Full wave Cockcroft-Walton voltage multiplier / Projects / Forums

Forums

4hv.org :: Forums :: Projects

Next topic »

Full wave Cockcroft-Walton voltage multiplier

previous 1 2 3 next

Move Thread

LAN_403

GeordieBoy

Sun Nov 15 2009, 11:47PM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

The problem with increasing the drive frequency to the NST is that the leakage inductance caused by the current shunts is going to result in a higher reactance at a higher frequency. So as you increase the frequency the available charging current will fall. e.g. At 240Hz, your 15kV/30mA NST becomes a 15kV/7.5mA transformer

-Richie,

Proud Mary

Thu Nov 19 2009, 09:35PM

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

I must agree with everyone else it looks really cute, but I'm a little concerned about the seeming lack of anti-surge resistors in the circuit - once one element goes down it'll run a like a ladder in a stocking through the whole lot, as I'm sure many here will have found to their cost.

wylie

Fri Nov 20 2009, 03:43AM

Registered Member #882 Joined: Sat Jul 07 2007, 04:32AM
Location:
Posts: 103

The gold hue of the transformer oil really sets it off.

That's what gets me too. Maybe add an LED to the bottom to really Sci-Fi it up? It's great without it though, so don't mess with success :)

(most of her points are nice and round)

More proof might be required ;) On a serious note, shouldn't hair present alot of charged points for corona losses?

teravolt

Fri Nov 20 2009, 05:59AM

Registered Member #195 Joined: Fri Feb 17 2006, 08:27PM
Location: Berkeley, ca.
Posts: 1111

If you want more output you could use a flyback circuit witch runs at about 15khz at what ever voltage you need. If you want fullwave you would need two identical flyback transformers. a mazzilli circuit has the moast performance. hear is a link

RogerInOhio

Sat Nov 21 2009, 01:03AM

Registered Member #1034 Joined: Sat Sept 29 2007, 12:50PM
Location: Chillicothe, Ohio
Posts: 154

I must agree with everyone else it looks really cute, but I'm a little concerned about the seeming lack of anti-surge resistors in the circuit - once one element goes down it'll run a like a ladder in a stocking through the whole lot, as I'm sure many here will have found to their cost.

That would be bad! The diodes in this thing where the biggest expense at $130.00 but so far I haven't had any problems.

That's what gets me too. Maybe add an LED to the bottom to really Sci-Fi it up?

I thought about that too and I still might do it.

shouldn't hair present alot of charged points for corona losses?

That's a very interesting point. It turns out that her hair is an extremely poor conductor of electricity. Otherwise I think it would bleed the charge down a lot. I know that if I put a small brush made out of copper wire on the electrode it bleeds the charge down so much that I can only get about an inch of spark but you can really feel the charge in the air that some people describe as an "ion wind".

If you want more output you could use a flyback circuit witch runs at about 15khz

Thanks for the link. I might consider doing something like that some time.

Proud Mary

Sat Nov 21 2009, 02:42AM

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

Hi Roger, as you are really only dealing in uA, the best retro-safeguard you can do is to increase the size of your output 'safety' resistor.

For a bigger splash, and so as not to undo all the hard work you have done, why not terminate your C&W with say 100M/75kV and use that to charge up a secondary capacitor, so if you have a dead short the storage, smoothing capoacitor will take the strain.

Obviously, asking direct sparks and arcs between HT+ and Earth will destroy almost any C&W given enough time, so you should always look to the needs of your application first, and then calculate all your component values after.

Your pal, Stella.

Errata deleted

RogerInOhio

Sat Nov 21 2009, 12:05PM

Registered Member #1034 Joined: Sat Sept 29 2007, 12:50PM
Location: Chillicothe, Ohio
Posts: 154

the best retro-safeguard you can do is to increase the size of your output 'safety' resistor.

I think the 1.5 mega ohms I have should be more than sufficient to protect the diodes. I'm guessing that the output voltage is about 100,000 volts so the max current would be about 66 ma. The diodes are rated for 100ma.

Proud Mary

Sat Nov 21 2009, 03:26PM

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

RogerInOhio wrote ...

the best retro-safeguard you can do is to increase the size of your output 'safety' resistor.

Have you tried running it under load continuously for 100 hours to see if you can get some empirical insights into the probable Mean Time Between Failures?

Don't me wrong, laddie, you've made a truly remarkable piece of kit there - the best looking C&W I've yet seen in my years at 4HV, and I'm sure all will second that. You've made something to be proud of.

But is there a possibility that while you've thought about shock loads on the output, what will happen if you get a really aggressive voltage spike coming in from the bottom?

I tried to make a simple analogy by talking of the way in which a ladder 'runs' in a pair of tights, but we could also think about it
by saying that what is true for n is also true for n + 1, such that if you blow out a single diode the entire circuit will go down like a row of dominoes.

Steve Conner

Sat Nov 21 2009, 03:49PM

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

I like that! We should call it "destruction by induction".

The failure mode with CWs, as far as I know, is as follows:

An arc between the top of the stack and the bottom forms an LC resonant circuit, where C is the capacitor stack and L is the inductance of the arc and wiring. Because this inductance is low, the impedance of the circuit is low and the resonant frequency is high.

Before the arc, the energy is in the capacitor stack. The arc then kicks the circuit into resonance. One-quarter cycle later, the stack voltage is zero, and the energy that was in it is now in the stray inductance. Because of the low impedance and high initial voltage, the current corresponding to this stored energy could be hundreds, maybe thousands of amps.

The oscillation then continues, with this high current trying to recharge the capacitors in the opposite polarity. The diodes are taken out as they try to clamp this reverse voltage.

The cure of course, is to add enough resistance in series with the CW's output that the resulting RLC circuit is overdamped. It's then a mathematical certainty that the voltage will never reverse. The resistance must be rated to stand the CW's full output voltage without arcing over.

Proud Mary

Sat Nov 21 2009, 05:37PM

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

Steve McConner wrote ...

Proud Mary wrote ...

I tried to make a simple analogy by talking of the way in which a ladder 'runs' in a pair of tights, but we could also think about it
by saying that what is true for n is also true for n + 1, such that if you blow out a single diode the entire circuit will go down like a row of dominoes.

I don't disagree with your theoretical analysis in general, Steve, but even very careful drawing board designs, will defy expectations with that dreadful crack near the top of the stack which tells you your diodes are kaput. This is no very big deal when SF6 is used, apart from the cost of replacement parts, but if someone has taken the trouble to pot in epoxy or what have you, then they are uggered, and in for financial and emotional loss.

As we all know, C&Ws aren't pulse power devices, so I'd go for resistive damping in every stage in the hope (not certainty) of limiting spark over damage.

Reverting back to my original query on the adequacy of the 1M5 output resistor, I realise that I had not fully made myself clear.
I'd agree with your 'adequate damping' analysis provided that the 1M5 resistor was correctly rated i.e. 1M5 141kV or what have you,
because with no other damping in the circuit diagram as illustrated, a single output flashover would banjax the entire effort under what I'll call 'critical circumstances' - the electronic equivalent of a spring tide occurring together with a full moon and a gale driving it cumulative, and someone forgetting to take the bath plug out, and so on. You may say
such an event is not likely to happen very often, but I'd say this lad has really put his heart into producing something special, and it would be uncharitable - false even - not to point out the vulnerabilities of the circuit as it is, whilst something can done to harden it up. Imagine if he was demonstrating it, as he surely will, and it all fizzled.

No doubt others will come forward to disagree, but I would always advise as good practice the de-rating of all C&W circuit elements by 66% at least for reliable continuous operation in SF6 or a good dielectric oil (dry kitchen- type sunflower) and by 80% or more in 'household' air.

I'd expect the output resistor to be a vitreous type rated at a notional 282kV (assuming Vout)max) is 100kV, (i.e. double PIV)

previous 1 2 3 next

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