Work on a 60kv+ supply. Gonna be a bear!

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.
Post Reply
User avatar
Richard Hull
Moderator
Posts: 14992
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Work on a 60kv+ supply. Gonna be a bear!

Post by Richard Hull »

I attach annotated images of my beginning work on a supply based on Wallis-HiVolt supply components. Mostly these are negative HV voltage doubler discs to be driven by the supplied giant flyback. The trick is driving the flyback with many ampere peak pulsed DC at about 300volt DC supplied directly from a full wave DC supply driven from a 0-240 volt 30 amp variac. The manufacturer used an IGBT bridge pumping 400 volts pulsed DC from their drive box that demanded 208 three phase power to get 100kv @ 60mA out of it. I will be very happy to get 60kv @ 20ma clean DC.

In my design there will obviously be no limiters to voltage or current. keep it simple. Still there is a long road ahead with a lot of high current high power testing along the way. This is not my normal electronics path, but it is time to move on a bit more voltage.

I did the calcs for resonance LC wise. The 800uh primary to the flyback needs 50uf to resonate at 25khz. I do have polypropylene 10 ufd 1kv caps laying around the lab. Seems a bit over board. What about operating at a higher harmonic?

Enjoy the annotated photos as an introduction to this effort in this post's thread.

Again, as usual, click images to enlarge

Richard Hull
Attachments
VDoubler disc.anno.jpg
Positive 100kv.anno.jpg
Giant flyback.anno.jpg
Main Hi-pwr plug.anno.jpg
Schematic.anno.jpg
Schematic2.anno.jpg
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
Matt_Gibson
Posts: 505
Joined: Thu Sep 23, 2021 10:36 am
Real name: Matt Gibson

Re: Work on a 60kv+ supply. Gonna be a bear!

Post by Matt_Gibson »

I’d go the DRSSTC route and use a full bridge of IGBTs.

My current PSU uses a half bridge CM300dy IGBT module, driven by a DRSSTC controller (eBay has a ton of these from a China that are all based of Steve Wards design) that uses a cheap function generator for frequency input (36khz). A large variac controls voltage to the half bridge. All of this is what drives my self wound transformer. I also use a series capacitor to resonate with my leakage inductance.

-Matt
User avatar
Rich Gorski
Posts: 133
Joined: Mon Aug 01, 2022 4:34 pm
Real name: Rich Gorski
Location: Illinois

Re: Work on a 60kv+ supply. Gonna be a bear!

Post by Rich Gorski »

Richard,

Wow, that's going to be a beast of a supply. It will rival my DXR3000 but without the hassle of unexpected shutdown for over current.

Just a though for the pulsed driver. Maybe consider a simple high voltage power MOSFET circuit. These go as high as 1.2kV at 20A drain current and cheap too under $10. I think you can drive them directly from a pulse generator. Should easily handle 25kHz.

Rich G.
User avatar
Richard Hull
Moderator
Posts: 14992
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: Work on a 60kv+ supply. Gonna be a bear!

Post by Richard Hull »

I plan to manufacturer from scratch the driver as a half bridge of MOSFETS although I have plenty of suitable IGBTs on hand as well. It is to be remembered that I am looking for only 60kv @20ma, (1200 watts), at full power, not the original component design of 100kv @ 60ma. (6000 watts)
I am currently experimenting with a tiny SOIC IR half bridge driver, (IR21091STRPBF),and it seems good and drives a couple of small DPAK MOSFETs to 5-10 amps at 28 volts on a bread board, driven by a variable freq/duty cycle Marlin P Jones LCD readout oscillator.

Images attached here

Richard Hull
Attachments
Breadboard.JPG
MPJA generator.anno.jpg
Basic half Bridge circuit.anno.jpg
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
Rich Gorski
Posts: 133
Joined: Mon Aug 01, 2022 4:34 pm
Real name: Rich Gorski
Location: Illinois

Re: Work on a 60kv+ supply. Gonna be a bear!

Post by Rich Gorski »

Yeah. I've got a few dozen of those little pulse generators laying around. Neat little things. I used them on an old project of mine driving a MOSFET circuit to pulse deep UV LEDs (235nm LEDs). Hard to believe LEDs can go that far down in wavelength. These prototype LEDs did although they didn't last long at any reasonable power level. That was 5 years ago so maybe the deep UV LED technology has improved and 235nm is common now.

Rich G.
User avatar
Richard Hull
Moderator
Posts: 14992
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: Work on a 60kv+ supply. Gonna be a bear!

Post by Richard Hull »

For those interested in that cool little ready to use variable freq/duty item, it is here for $6.95

https://www.mpja.com/Variable-Frequency ... /34758+TE/

It is nice in that it will take a wide range of voltages for power and will output pulses at near that power voltage. It can drive a reasonable digital load. Read the specs.
I am spending on the order of $800/year with MPJA and a bit less with Electronic Gold mine. I am constantly experimenting with Arduino and projects related to nuclear instrumentation and an occasional odd ball project as needed around the home.

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
Peter Schmelcher
Posts: 228
Joined: Sat Nov 13, 2010 1:56 am
Real name: Peter Schmelcher

Re: Work on a 60kv+ supply. Gonna be a bear!

Post by Peter Schmelcher »

Nice circuit designs are in the eye of the beholder. For power stuff I like resonant converters with a single SiC (silicon carbide) power switch. SiC is today’s power choice for electric cars. They switch fast 25-100ns and they have a built-in diode so two parts in one. For example a 4A 3.3kV is $30 from Digikey.

Tear down of Ikea inductive heater. The guy mentions Tesla coils.
Hacking IKEA 2kW Induction Cooktop, Teardown (Part 1 of 5) - YouTube
https://www.youtube.com/watch?v=Mjsws9rZZO8

I think this old Unitrode app note is near the start of the resonant converter idea.

-Peter
Attachments
Ti U-138 Zero Voltage Switching Resonant Power Conversion.pdf
(794.2 KiB) Downloaded 216 times
User avatar
Richard Hull
Moderator
Posts: 14992
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: Work on a 60kv+ supply. Gonna be a bear!

Post by Richard Hull »

Thanks for the info. I will investigate. I will give the single monster IGBT resonant circuit a try.

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
Matt_Gibson
Posts: 505
Joined: Thu Sep 23, 2021 10:36 am
Real name: Matt Gibson

Re: Work on a 60kv+ supply. Gonna be a bear!

Post by Matt_Gibson »

I think you’ve got one awesome project here. That transformer and multiplier is something to really appreciate. Glad to see you coming to the dark side (switching power supply).

I have a hitek xr2000 (100kV and 20mA) coming that I got for cheap on eBay. Hope to have some beefy transformer and multiplier inside to scavenge and incorporate into my existing setup.

Keep the updates and pictures coming!

Switchers FTW.
User avatar
Richard Hull
Moderator
Posts: 14992
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: Work on a 60kv+ supply. Gonna be a bear!

Post by Richard Hull »

I have a basic block diagram. A Zero volt buss is established to isolate all electronics from ground or neutral. Allowing the flyback to supply true isolation between ground and neutral in the high voltage output, yet in the end, allowing a true ground reference need for the fusor and safety.

The switching device can be an IGBT or an FET.

Note that the drive electronics can be independently set to some ideal resonant operational state and the variac can control the actual DC power to the flyback. Some tweaking might be demanded at power. The inductor in the power circuit is to be a large henry inductor capable of handling the current. I hope this works well and doesn't, itself, create and issue in the power filter circuit. I am not well versed on HF DC power circuit filtering.
Classically one might want the highest henry value to choke out the HF energy.

I did the calcs for 25khz resonance and a 50uf cap is needed to work with the 800uh flyback primary. I recalc'd and 10uf is resonant at 55kz.
I have a bunch of 10uf polypropylene 10uf 1kv caps in the lab. Is this reasonable?

Any issues that you HF design boys can see here.


Richard Hull
Attachments
Basic driver block diag.jpg
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
Peter Schmelcher
Posts: 228
Joined: Sat Nov 13, 2010 1:56 am
Real name: Peter Schmelcher

Re: Work on a 60kv+ supply. Gonna be a bear!

Post by Peter Schmelcher »

I think the resonating cap should be across the transformer primary and the inductor. That way the 60Hz input filter capacitor is not in the main path of the high Q resonating currents. Guessing the inductor will have an iron powder core and also the catch diode is internal to the switch. During debug I would add a fuse and resistor to protect the switch, makes less expensive magic smoke.

A general design concern is the reverse recovery time of high voltage multiplier diodes. This and the multiplier capacitors series inductance practically limits the maximum operating frequency of the multiplier stack. A real pain to measure.

A lazy sloth like myself might spend $50 on the Ikea inductive cooker and tune my load to match its operating frequency and add a capacitive divider on the output to reduce the multiplier input voltage during testing, guessing the variac will have limited adjustment with the inductive cooker. A cast iron pan with some water in it would make a quick dummy load during debug testing.

A fun project
-Peter
User avatar
Richard Hull
Moderator
Posts: 14992
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: Work on a 60kv+ supply. Gonna be a bear!

Post by Richard Hull »

Thanks for the info. I will follow the direction.

I found two giant half bridge IGBT blocks I bought from Steve Ward some years back at one of my old Teslathons (now re-named HEAS events). They are rated 600volt @400 amps. Could I just use one of those IGBTs of the bridge? I am currently looking to grab my box of high power IGBTS from storage and see what is in there. All IGBT's I have ever bought have internal hexfred reversing diodes. I have grown use to modern MOSFETs in my under 60v projects with .002 ohm Rds on ratings and no heat sink needed at 5 amp draw.

I was always planning on fusing the hell out of the mockup in stages. Making the bridge for 600 volts with filter, but using only a 120 variac, bumping and scoping several points with the three Pearson 10:1 HF current transformers plus, a normal line voltage and current monitor.

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
Peter Schmelcher
Posts: 228
Joined: Sat Nov 13, 2010 1:56 am
Real name: Peter Schmelcher

Re: Work on a 60kv+ supply. Gonna be a bear!

Post by Peter Schmelcher »

I am no IGBT expert but from memory the older IGBTs have a long turn off tail so they might not switch at 25kHz. Also, the IGBT gate input capacitance might be rather large with high current parts. That said a resonant converter would result in the highest switching frequency for the older parts. The break-before-make time sequencing in an H bridge topology disappears in a resonant converter. If the old parts turn out to be marginal plan B might be an upgrade to a SiC fet in the exact same circuit. I cannot imagine any dual IGBT modules internal wiring that would prevent using half of the module provided the heat sink mounting is electrically isolated.
-Peter
User avatar
Richard Hull
Moderator
Posts: 14992
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: Work on a 60kv+ supply. Gonna be a bear!

Post by Richard Hull »

I looked at Mouser and the 1200 volt, 60A continuous, 120 amp pulsed .030 ohm Rds on, SiC MOSFET is $34.45 each. I probably will wind up buying one, OUCH! According the the data sheet, this is a real brute compared to even the best common Si MOSFET. I had no idea such low Rds on resistances were to be had at such voltages!! Incredible! With me babying it all along the process, hopefully it will survive low power faux pas on my part.

Regardless of first pass testing, I plan on powering up only 1 "pie" or disc first and measuring the doubled DC voltage out of the fly back and then load its output to 20ma ( I have a 1megohm 200 watt variable resistor bank.). Baby steps....

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: Work on a 60kv+ supply. Gonna be a bear!

Post by Bob Reite »

IGBTs should be able to do 25 KHz, but not much higher.
The more reactive the materials, the more spectacular the failures.
The testing isn't over until the prototype is destroyed.
Matt_Gibson
Posts: 505
Joined: Thu Sep 23, 2021 10:36 am
Real name: Matt Gibson

Re: Work on a 60kv+ supply. Gonna be a bear!

Post by Matt_Gibson »

I’m running my cm300dy at 36khz. Everything seems to be running fine…I had to reduce my gate voltage from +\-24v to +\-12v due to the huge gate charge on them.

-Matt
User avatar
Richard Hull
Moderator
Posts: 14992
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: Work on a 60kv+ supply. Gonna be a bear!

Post by Richard Hull »

As an ex-electronics engineer (old school), I will tell a tale of this afternoon of humor and sadness.

The setup:
As most of you know my lab is a filled to capacity as a storied history of my acquisitions of a lifetime of electronics work. Specifically, my Tesla coil work in this tale's case. In my quest to build this HF HV supply, I am trying to do this the easiest way. I am now convinced that modern electronics can now avoid the bridge and half-bridge drives of old. I envision a single SiC MOSFET blasting into a resonant circuit arrangement of my giant fly back a PWM gate signal to create up to a kilowatt drive of pulsed DC of from 0-300VDC.

The key is doing the resonance correctly as noted above. With a measured 800uh primary on the flyback, I calcualte I need a 50.6 ufd capacitor to resonate at 25khz. (my design goal) This will be placed across the primary of the flyback.

The tale:
Casting about among the rubble of my Tesla coil days, in the hot upstairs storage area, I found about 40 - 1uf 2000volt polystyrene snubber caps (ideal).
I also found 4- 60ufd 500volt polypropylene capacitors. Both dielectrics are good for pulsed work. The big capacitors are made by Condenser Products. Just for a back up and enquiry, I called them. Their website had verified for me that the big caps were polypropylene series. It noted that they have been a prime manufacturer of capacitors since 1934! I wanted to ask an engineer if they were series connected to make a 30ufd @ 1kv could they work well at DC pulsed 1KW @ 25khz?

I finally got transferred twice and the third guy listened to my tale as you all have just now. I will try and quote....
"I'm kind of new at this and don't feel I could answer that, you see, our engineer died a couple years ago and we haven't found another yet"
I thought, What! They have or had only one engineer!?? In sympathy, I replied. " I am sorry to hear that, did covid take him out?"
His reply was "No, he was a chain smoker and a heavy drinker. He had COPD and his lungs just failed him in the end"
I thanked him for his honesty and that was that. Well it was going to be his lungs or his liver or both, I figured. To me this was amazing and humorous in that a condenser manufacturer would have only one engineer on site. Was the work so stressful (engineering can be stressful), that it drove him to chain smoke and drink or did he bring it with him as a hire.

A minute or two more though might suggest the answer and why no real engineer might need to be hired immediately. Designing capacitors for the possible 6 major lines of capacitors and learning all the gotchas of all the disciplines and uses your resultant effort might be is a time won knowledge. Many engineering jobs are found to be an OJT for wet behind the ears, newly minted college grads. They learn that many calculations are thwarted by a myriad of unforeseen practicalities which must be dealt with. Fortunately, for established companies, there are a lot of techs and folks on the floor who will quickly knock the rough edges off the newbie engineer.

Further thought on this had me go into the process, itself. There are only so many dielectrics! In 1934 it was paper and mica. As the years progressed, new plastic dielectrics completely changed the capacitor business. Engineers of competence already there had to learn the new dielectrics and design based on them as they came along. I thought, lets face it, after about 2010 their are no new dielectrics of late. If you have been in business since 1934, you have made about every value of capacitance in every form factor and at every voltage over the years. you have a recipe for every thing you have ever made! Furthermore, the computer age has certainly brought forth programs where you punch in voltage, capacity and dielectric and it spits out the turns, thickness and finished volume of the capacitor for you. Now the techs can figure out the packaging, test wind, finalize and adjust. Then and the floor people will make it happen. Who needs an engineer? Well, my questions went begging! It is still nice to have a brilliant mind who has seen it all and survived it all with specific knowledge of the product to interface with engineers who wish to commune at that level in queries.

End of the tale: Well those in the know might make a suggestion. Are both caps gonna' do fine here? I ain't no switcher design engineer. I would think either would do fine and would certainly believe the snubbers are ideal. Well?

Oh I have some photos of my finds in and amongst the rubble.

Richard Hull
Attachments
Cap1.JPG
Cap2.jpg
Cap3.jpg
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
Matt_Gibson
Posts: 505
Joined: Thu Sep 23, 2021 10:36 am
Real name: Matt Gibson

Re: Work on a 60kv+ supply. Gonna be a bear!

Post by Matt_Gibson »

Only thing that I can add is that I sized my capacitor and frequency around my leakage inductance. Once I got close, I adjusted frequency until I saw primary current spike.

-Matt
Peter Schmelcher
Posts: 228
Joined: Sat Nov 13, 2010 1:56 am
Real name: Peter Schmelcher

Re: Work on a 60kv+ supply. Gonna be a bear!

Post by Peter Schmelcher »

Have a vintage smile “Plastic Capacitors Inc.” today known as CDE. Plastic is advertised on the film inside the hermetic 1.25" glass enclosure.
Glass Capacitor.jpg
-Peter
User avatar
Bob Reite
Posts: 576
Joined: Sun Aug 25, 2013 9:03 pm
Real name: Bob Reite
Location: Wilkes Barre/Scranton area

Re: Work on a 60kv+ supply. Gonna be a bear!

Post by Bob Reite »

The resonator capacitor is going to have a lot of circulating current IMHO, I'd use the physically bigger one.
The more reactive the materials, the more spectacular the failures.
The testing isn't over until the prototype is destroyed.
User avatar
Richard Hull
Moderator
Posts: 14992
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: Work on a 60kv+ supply. Gonna be a bear!

Post by Richard Hull »

I was leaning towards the big boy for starters. I have a busy June ahead of me so the going might be a bit slow.

I have no real power plug and the key point is to connect to those two big, ugly .240 pins. I have already machined suitable aluminum 3/8-inch rod stock with suitable hole and the opposite end to accept a large banana plug. The primary lead to the external resonant capacitor need to be extremely short. The mechanical setup in this effort is as important as the electronics. We can't have any high current, HF arcing at this point.

You can see those two primary pulsed power pins in the original post's photos

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
Dennis P Brown
Posts: 3159
Joined: Sun May 20, 2012 10:46 am
Real name: Dennis Brown

Re: Work on a 60kv+ supply. Gonna be a bear!

Post by Dennis P Brown »

An aside relative to large Cap manufacturers. Getting super sized caps - in the 200 -300 lbs, and 50 KV range takes a year and a half lead time. Don't know if it has all been off-shored or just no one keeps even a single unit in stock.

I had a field engineer come to my lab in order to install a large cap based pulsed laser system (Ruby.) His arms were covered in scar's. Puzzled, I went ahead and asked him what had happen. He had both arms inside the power unit adjusting wire connections at the time (the unit was new and uncharged. So I wasn't distracting him during a hazardous time.) He just answered in a simple tone that he often (!) had massive amp discharge loads (3 kV!) run down his arms creating new scares! Needless to say, I both didn't say anything about that methodology of work and was simply shocked beyond measure (no pun really intended.)
User avatar
Richard Hull
Moderator
Posts: 14992
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: Work on a 60kv+ supply. Gonna be a bear!

Post by Richard Hull »

New. large caps are delivered shorted. When banking them, the bank rails are shorted. In most all pulsed electronics in the 300 joule plus range will see from 5-10% of the energy retained after discharge!! This is more true for giant single shot systems. In all cases, at shut down, a suitably sized resistive crowbar is dropped across the bank and once the voltage is at or near zero, a solid copper crowbar is dropped across and locked down.

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
Matt_Gibson
Posts: 505
Joined: Thu Sep 23, 2021 10:36 am
Real name: Matt Gibson

Re: Work on a 60kv+ supply. Gonna be a bear!

Post by Matt_Gibson »

I’ve got a Ross relay (normally closed) that shorts my capacitor bank via some beefy resistors when I have my ruby laser (4kJ capacitor bank) powered off. I treat my capacitor banks very similarly to my guns, always assuming both are loaded and ready to fire.
User avatar
Richard Hull
Moderator
Posts: 14992
Joined: Fri Jun 15, 2001 9:44 am
Real name: Richard Hull

Re: Work on a 60kv+ supply. Gonna be a bear!

Post by Richard Hull »

For those who are curious and capable of reading a schematic and dare to follow my path in this quest, I attach a full schematic of the Hi-Tek flyback cage, interface board and stack discs with the 10 section epoxy block filter cap rated at 150kv. I show in Red the limited 4 resistors needed to not use the flyback (stored for years) but to fed the stack using a 9kv center tapped 60 hertz neon transformer hooked to a variac to output from the stack of discs the full 0-100kv @ mere filtered microamps for use in electrostatic experiments.

I do plan to use their interface board Schematic also shown solely for the nice well controlled output of the load current to the fusor and the output
high voltage as usable analog signals. I have already removed the resistors in red from the little board also shown with them attached, prior to removal.

Richard Hull
Attachments
The interface board... those can see how this loinks to the stack and little board in the next print of the overall block type diagram.  I will have to lift some components to kill the non used center section
The interface board... those can see how this loinks to the stack and little board in the next print of the overall block type diagram. I will have to lift some components to kill the non used center section
The way I modified the stack for Neon xfrm use.  all resistors are now removed.  Imagine them gone! then this is the schematic as Hi-Tek meant it as delivered to work with their interface board below in the next schematic.
The way I modified the stack for Neon xfrm use. all resistors are now removed. Imagine them gone! then this is the schematic as Hi-Tek meant it as delivered to work with their interface board below in the next schematic.
This is the bottom of the plate that mounts the disc stack on the other side.  The is the bottom view that sits on the flyback/interface board cage.
This is the bottom of the plate that mounts the disc stack on the other side. The is the bottom view that sits on the flyback/interface board cage.
The little board on the bottom of the plate seen above. the closeup shows my old mod to work with the neon sign transformer.  Now removed.
The little board on the bottom of the plate seen above. the closeup shows my old mod to work with the neon sign transformer. Now removed.
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
Post Reply

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