Re: Full wave Cockcroft-Walton voltage multiplier
rp181, Sun Nov 15 2009, 02:24AM

very nice construction, the glass tube looks great! where did you get the tube?

Re: Full wave Cockcroft-Walton voltage multiplier
Mads Barnkob, Sun Nov 15 2009, 09:26AM

Looks good, clean construction and good results ;)

How do you find the performance of those capacitors? They look like some I bought long back from a china ebay'er, was wondering if you got yours from a similar place.

Re: Full wave Cockcroft-Walton voltage multiplier
RogerInOhio, Sun Nov 15 2009, 10:53AM

very nice construction, the glass tube looks great! where did you get the tube?

I got the glass tube at an arts and craft store called Hobby Lobby. It was meant to be a flower vase and it is closed at what is now the top so I had to drill a hole for the output electrode. The bottom is sealed to the base with silicon rubber caulking.

How do you find the performance of those capacitors?

You are right. The capacitors are from China and I got them off of Ebay. They work just fine.

Re: Full wave Cockcroft-Walton voltage multiplier
Proud Mary, Sun Nov 15 2009, 12:34PM

Hello Roger, do you know what the sag under load is?

Re: Full wave Cockcroft-Walton voltage multiplier
brtaman, Sun Nov 15 2009, 12:54PM

Great work, roger. The construction is top notch, looks as if it was built commercially! The gold hue of the transformer oil really sets it off.

Re: Full wave Cockcroft-Walton voltage multiplier
RogerInOhio, Sun Nov 15 2009, 02:23PM

Thanks guys for all the complements.

Hello Roger, do you know what the sag under load is?

I don't have any figures but I can say it sags a lot. If you connect something to it that has a lot of sharp points the corona losses are so great that the voltage drops down to a fraction of what it normally is. The girl in the bottom picture doesn't have too many sharp points on her (most of her points are nice and round) and if you where to try to touch her hand you would get hit by a three or four inch long spark.
I would say that if you where going to make something like this to drive an x ray tube or a partial accelerator you would want to design it to have a lot less sag than what I'm getting.

Re: Full wave Cockcroft-Walton voltage multiplier
Electroholic, Sun Nov 15 2009, 05:29PM

how fast are the diodes? maybe pumping it at higher frequency will help?

Re: Full wave Cockcroft-Walton voltage multiplier
Proud Mary, Sun Nov 15 2009, 06:43PM

Hello Roger,

if you could specify V(in), f, C, and number of stages, performance under load can be easily estimated, (i.e. before corona and other leakage losses)

Stella

Re: Full wave Cockcroft-Walton voltage multiplier
RogerInOhio, Sun Nov 15 2009, 10:00PM

how fast are the diodes? maybe pumping it at higher frequency will help?

The number on the diodes is 2CL2FM and the recovery time is 100ns. I suspect that is fast enough that I could increase the pumping frequency quite a bit and I might want to do that sometime but it seems a little hard to do. Do you have any suggestions on how I might raise the frequency?

if you could specify V(in), f, C, and number of stages,

Frequency is 60 hz voltage in is 7500vac , 5 stages first stage uses .003uf caps. The other stages use .002uf caps.

Re: Full wave Cockcroft-Walton voltage multiplier
GeordieBoy, Sun Nov 15 2009, 11:47PM

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,

Re: Full wave Cockcroft-Walton voltage multiplier
Proud Mary, Thu Nov 19 2009, 09:35PM

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.

Re: Full wave Cockcroft-Walton voltage multiplier
wylie, Fri Nov 20 2009, 03:43AM

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?

Re: Full wave Cockcroft-Walton voltage multiplier
teravolt, Fri Nov 20 2009, 05:59AM

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

Re: Full wave Cockcroft-Walton voltage multiplier
RogerInOhio, Sat Nov 21 2009, 01:03AM

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.

Re: Full wave Cockcroft-Walton voltage multiplier
Proud Mary, Sat Nov 21 2009, 02:42AM

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

Re: Full wave Cockcroft-Walton voltage multiplier
RogerInOhio, Sat Nov 21 2009, 12:05PM

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.

Re: Full wave Cockcroft-Walton voltage multiplier
Proud Mary, Sat Nov 21 2009, 03:26PM


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.

Re: Full wave Cockcroft-Walton voltage multiplier
Steve Conner, Sat Nov 21 2009, 03:49PM


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.

Re: Full wave Cockcroft-Walton voltage multiplier
Proud Mary, Sat Nov 21 2009, 05:37PM


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)

Re: Full wave Cockcroft-Walton voltage multiplier
RogerInOhio, Sun Nov 22 2009, 11:56AM

Thanks for all the replies on possible failure modes. It has been an interesting and thought provoking discussion.
I'm happy with how this thing has turned out but as is often the case with these things I have a craving for more power. Most people that build these use things like flyback transformers that run on higher frequencies and I was wondering if there would be any chance that the newer inverter type neon sign transformers would work for this. Probably not I guess since it may be more like a florescent light ballast that puts out a high voltage pulse to start the tube and then runs at a much lower voltage. I have never heard of anyone doing anything useful with one of these.

Roger

Re: Full wave Cockcroft-Walton voltage multiplier
Arcstarter, Sun Nov 22 2009, 06:53PM

Yes, inverter NST's will work fine. As far as i know they are rated for a constant output. I had a small one, 12v powered, and it was a constant output, no resonant capacitor to give a high starting voltage/ballast it. Most CFL supplies seem to use a current fed royer, and a resonant capacitor which gives the quick burst of voltage to ionize the tube, and then once there is a load the freq shifts and it is no longer resonant, so it ballasts.

Not to mention, someone from here made a small multiplier with a little inverter NST. I think it was a 4kv model.

Re: Full wave Cockcroft-Walton voltage multiplier
kimbomba, Wed Aug 25 2021, 04:57PM

If I want a negative voltage output all I need to do is to invert the diodes right?