSFET - IRF9Z24NS
S3 - IGBT - IXEL40N400
78M05B - 5V Positive Fixed Voltage Regulator
OM7646 - 1.25V-57V Positive Adjustable Voltage Regulator
AD8400 (R-8 package) Digital Potentiometer

The max dissipation of the piezoelectric transformer (PT) is 7 watts, but because of the capacitive discharge circuit the output is much greater then the input. (65.5kHz is the resonant frequency of the PT)

PT INPUT: 6.88watts - 1.25V to 41V @ 0.168amps (constant current when voltage is adjusted)

PT OUTPUT: 2.46kV @ 0.0028amps

I do not know how to calculate the capacitive discharge output (dependant of frequency), but I think that is all the information that I know at this time.
Optimally, I am looking for a state of the art device or unique circuit which will serve the purpose for my application and the necessary requirements.

(Note: the rest of the circuit is approx 4x4x.5 inches using surface mount components and tight packaging)

Regards,

Uriah George
**link**

...

Sun Jul 15 2007, 09:18AM

Registered Member #56 Joined: Thu Feb 09 2006, 05:02AM
Location: Southern Califorina, USA
Posts: 2445

I think that for an application like this, you really would be better off with a conventional smps supply... Say you wind a transfomer to give 10kv and use a few doubler stages to get the required 50kv. Since you only seem to need a few watts of power, it should be trivial to use a small ferrite core... You would mainly be limimted by the number of layers of insulation... But I bet that you could make a transformer weighing >10g capable of giving sufficient power--and now you have a nce cw source of hf juice. Heck, you best bet might be to just get a ccfl supply 'out of the box' and add a few doubler stages on it.

Steve Conner

Sun Jul 15 2007, 09:49AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Putting a multiplier after the IGBT switch won't work. It's not possible to C-W multiply a single pulse. So, you have to put the multiplier before the IGBT switch, where it will multiply the 65kHz AC, and now you need a 50kV IGBT. That obviously won't work. I see you got a 4kV IGBT, but you won't get a 50kV one.

In a situation like this I'd do the IGBT switching and energy storage at 600V or whatever, and use a custom pulse transformer to generate the 50kV output.

If you can't stand to use a pulse transformer, I'd say your best remaining option is some sort of solid-state Marx arrangement with IGBT or SIDAC switches. A Marx generator of course charges its capacitors in parallel and discharges them to the load in series. So with a 12 stage Marx you could charge at 4kV and pulse 48kV to the load, without needing any switch rated over 4kV. Greg Leyh helped design a huge device like this to generate pulses for a particle accelerator.

thedatastream

Sun Jul 15 2007, 03:11PM

Registered Member #505 Joined: Sun Nov 19 2006, 06:42PM
Location: Yorkshire!
Posts: 329

Reading between the lines, are you saying that the application requires a repeated pulse discharge of a 50kV capacitor into your load (whatever that might be - would be useful/interesting to know)? Reliably switching 50kV is no easy feat, especially if you have space constraints (no spark gaps!).

Can you post some more info about the application? There's not really enough in your first post to allow us to offer what could be more constructive solutions. Size and weight constraints, application, relaibility and servicing requirements, etc. Any info you can give will hlp us to help you.

Rgds,
James

Marko

Sun Jul 15 2007, 03:16PM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

I'm not trying to be a bastard here, but you really need to give us some more information if you expect help?

You didn't mention most important things like energy and peak power you need to achieve.
Knowing what is going to be final load is also going to be very helpful.

If you are driving externally controlled pulsed load, input power is not much of a meaning.

Only existing device that could actually switch your CW multiplier is a triggered spark gap.
Now, if your load is very low impedance you are going to have very high peak current wich will destroy the diodes.

Steve's marx recommendation would do much better in that case.

The another problem you put out is, that 50kV isn't a trivial voltage. If you are going to seal such a circuit into a small 1cm wide box you again need to assure that there are no conductive objects several centimeters in vicinity of the box which may provoke arcing, and in turn making the box so thin is meaningless.

The weight of the circuit will, mostly, be directly proportional to energy you want to store in capacitors. Electrolytic capacitors have higher energy density than other types of caps, but are limited to low voltages (<450V).

For described size of the circuit, using a small ferrite transformer is not a problem, but piezo is surely more compact and hassle-saving if you can obtain it.

It should also be driven by a bridge, and not in flyback mode as you drew. PWM regulation would get you rid of the linear regulator I see no purpose in.

Sulaiman

Sun Jul 15 2007, 05:08PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

Your piezo transformer has a very high effective output impedance requirement;
2.46kV / 2.8 mA = 878 kOhm.
SO quite low value capacitors could be used for a C-W multiplier (2.8 pF = 878 kOhm @ 65.5 khz)
I'm sure you could find some small 100pF 7 kV capacitors.
The problem then would be what to use as a switch.
I guess the entire 50 kV section will need to be potted due to low atmospheric pressure, potentially below the dew-point.

To me the most likely solution does seem to be a suitable pulse transformer for direct conversion
from dc bus voltage to 50 kV direct.

Uriah

Sun Jul 15 2007, 06:05PM

Registered Member #898 Joined: Sun Jul 15 2007, 07:50AM
Location:
Posts: 20

You guys are great! Thank you all so much for the replies so far.

I am working on plasma propulsion methods, such as U of Tennesse, NASA, Atmosphere Glow, and several other organizations have been working on in the past few years. I am sorry, but the most I can give you right now is a CAD rendering and a brief description. I am headed to the hospital because of a foot injury and not in the best mind frame to explain my project.

Firkragg, can you please explain a little about this...

"It should also be driven by a bridge, and not in flyback mode as you drew. PWM regulation would get you rid of the linear regulator I see no purpose in."

I was trying to get the message accross that my peak load is very high. Bringing the frequency down to below 100Hz will bring the current way up. Can someone explain how to calculate this please.

So a marx generator is sounding better and better all the time.

I'll get back on later after the hospitol. Thanks again to everyone.

Regards,
Uriah

Sulaiman

Sun Jul 15 2007, 07:20PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

Before going too far with the Marx development,
I think that you should consider emi/rfi interference suppression/containment
since it's airborne and there will be other sensitive equipment about,
Marx generators are extremely impulsive.

Steve Ward

Sun Jul 15 2007, 09:20PM

Registered Member #146 Joined: Sun Feb 12 2006, 04:21AM
Location: Austin Tx
Posts: 1055

Before going too far with the Marx development,
I think that you should consider emi/rfi interference suppression/containment
since it's airborne and there will be other sensitive equipment about,
Marx generators are extremely impulsive.

Why would a marx generator produce any more EMI than any other method of generating the 50kV pulse. They will all have similar dv/dt, so how does the EM field know what kind of circuit generated it?

If you need high reliability, you are better off with a properly designed pulse transformer as others have mentioned. Ive built a type of "marx generator" out of mosfets (to boost the input by 4X) and it was not only a pain in the ass to drive all of those MOSFETs but it also seemed unreliable. I think a pulse transformer will be smaller, and more reliable but it has to be properly made (vacuum impregnation, and proper interwinding insulation).

Sulaiman

Sun Jul 15 2007, 10:00PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

I wasn't thinking so much of the output dv/dt etc. as that can be tailored to suit,
more of the parasitic oscillations of each individual spark gap. (vhf uhf etc.)

1 2 3 4 next

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