Development of a Saturable Inductor Pulse stage.

[Peterae]

Time to build an HV pulser stage for my Controller.

scrubbed my previous post, back to the drawing board.

[Peterae]

OK found quiet a good paper here on a saturable inductor chain design.
http://pps.coe.kumamoto-u.ac.jp/streaming/PulsedPower/generator/system1.htm

So i can use decreasing inductance along the chain with the same capacitance and achieve 100% energy transfer with a decreasing pulse width but the peak voltage will remain the same on the output pulse.

The other way is to have the same inductance but decreasing capacitance, this will give me increasing peak voltages and the decreasing of the pulse width with an efficiency of 3.6% in the example provided.

Looks to me if this was ever to be used to produce an efficient pulse driving system, then i can only use design a) which means i cannot use the saturable inductor chain for voltage multiplication with width sharpening.

EDIT
Spheric would never have used saturable inductors to pulse his coils in voltage magnification mode because he would have too much energy dissipated in the saturable chain, this still doesn't rule out Spherics use though as he was after fast pulses and this would do that efficiently.

I have a lot of trouble believing this chain would have been used above 500V because of the insulation problems associated with winding enamelled wire onto the cores.

My dilemma is whether this is a  viable way forward for me, i have fast pulses, what i need is high voltage, even if i did need faster pulses to get a reduction from 100nS to 10nS would require the first inductor to be wound with 100 times the inductance of the final stage, OK we could settle for a times 4 decrease, but then why go to all this trouble.

[Grumpy]

spherics may have mentioned the saturable reactor as a possible way, but never actually tried it.  Some of the stuff he says about the actual construction of SM TPU's doesn't seem accurate.  He may have actually used other methods but didn't want to discuss them out of fear.

Avalanche transistors work well if you can keep them working.

[muDped]

OK found quiet a good paper here on a saturable inductor chain design.

http://pps.coe.kumamoto-u.ac.jp/streaming/PulsedPower/generator/system1.htm

Very interesting tutorial link.  Many thanks Peterae!

For any who may be interested in further study of this excellent
material, the Main Page of the Pulsed Power e-book is found

HERE

[Peterae]

Cheers Dumped, yes very handy and easy to understand site.

G well done getting the Avalanche stage going, i am not sold enough on it's robustness or current switching ability

Anyone ever tried a cascode stage
http://www.pes.ee.ethz.ch/uploads/tx_ethpublications/biela_SicSwitch_PMC08.pdf

[the_big_m_in_ok]

http://www.pes.ee.ethz.ch/uploads/tx_ethpublications/biela_SicSwitch_PMC08.pdf

@Peterae
Would these be somewhat similar?

http://www.google.com/patents/about?id=TGkaAAAAEBAJ&dq=cascade+inverter&as_drrb_ap=q&as_minm_ap=0&as_miny_ap=&as_maxm_ap=0&as_maxy_ap=&as_drrb_is=q&as_minm_is=0&as_miny_is=&as_maxm_is=0&as_maxy_is=&num=100

http://www.google.com/patents?id=oC0EAAAAEBAJ&printsec=abstract&zoom=4#v=onepage&q&f=false

--Lee

[Grumpy]

Cheers Dumped, yes very handy and easy to understand site.

G well done getting the Avalanche stage going, i am not sold enough on it's robustness or current switching ability

Anyone ever tried a cascode stage
http://www.pes.ee.ethz.ch/uploads/tx_ethpublications/biela_SicSwitch_PMC08.pdf

don't need a lot of current and 2kv is enough

[Peterae]

the_big_m_in_ok

I cannot seem to find a HV cascode amplifier circuits to show you, you can buy chips with 3 stages built in for driving RGB cathode ray tubes, they work at 200v and can have very fast rise and fall times, but we are trying to switch 1kv upwards


anyway i keep tripping over G's avalanche stage LOL, i better have another go at one.

[Peterae]

OK Plan B
Marx Avalanche Stage taken from a paper Grumpy posted some time ago.
Grumpys Thread http://www.overunityresearch.com/index.php?topic=378.0

and the paper http://icecube.wisc.edu/~kitamura/NK/Flasher_Board/Useful/research/RSI00308.pdf

I wonder if i could connect up 10 of these stages instead of the 4 used below, if so it would give a fast 9kv Pulse

[Grumpy]

I'd try a lower voltage first and work out any bugs.  Keep the current as lows as possible.

EDIT:

I have found that triggering the first transistor in the stack makes for a "crack the whip" effect through the other transistors.

[Peterae]

Hi G

The good thing about the Marx version is there are only 3 transistors in each stack, so i can test at 900V with 3 transistors and gradually add the stacks in 3's each time adding 900V to the output pulse.

[Grumpy]

Marx allows you to use 900v and multiply that, but requires additonal caps and resistors.  Nice versatility.

Same number of transistors either way.  Come down to what sort of PS you have.

[Peterae]

I like the Marx config, i can built stackable boards, 1kv each board.

I discovered i can make my own HV caps, by taking a piece of 1mm enamelled and winding 0.5mm enamelled axially down it's length, i could connect my cap meter to the ends and wind away watching the capacitance go up, about 4cm i had 80pF. 

[BEP]

I like the Marx config, i can built stackable boards, 1kv each board.

I discovered i can make my own HV caps, by taking a piece of 1mm enamelled and winding 0.5mm enamelled axially down it's length, i could connect my cap meter to the ends and wind away watching the capacitance go up, about 4cm i had 80pF. 

Do you have any problems with dielectric breakdown of the conductor insulation?

Side note: Your home brew capacitor should also make a good linear delay line by using four connections instead of two.

[Peterae]

Hi WW
I hope that the Shellac on both pieces of wire will be good for what i am using.

Your home brew capacitor should also make a good linear delay line by using four connections instead of two

Indeed i am not yet sure how long gives what delay

[the_big_m_in_ok]

the_big_m_in_ok

I cannot seem to find a HV cascode amplifier circuits to show you, you can buy chips with 3 stages built in for driving RGB cathode ray tubes, they work at 200v and can have very fast rise and fall times, but we are trying to switch 1kv upwards


anyway i keep tripping over G's avalanche stage LOL, i better have another go at one.

Peterae,
Finally got back to this thread.  This time I'll remember to 'check' the Notify and Return to this thread boxes.

I imagine the devices you need will be pricey in any event---to put it mildly.  Have you tried an auction of surplus Gov't or (especially) Military materiel?

This ploy may still be a long shot, though.  I was in the US military and was a R&D electronics technician for awhile.  Specialized components may be hard to come by.   I wind my own coils by hand myself.

Luck,

--Lee