A 50kV supply with 33mA drive ability

This forum is for specialized infomation important to the construction and safe operation of the high voltage electrical supplies and related circuitry needed for fusor operation.
User avatar
Finn Hammer
Posts: 298
Joined: Sat Mar 05, 2016 7:21 am
Real name: Finn Hammer
Contact:

A 50kV supply with 33mA drive ability

Post by Finn Hammer »

It has been a while since I posted something here, so I guess I'd better not come empty handed..

Over the years there has been a major problem in the Fusor community: How do I get my hands on a suitable power supply?
I have felt the same restriction, so when the corona lockdown sent me home in March, I decided, that now is the time to get serious and have a poke at designing the driver for such a supply.

I don't think I should go into too much boring detail about the many setbacks I had during the development,
Only, show you a couple of steps on the way:

IMG_20201217_112915.jpg

And perhaps also a basket full of the some of the stuff that didn't make it to, but did lead to, the final product:

IMG_20201212_001249.jpg

So to cut to the chase: the supply consists of a SiC mosfet based LLC converter, on a 100mm by 80mm board.

The attachment IMG_20201212_001249.jpg is no longer available

Converter to the left, enpty board to the right, Mosfets on the back side.

I have successfully driven the driver beyond 2kW for prolonged periods, and it is thermally stable. This means that 60kV at 33mA is a viable possibility.

The driver drives a transformer with 2 pie basketwound coils of my own manufacture, 400 turns, primary has 24 turns, so with 320Volts in, and a bit of gain in the converter, I get 5kV out, into a CW multiplier. The leakage inductance in this transformer is high, and the magnetizing inductance is reduced by an air gap, to suit the LLC operating point.
Prototyping the multiplier was a challenge due to a need to figure a suitable derating of the diodes.
I had asked VMI for a quote, and their 25mA/25kV diodes costs 23USD a piece, which I find to be excessive, I would need 20 pcs. for a 50kV supply which ads up to 460 bucks.
Some chinese diodes, designated 2CL2FP, promises 30kV, 100mA with 100nS reverse recovery, however:
At 20mA they cave in between 90-100kHz, and at 33mA they stop at 75kHz, so I made a compromize to drive them at 50kHz and 33mA, which works fine under oil.
Oil is messy, though. Prototyping under oil is particularly messy:


The attachment IMG_20201211_122807.jpg is no longer available

Not to talk about the high risk operation of getting the parts cleaned in between iterations (just don't tell your wife):

The attachment IMG_20201211_123042.jpg is no longer available

I am using rapseed oil, but still.....


Anyway, I am working on a potted version, and I hope the thermally conductive potting material which is in the mail will provide sufficient cooling to the whimsy diodes, which are the weak part of the chain.

This is how the project stands right now:

The attachment IMG_20201217_104625.jpg is no longer available

The driver has 4 optically isolated inputs, one for a kill switch, one for a pause switch, one for a reset switch and finally one for an overtemperature sensor. Furthermore, it has a couple of 5V outputs to drive LM35 temperature sensors for the transformer core and an arduino, which will monitor the current in the CW ground lead. I hope to be able to sense a spike here if there is a spark, and then pause the supply for a brief moment, before letting it come up again with its soft start.
Internally, the driver has primary current sensing, so it is protected in as many ways as I could think of, and which was made possible by the Fairchild FAN 7631 driver chip, which is a fine piece of machinery.

A note on the LLC operating point: In a hard switched converter, it is normal to have to deal with some rather nasty ring-a-ding voltage transitions across the main switches, and this picture shows what that looks like, where the cyan trace is the voltage across the bottom switch:

The attachment noicyswitch.png is no longer available

However, by adding the right combination of leakage- and magnetizing inductance as well as drive frequency, it is easy to get to this level of excellence, with absolutely no overshoot and ringing. Quite a revelation.

IMG_20201212_001043.jpg

I should add, that the converter is fed by a variable transformer, and is running open ended, so the voltage regulation is the one you get from the internal impedance of the converter. On the other side, you get a fully variable output from 0 to whatever upper limit you chose to invest in, and there is a fine tune option by varying the drive frequency as well.

This is it for the moment, I still have not quite made up my mind about how far I should go (if at all), to protect my IP, but for now, hold your horses and wait for the fully assembled potted version.

Anyway, I guess it won't hurt to give the schematic. Nothing much to it, really but as usual, the devil is in the detail.
IMG_20201217_104625.jpg
Cheers, and happy hollidays to you all, Finn Hammer
Last edited by Finn Hammer on Wed Mar 02, 2022 8:08 am, edited 1 time in total.
User avatar
Richard Hull
Moderator
Posts: 14975
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: A 50kV supply with 33mA drive ability

Post by Richard Hull »

Finn, welcome back! Your high quality posts and superlative work have been missed!!

The high voltage supply is, indeed the holy grail in the fusion effort. Your efforts will be appreciated by all who are in tune with your effort. I always knew that with the demise of giant line operated HV transformers which are the ideal one stop solution of yesteryear, that only a modern electronic switched solution would do. Of late, the Chicom two and three flyback HV precipitator supplies have been nursed to their limits by careful hands. Most notable is the FAQ by Mark Rowley in this forum who has turned these rather low cost solutions into a workable system. Their output is limited, and as Mark notes, ham fisted operation will not work as they are not bullet proof and for our purposes are operated at the "bleeding edge". As I remember, just before you left us with the promise of, "I shall return", you were investigating this type of Chicom supply.

Now to your work. The diodes are a critical factor, as you note. As the world drifts towards operating electronics that has left the high voltage world of 5 volts to fully establish itself at 3 volts as its standard and yet is finding itself snapping at voltages just above the silly con junction's potential hill, the needs for high voltage, fast, moderately high current diodes is a backburner effort. Those semi manufacturers who dare to play in the rarified atmosphere above 20kv demand a suitable pound of flesh for their suitable diode product. Your report on the Chicom diodes you have settled on is unsettling. But one doesn't bread board in the unknown with 20 high priced top tier diodes. Then their is always that perineal custom transformer issue as we go from 60hz to 25khz. I fear we will always be tied to some "special" transformer regardless of frequency! The switching electronics is somewhat stock and beyond fighting inevitable kickback spikes is secure. Diodes and transformers remain the issues.

As regards to your return, you might do a bit of "catch up" reading of recent FAQs and postings in the various forums, if interested.

For those here who do not remember your work, let me say that this guy has the "chops"! A hearty welcome back with the hope of total success at what at times seems the impossible dream. As always, great work!

Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
John Futter
Posts: 1848
Joined: Wed Apr 21, 2004 10:29 pm
Real name: John Futter
Contact:

Re: A 50kV supply with 33mA drive ability

Post by John Futter »

Finn
You make it look easy
I have been counseling the newbies that making one of these is not easy and that you will make many Silicon sacrifices to the gods of high voltage
TAKE NOTE, ALL WHO WANT TO FOLLOW
Finn is very experienced and even He has a small bin of sacrifices.
and yes the chinesium diodes are very good if not overdriven in both current and / or frequency and they are in the good price bracket.
Your pie basket wound coils are are A1
User avatar
Richard Hull
Moderator
Posts: 14975
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: A 50kV supply with 33mA drive ability

Post by Richard Hull »

Thanks for the wise cautionary note to all the noobs and non-electronic types, John. Indeed, Finn is not only talented but expert at his task and he does make it look easy. As I noted the diodes are always an issue in HF, HV efforts. Your praise of Finn's basket weave coils points to my comment that it is always going to be about a transformer be it 60hz or 25khz. Traditionally, semiconductors and high voltage have been at odds. Somehow trying to used solids within quantum mechanical arrangements have not favored speed and high voltages compared to fast enough electrons in a vacuum fire bottle. The battle has only now come within reach with all the advantages of nicely controlled smaller high voltage systems using the latest semiconductors pushing custom ferro-magnetics to their limits. It is almost impossible to destroy a 60hz, 1958, 130 lb dental x-ray system supply that employs vacuum kenotrons for rectification under oil, but those days and systems are gone forever, save for the lucky rare purchase of surplus or abandoned gear.

It must be realized that unless Finn can supply a transformer and circuit boards that will allow the adroit among us to populate the boards and build his setup, once codified by him, this will be a labor of love.

Finally, all must realize that any semiconductor HV supply that can survive fusor builders with ham fists, 5 thumbs of each hand, coupled with zero electronic experience is a pathway to tears. As Mark Rowley wisely warns in his FAQ on the Chicom supplies, destruction and sorrow are but one major mistake away in their operation.

Hopefully, Finn's work will prove to be a bit more bullet proof. I know that is what he is aiming for and we would all love to see it happen.

Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
User avatar
Finn Hammer
Posts: 298
Joined: Sat Mar 05, 2016 7:21 am
Real name: Finn Hammer
Contact:

Re: A 50kV supply with 33mA drive ability

Post by Finn Hammer »

Richard, John, all

Thank you for your kind words and comments.
This post contains random babble, as well as three questions. One about the impedance of the plasma, one about the need for a FAQ page, the latter about kickback from the plasma.

Transformers! I have been interested in the subject for 25 years, and the first one I wound was for a big tesla coil that I built for the Danish Electricity Museum. A 20kV monster which still functions to this day. The basket with wire from failed windings has been emptied many times since then. If I had known then what I know now, it would have been smaller.
This whole project started when I realized that I did not have the gumption to wind another coil for another mains frequency X-ray like transformer with 100K+ turns of thin wire, and although I could have completed several of those coils in the 9 months this project has taken up til now, and at a fraction of the cost, still I do not regret a single moment.

A switcher is the way to go.

I started out with a couple of silly restraints, which all came from inspirations I gained inside a Spellman 30kV, 600W supply: Small basketwound coils, 100kHz operation etc. The 100kHz was no problem as long as I was prototyping with a 1kV secondary, to suit my big wirewound load resistor, but as soon as I started to build the CW doubler, the limitation of the diodes kicked in.
Another silly limitation was the size of the transformer, I wanted to keep it small, but in the end had to let go of that restraint too.
By now, I think I will even drop the frequency to 25-30kHz, and perhaps get 50mA out of the diodes instead, and then add a couple of stages more in the doubler.
My doubler is built from series connected 1600V 68nF polypropylenes and this type of cap has been used and abused in tesla coils for decades, and it is well known how far they can be overvolted, even in pulsed applications. I had a big stash of them from a tesla coil project that did not make it beyond the hoarding stage, so they come to good use now, providing a 17nF stage capacitance,which results in a low 30V ripple.
I have a parametric LTspice model running and it is a good way to get insight in the direction current flows in the stack, for example to understand why the bottom diodes do not carry the full input current, but only the output current.
It is my advice to all who want to build a CW stack: get LTspice, learn to use it by entering a parametric model and gain insight in the circuit without burning too many diodes.

At present, I am working on the spark sense circuit, I think it will be old school current transformer/comparator and triple five timer.

I have a lot of experience in power electronics, because I got my feet wet in the tesla coiling community where I was lucky enough to join in by the time hand wound capacitors were discarded in favour of the Multi Mini Capacitor (MMC) and have taken an active part in the development of the solid state tesla coils. As such, my experience lies in the application of Oscillators with current feedback.
This is my first real switching supply.

A Danish PSU expert has just told me that this is probably not an LLC, but rather an LLCC supply, the latter is hampered by an unfortunate tendency to increase gain at low loads. Since a fusor does not apply a load untill the plasma strikes, this may be a benefit, because the suply will automatically back off voltage when the plasma strikes, and starts to consume some power, and it will be a matter of experimentation to find a useful level of backoff.

There is one question I have been mulling over for a while, and that is this: Is there a record of plasma impedance relative to pressure? I do realize that pressure is supposed to fluctuate wildly in a fusor, (and I guess here: mainly in the start phase) but anyway.
An impedance magnitude would be usefull in predicting which is a more urgent requirement: Current or Voltage.
Lets say I want to burn 1kW off in the fusor, and it happens at 5mA, then the impedance of the plasma will be 40Meg and it will require 200kV to drive it, in other words: this is not likely to happen anytime soon.
If the 1kW is delivered at 20mA, the impedance works out to 2.5Megs and the voltage requirement lands at 50kV.

But is there some record of current/voltage (measured after the ballast resistor), this would be a big help.

Kickback from the plasma! I have been fighting to understand the mechanics behind this reported phenomena, said to chase the smoke out of CW diodes and switcher PSU's in particular. I would appreciate if someone would offer a comprehensible explanation of this, since it is really hard to protect a supply against a failure mode that is not understood by the designer.

I am working on a Transformer FAQ, with the purpose of presenting insight into the magnetics without too much magnetic jumbo-mumbo, to appeal to the newbies with misconceptions about core saturation, but also covering topics like core loss limitation. Does that sound like a reasonable idea, or does it fall outside the scope of this forum?

Happy hollidays to all!
Cheers, Finn Hammer
User avatar
Richard Hull
Moderator
Posts: 14975
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: A 50kV supply with 33mA drive ability

Post by Richard Hull »

Finn a Transformer FAQ by you would be welcomed. You ask for plasma impedance measurements, a very reasonable request as you note for a power supply designer. I have only used a 150lb dental x-ray transformer from the 1950's with my system so it is bullet proof, I uses a simple bridge rectifier system with no capacitors and on a variac smoothly advanced from 0-about 45kv max. I have a single 62k ohm 100 watt ballast resistor to the output of the supply, all under oil. Thus it is fully involved with 120 hz full unfiltered DC. Any data from me would not be worth beans, I fear.

Once my fusor is "conditioned" as I raise the voltage to about 28-30kv on my digital voltmeter, the plasma just strikes and the current roars to 50-60ma with a starting gas pressure of 12 microns. the voltage plunges due to the ballast and then in about second the current plunges to 4ma and the voltage returns to 28kv. The plasma in now lit and I must quickly advance the voltage to just maintain the 4-5ma and lit plasma. The gas behaves funny in that it now allows me to zip up to 34kv at 6ma and the gas pressure no drops slowly forcing me to leak in more gas to raise the pressure as the current is dropping! in the end things smooth out around 13-14 microns with 10-15 ma and a steady unwavering gas pressure, stable voltage of 40kev. The only rough patch is the plasma ignition. Lots of instantaneous current saving the supply via the ballast resistor but the voltage pops right back up and the current falls is left untouched as does the gas pressure. This forces that required upping of the gas pressure a voltage quickly to keep it lit. from this point on there are no current spikes no voltage spikes and only a slow and steady pressure drop to which more must be leaked in to maintain the current. (gas is being buried in the walls of the fusor...A good thing).

As mentioned many times for the last 20 years here, we do not operate in any manner of broad stable regime save for a trained operator balancing the operational system on the razor's edge between electron runaway via a heavy townsend avalanching discharge and a simple stable glow mode region in the gas. Glow modes work smoothly in the 100 microamp range. Fusion needs current! For current increases the number of ions, (deuterons), needed to do fusion, a plus. More pressure, more fusion fuel, a plus. More voltage means one moves into the better part of the fusion cross section curve, more fusion, another plus. All of these pluses drag the operator ever closer to a disaster in the gas environment, (townsend runaway discharge). The razors edge....The balancing act against a runaway sudden end of fusion.

Recently plotted voltages pressure and currents over a 20 minute successful fusion run check out this URL and glean what you can. Drag the page up and down for more data on later runs. This is real operation and real time data. We must remember, my supply has at least a 62k ohm invariant impedance (resistor ballast) figuring the gas plasma loads should be a snap throughout my data.


viewtopic.php?f=13&t=13568&p=89735#p89735


Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
User avatar
Bob Reite
Posts: 576
Joined: Sun Aug 25, 2013 9:03 pm
Real name: Bob Reite
Location: Wilkes Barre/Scranton area

Re: A 50kV supply with 33mA drive ability

Post by Bob Reite »

I have a filtered supply, so I can give a DC resistance for the plasma. Now it's going to vary depending on pressure!

I generally operate at 12-14 microns. At that pressure, the system draws 7 mA with 40 KV applied. So that's a DC resistance of 5.7 megohms in round numbers.
The more reactive the materials, the more spectacular the failures.
The testing isn't over until the prototype is destroyed.
User avatar
Finn Hammer
Posts: 298
Joined: Sat Mar 05, 2016 7:21 am
Real name: Finn Hammer
Contact:

Re: A 50kV supply with 33mA drive ability

Post by Finn Hammer »

Thank you, Bob. That number is a great help. Do you have a resistor in series, and if so, what value?

I tried to lower the frequency to 30kHz, and this had the effect that the diodes now can last, even at 50mA draw. I don't feel like testing more diodes to destruction, and 50mA should be more than enough, with amble headroom.
An old hand in PSU design once told me, there are no good reasons to raise the frequency beyond 20kHz, and I guess the reality of diodes are starting to prove his point.

Cheeers, Finn Hammer
User avatar
Richard Hull
Moderator
Posts: 14975
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: A 50kV supply with 33mA drive ability

Post by Richard Hull »

Finn, Based on the data I supplied and using my reported 62k ohm ballast resistor you should have a lot of info over a time ordered run. I believe Bob does have a ballast. It is almost impossible not to have one.

Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
Victor Gonzalez
Posts: 28
Joined: Tue Jul 16, 2019 10:16 pm
Real name: Victor Gonzalez

Re: A 50kV supply with 33mA drive ability

Post by Victor Gonzalez »

Finn can you show us how you wound the secondary pancake?

That's something I'm searching and can't find
User avatar
Richard Hull
Moderator
Posts: 14975
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: A 50kV supply with 33mA drive ability

Post by Richard Hull »

I doubt if a pressure versus voltage will be of any value based on fusor type and size and special circumstances due to this. We have observed this forever, here. Every odd fusor from big to small demands the operator to learn how his pressure voltage and current respond to actually do fusion.

As to kickback in a struck plasma, this never happens in a stable, fully functional power supply that is not some sort of current controller mess of a supply that is being operated on the ragged edge of its capability.
A huge strong supply of 4kw of continuous capacity will never burp once the plasma is lit. It has never happened to me with my heavy x-ray transformer. There is no multiplier and no storage capacitor in the entire system! It is 120hz full bore ripple! I will let others chime in who have had it happen to them and I recommend when they chime it, they give full data about their fusor, shape, size, etc., and what supply they are using.

I am unaware of any living being here doing significant fusion who is using non-multiplied 40-50kv full wave rectified, non-filtered HV at 120hz full ripple. Most with issues are using HF switchers professional or of the precipitator type that are ostensibly pure smoothed DC. while at the same time, they are working close to the ragged edge of its capability. Some seem to have mastered this cliff-hanger supply situation like Mark Rowley, which means he is a plus-ultra operator and has the skill set for his specific device under his full and able control. Mark has had a few smoke plumes in his early work, but not lately. He has learned the "rope limit" while circumnavigating his switcher supply, with a suitable supply of spare components in case he dare challenge the supply and lose that bout in the ring.

The big hassle for many is that monster startup current drain. All gas glow systems have a large start up hysteresis. In the case of my fusor, once conditioned, this can be 15kv work of differential. wants to smoothly run at startup at 15kv but strikes at 26kv. the 63k ohm 125 watt resistor takes it on the chin until I dial the voltage back. Due to my skilled operation, I usually have it functioning at 30kv glow a minute or two after striking. From there I do my spider monkey routine adjusting the pressure smoothly to meet a "feel" for the current as I creep the voltage and fusion ever upward. It is an art. It makes it easy when you need not even let your power supply capability enter your stream of consciousness.

Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
User avatar
Finn Hammer
Posts: 298
Joined: Sat Mar 05, 2016 7:21 am
Real name: Finn Hammer
Contact:

Re: A 50kV supply with 33mA drive ability

Post by Finn Hammer »

Victor,

You are right, basket coil winding is a lost art, and a complete separate project.

I bought a MoReCo winder on EBay, and it is a piece of junk, really, but if you nurse it a bit, it can be brought to work. There is a lot of unwanted slack in it though.

It looks like this:

IMG_20210119_145318.jpg

The brass handle on the *drive shaft* is the center of action, and turning it creates the two required motions, the rotation of the bobbin and the reciprocating motion of the wire guide.
The drive shaft also has a cam which produces a reciprocating motion on the (shall we call it *cam shaft* ;-) ) which transmits it to the wire guide.
The bobbin is fitted between 2 wooden cones at the far left on the *winding shaft*. the drive shaft and winding shaft are meshed together with a set of gears, where the driving gear is the smallest, and this is important, because to get a proper lay of the wires, the winding shaft has to lag the drive shaft.
You can see it on the sketch here: each successive layer must not cross the preceding layer at the edge where the wire changes direction. A very important point.

IMG_20210119_161838.jpg


The bobbin has a layer of double stick tape to insure the first layer stays snug on the bobbin.

firstlayer.jpg



The wire must be cotton or silk covered to create enough friction between windings, otherwise the wire slips off and creates a mess.
I wish you good luck with your coil winding endevour, and do not be discouraged when your junk basket starts to look like this:

pitstop.jpg

It is the norm, rather than the exception.

Cheers, Finn Hammer
User avatar
Richard Hull
Moderator
Posts: 14975
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: A 50kV supply with 33mA drive ability

Post by Richard Hull »

Really cool old basket winder mechanism. Even with the machine, slack in the mechanism forgiven, it is just another art that must be mastered, just like running a fusor. The box full of fails shows you have the stuff needed to get the job done. Adversity abounds in any thing worth doing.

Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
Frank Sanns
Site Admin
Posts: 2117
Joined: Fri Jun 14, 2002 2:26 pm
Real name: Frank Sanns

Re: A 50kV supply with 33mA drive ability

Post by Frank Sanns »

Winding fine wire on coils is challenging. Too loose and they slide off. Too tight and they pull to the side as the layers increase and they slide off.

A solution that I have used with nearly 100% success is to spray the windings with some clear fast drying urethane or lacquer. Just enough to make it tacky and hold. I do this each time I get enough work into the coil that I don't want to loose the last portion of work. A light spray of the entire coil when you finish will make it look uniform.
Achiever's madness; when enough is still not enough. ---FS
We have to stop looking at the world through our physical eyes. The universe is NOT what we see. It is the quantum world that is real. The rest is just an electron illusion. ---FS
User avatar
Finn Hammer
Posts: 298
Joined: Sat Mar 05, 2016 7:21 am
Real name: Finn Hammer
Contact:

Re: A 50kV supply with 33mA drive ability

Post by Finn Hammer »

Frank,
Thanks for the tip. So you have a basketwinder too? Which brand is it?

I have used CA glue for the same purpose and I gather there is even prefabricated wire with a coating which can be made sticky, by passing it through a bath with solvent, but up to now, have not managed to procure any of it.
I am just at the end of restoring an old lathe,

IMG_20210119_145508[1].jpg

and when it gets operationable, I may try to make multi section bobbins, although I am not sure they will help me in any way, but just take up more space in the winding window. The ability to accomodate 2 separate windings on a single core really appeals to me, and this may not be possible with multi section bobbins, which tend to be wider and not so tall.

Cheers, Finn Hammer
Frank Sanns
Site Admin
Posts: 2117
Joined: Fri Jun 14, 2002 2:26 pm
Real name: Frank Sanns

Re: A 50kV supply with 33mA drive ability

Post by Frank Sanns »

Finn,

My wire winder was the two handed one. Early on I did use my father's small lathe but it was too powerful and I did not want to have the wire grab and pull my hand in or slice through it. The lathe was only

The biggest project I every did by hand was around 4 miles of 22 gauge magnet wire on an 18" phenolic tube. It was a pseudo Tesla coil but it really was a pulsed air core transformer. Easily 5 foot sparks with each pulse. Plenty of fine tendrils.

Well worth the effort.

My technique was to make wooden plugs for each end of the tube with a hole in the center. A brass rod ran the length. At one end was a V block for the rod to sit in. At the other end, it was bent to form a handle. Using gloves, I would guide the wire in single layer as my wife turned the crank. I think we finished in in around 4 non-consecutive nights. It was tedious and I learned very quickly the trick to pull the windings tight and tape after every several turns. Then before the tape was removed, I sprayed it down untaped areas with insulating varnish, Once it dried, I removed the tape and sprayed the entire coil. I figured a little more dielectric between the windings was a good thing.

Yours look mighty pretty. Wish I had one of those machines when I was making smaller coils. Unfortunately I did not have the finances or space for such things when I was working on them.
Achiever's madness; when enough is still not enough. ---FS
We have to stop looking at the world through our physical eyes. The universe is NOT what we see. It is the quantum world that is real. The rest is just an electron illusion. ---FS
John Futter
Posts: 1848
Joined: Wed Apr 21, 2004 10:29 pm
Real name: John Futter
Contact:

Re: A 50kV supply with 33mA drive ability

Post by John Futter »

Nice Lathe Finn
I had one of those until about 16 years ago sold it because I bought a Colchester Chip master lathe and an Bridgeport mill to replace it. I sold the milling attachment seperately
The Timken head bearing is very expensive in those Boxfords.
User avatar
Joe Gayo
Posts: 404
Joined: Sun Jan 06, 2019 9:34 pm
Real name: Joe Gayo
Location: USA

Re: A 50kV supply with 33mA drive ability

Post by Joe Gayo »

Finn,

Is your supply current/voltage mode controlled, or is it an open loop? If you are using feedback, how did you design the resistor divider for the voltage feedback? I find that this is the most critical part when designing a closed-loop high voltage supply. It has to be the right balance of DC and AC feedback. Have you performed any stability analysis?

Joe
User avatar
Finn Hammer
Posts: 298
Joined: Sat Mar 05, 2016 7:21 am
Real name: Finn Hammer
Contact:

Re: A 50kV supply with 33mA drive ability

Post by Finn Hammer »

Joe,

Open loop, I'm afraid.
For one, I fear that the plasma would react faster than any feedback system, and could lead to oscillations.
Another reason is that I am not so good with bode plots and phase shift.

Cheers, Finn Hammer
User avatar
Richard Hull
Moderator
Posts: 14975
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: A 50kV supply with 33mA drive ability

Post by Richard Hull »

System hysteresis at such energies is a big issue. Closed loop systems that oscillate are bad. Nothing beats the "bungy cord" of a nice high watt, suitable resistor in the HV line in starting a fusor plasma. The terror is momentary and experienced only at startup in a good system with a suitable supply under the control of a human being who can reduce the supply voltage quickly once started and adjust the pressure and voltage to a current the supply is happy to supply continuously.

Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
User avatar
Finn Hammer
Posts: 298
Joined: Sat Mar 05, 2016 7:21 am
Real name: Finn Hammer
Contact:

Re: A 50kV supply with 33mA drive ability

Post by Finn Hammer »

All,

After a lot of beginners rant in the appropriate beginners forums, I am happy to be back here where I feel I am a bit more in controll of things, and then perhaps not. This delay in posting is the result of unforseeable trouble in obtaining a proper supply of high thermal conductive potting material.
But now I am in posession of 10Kg og Robnor PX700k-1/BK, and that ought to last my remaining lifetime.

I had previously potted one multiplier, a 30kV model and it functioned as desired, putting 50mA out without problems:

First prototype to look alright
First prototype to look alright

I also had a 60kV prototype on the floor, waiting to be potted:

Guts of the 60kV variant
Guts of the 60kV variant

This is the connector I use: a 4mm banana plug, some heatshrink and some pneumatic fitting. It works fine but I should probably put my lathe to some good use and produce something that looks more serious.

Plug detail.
Plug detail.

After the potting had cured It was time to hook it up to the trusty 1.6MOhm load:

60kV potted and ready to go
60kV potted and ready to go

And brace myself before starting to wind the variac towards the endstop. I admit to being scared this time, the ion wind from the hissing resistive tower load was noticable, but perhaps it was just the fear of a failiure that got me.


Scope readout
Scope readout

But then not so shabby! 63.8kV @ 37.5 mA. That ought to doo even for the most power hungry fusor. (for a while, at least ;-) )
Cyan and Magenta traces carry the real magnitude labels, although Magenta is in Ampere, not volts, for gate signal and primary current respectively, the yellow voltage trace magnitude labels are a result of a 20000:1 voltage probe and a probe setting of 20X so it reads 20kV/division.

So far so good, what now?. Obviously, a 100kV unit is within reach, just a matter of field controll here and there. I am willing to share, the question is how much more is needed before I have told enough for others to replicate, should they want to.




Cheers, Finn Hammer
User avatar
Mark Rowley
Posts: 908
Joined: Sun Dec 29, 2002 12:20 am
Real name: Mark Rowley
Location: Sacramento California
Contact:

Re: A 50kV supply with 33mA drive ability

Post by Mark Rowley »

Looking great Finn!
Potting is definitely the way to go for optimal performance. Tbh though, I’m not brave enough to do it. Just too worried about one component failing and not being able to replace it. It looks like you spared no expense by utilizing top quality caps and diodes so failure maybe be very unlikely.

It’s great to see more neutron capable power supplies using multiplier circuits. Undoubtedly the future PSU of the fusor hobby.

Mark Rowley
User avatar
Finn Hammer
Posts: 298
Joined: Sat Mar 05, 2016 7:21 am
Real name: Finn Hammer
Contact:

Re: A 50kV supply with 33mA drive ability

Post by Finn Hammer »

Thanks, Mark.

Spellman and all the other big ones have done it for decades, time to catch up.
One thing I'm pondering is the benefits and tradeoffs by having either a small or large stage capacitance.
A small capacitance will have a larger ripple, and I imagine this will lead to a neutron output determined not by the peak voltage, but rather peak minus half ripple. A problem in this scenario is, that the breakdown of the feedthrough is determined by the peak voltage, so this component will have to be overengineered related to neutron output.
A large stage capacitance otoh, has small ripple, so the neutron output will relate directly to the quality of the feedthrough. But the energy released in a short from grid to shell may shatter or even evaporate the grid.
The latter would be preferable with respect to the chances of schrapnell ending up in the turbo, already coasting down since the controller got fried by the EMF wave from the shorted supply.
Perhaps I worry too much.
These multipliers of mine have 22nF stage capacitance. I solemny promise to build a version with 3.3 nF stage capacitance, so some time in the not so far future, I will be able to answer that question myself.

Cheers, Finn Hammer
User avatar
Richard Hull
Moderator
Posts: 14975
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: A 50kV supply with 33mA drive ability

Post by Richard Hull »

All of your work is appreciated my those who seek to make everything in the power circuit more bullet proof. Ripple vs. pure, or close to it, DC with its attendant stored energy capacity is always going to be a tradeoff. The critical breakover ionization potential is always an issue. That is why the ballast resistor comes into play to save the day and the supply and grid.

Richard Hull
Progress may have been a good thing once, but it just went on too long. - Yogi Berra
Fusion is the energy of the future....and it always will be
The more complex the idea put forward by the poor amateur, the more likely it will never see embodiment
User avatar
Finn Hammer
Posts: 298
Joined: Sat Mar 05, 2016 7:21 am
Real name: Finn Hammer
Contact:

Re: A 50kV supply with 33mA drive ability

Post by Finn Hammer »

All,

It is a big day, when a selfmade PSU goes into the cabinet of it's own, and as always, this cabinet shows up to be just a tad bit to small, but for now it is ok.


IMG_20210526_110900.jpg

Sitting nicely in the fusor's frame it has readout for Mosfet temperature, transformer core temperature and output current. The switch to the right is a combined Kill/reset switch.

IMG_20210528_091038.jpg

It powers the fusor with great authority, so great that I have plans for an upgrade consisting of a PFC frontend booster, and an intermediate buck voltage control.


The PCB below is from a TI reference design, for a 3kW PFC, and I will populate this board, but that is going to be a project for the next winter.

IMG_20210529_190137.jpg


Cheers, Finn Hammer
Post Reply

Return to “High Voltage - Fusor Input Power (& FAQs)”