Custom Fusor Power Supply

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
Mark Rowley
Posts: 878
Joined: Sun Dec 29, 2002 5:20 am
Real name: Mark Rowley
Location: Sacramento California
Contact:

Custom Fusor Power Supply

Post by Mark Rowley » Mon Mar 08, 2021 8:57 am

After having good results with the 60kV Precipitator power supply, I’ve decided to build a better version based on the same 4 flyback series arrangement.

The major downside of the Chinese PSU is its lack of serviceability. Each of the four flybacks are 15kV DC output. Unfortunately, the secondary coil and rectifier diodes are encased in resin which basically means the transformer is junk if any component within fails. Luckily, most of the failure occurs within the driver circuit which is considerably over engineered for basic fusor use. However, the flybacks do occasionally go bad resulting from diode failure or to a much lesser degree, coil damage. The coil appears to be wound with 36awg which is about as thin as you can get for this type of application. It’s just not designed for the rigors of fusor use. Again, this does not imply the 60kV precip isn’t a great starter or mid-grade psu. It just requires very careful application and operation as explained in the FAQ.

The Upgrade Design
The new design incorporates 4 custom flybacks driven in parallel with the outputs in series. Rather than utilize the damage prone 36awg, each flyback is wound with 1400 turns of #28 AWG. This results with an output of approx 8kV AC. A 2nF single stage CW multiplier on each flyback will bump up the voltage to 16kV. The 2nF’s will theoretically allow for 25mA however 7mA is about all that’s needed for the size of a cube or cylinder fusor (Fusor2020). All four outputs will be in series resulting with an expected potential of ~64kV.

The flybacks are driven with a common $40 1000 watt / 50a Chinese ZVS module.

Except for the driver, the entire assembly is operated under oil.

Unlike the precipitator, this assembly has easily replaceable components as well as serviceable flyback secondaries. The multiplier circuits are modular making for easy removal and servicing.

Shown are the segmented flyback bobbins and soon to be machined blanks. The cores are from the old 60kV precipitator supply.
7EBAB9AA-DC32-4BB2-9C41-AFE7929582B6.jpeg

Machining the bobbin. It’s highly recommended to use HDPE as it’s easier to work with. Unfortunately these four are made from UHMW which, as someone correctly put it, is like machining Silly Putty. It was not a enjoyable process. Shown below is HDPE from an earlier project.
30631983-D406-4C90-8172-22D53E8378CC.jpeg

Completed bobbins fitted to the ferrite cores
04716F54-36AF-4C3A-804E-4EBBB6327367.jpeg

Winding the bobbins was a snap with this inexpensive $40 coil winder.
A5BBF54B-6B86-4A57-B606-B405F21CF771.jpeg

Skipping ahead, this is the completed supply with the four separate single stage multipliers.
9A11846F-1D24-46A0-B57E-6F329A180FFB.jpeg

This is the cheap $40 ZVS driver. As a side note, the production quality is lacking. No heat sink compound as well as a few cold solder joints. Best to give these the once over before using them.
31263EB6-7354-48BC-BBA8-ADD13B98C0C5.jpeg

Similar to the precipitator, no load voltage testing results in an output approaching 100kV. Under load / fusor conditions it idles along nicely at 65kV with less than half the input rating for the driver. I’ve yet to go past this point but will continue with rigorous testing over the next few weeks. A sustained 30-45 minute run is the end goal.

Today’s fusion run was far from anything I had previously done due to no active cooling (I’ve yet to hook it up) and a total absence of wall loading. Once I started the BoT2021 project the Fusor had been sitting on a table collecting dust. Today’s test was literally plug-n-play with no prep or chamber servicing. However, the numbers were good enough to activate silver with no issues.

Not the best pic but it’s a neutron producing Star Mode from earlier today.
E8FF1294-BEB5-499A-88DD-598967BEE932.jpeg

I’ll update as I continue through the testing process.

Mark Rowley
Last edited by Mark Rowley on Mon Mar 08, 2021 6:13 pm, edited 1 time in total.

User avatar
Richard Hull
Moderator
Posts: 13305
Joined: Fri Jun 15, 2001 1:44 pm
Real name: Richard Hull

Re: Custom Fusor Power Supply

Post by Richard Hull » Mon Mar 08, 2021 2:12 pm

That is a lot of great effort. I wish you well with it and hope it pans out to be all you expect from it after all the effort that has gone into the idea.

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
Retired now...Doing only what I want and not what I should...every day is a saturday.

User avatar
Bob Reite
Posts: 498
Joined: Mon Aug 26, 2013 1:03 am
Real name: Bob Reite
Location: Wilkes Barre/Scranton area

Re: Custom Fusor Power Supply

Post by Bob Reite » Tue Mar 09, 2021 1:52 pm

Nice job on those transformers!
The more reactive the materials, the more spectacular the failures.
The testing isn't over until the prototype is destroyed.

User avatar
Mark Rowley
Posts: 878
Joined: Sun Dec 29, 2002 5:20 am
Real name: Mark Rowley
Location: Sacramento California
Contact:

Re: Custom Fusor Power Supply

Post by Mark Rowley » Tue Mar 09, 2021 4:40 pm

Thanks Bob and Richard. It’s met it’s goal of 60kV and can power my fusor all day at that voltage. Nothing gets hot and throttling from the variac is very stable. Quite honestly, I could leave it as is and be perfectly happy.

Testing requires intentionally pushing the design to failure. So as I was hovering near the 1E+06 n/s mark (*60kV/3mA/37mTorr D2), I dropped the pressure to around 30mTorr so I could ramp up the voltage without incurring a big increase in current draw... a common practice with the precip supplies. As expected, the neutron count jumped way up as I got close to 80kV. It was at this point flyback #4 began to arc (under oil) between the primary and secondary windings.

Whats really nice about this design is the ease of serviceability. Removing the supply from the fusor cart took one minute and pulling off the bad flyback took about 5 minutes.

At first I thought the arc-over was caused by an air bubble trapped behind some Kapton tape. So I removed the tape, rewound the bobbin and reinstalled. Same problem at 80kV. So it’s fairly evident the dielectric properties of mineral oil aren’t strong enough to ward off the potential in such a narrow space. Before the windings became too damaged I slid a small piece of acrylic sheet between the primary and secondary. This helped immensely until eddies in the oil caused it to slip out and the arcing resumed.

Over the next day or so the plan is to fortify the dielectric strength in the arc-over region with acrylic or pvc shielding as well as some silicone rubber (or similar) over the windings on that side of the bobbin. I’ll reserve epoxy encasement as an absolute last resort.

During the fix, I added an extra nF of capacitance to each multiplier. The flyback and driver are more than capable of handling the extra load and as expected it provided more reserve current during operation.

Even though it’s a stable supply at 60kV, I’m going to raise the bar to 80kV so it can have a 20kV safety margin.

*When loading is optimal, those parameters typically result in more than double the neutron count.

Mark Rowley

User avatar
Richard Hull
Moderator
Posts: 13305
Joined: Fri Jun 15, 2001 1:44 pm
Real name: Richard Hull

Re: Custom Fusor Power Supply

Post by Richard Hull » Tue Mar 09, 2021 6:27 pm

Really did not want to say anything at the time, but I thought that even under oil the primary spacing to the secondary looked kind of tight. I used a lot of oil in the tesla coil days and never trusted just oil. Spacing is still critical even under oil. (oil stirring is a sure sign of too much field between items.) High frequency operation really serves to heat and stir oil under tension. Same goes for air, of course. High frequency in any dielectric tends to heat it and lead to failures. I think Bob will concur. I have seen mica G-5s heat to where you can't touch them in RF service. They can take the heat unless they get so hot they lose their water of crystallization. ~900deg F

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
Retired now...Doing only what I want and not what I should...every day is a saturday.

User avatar
Mark Rowley
Posts: 878
Joined: Sun Dec 29, 2002 5:20 am
Real name: Mark Rowley
Location: Sacramento California
Contact:

Re: Custom Fusor Power Supply

Post by Mark Rowley » Sat Mar 13, 2021 5:24 am

Excellent results today. 30 minutes of testing without any issues. Operationally speaking, it’s similar up the precipitator unit with some subtle differences. I’m still learning how to drive it and find the operational sweet spot. As of now it’s clear I have to get the chamber cooling back in order as my neutron rate is getting poisoned after about 3-4 minutes of operation.

This video is of today’s cold start. The raw count rate was about 4500cpm on the Ludlum 2221. Voltage was slowly ramped up from 27kV to just under 40kV before I switched it off. Deuterium pressure was 37mTorr. I’m still not at the numbers I was at last year but am slowly getting there.

https://youtu.be/Qdrmq-Er8AM‬


Star Mode pic from the video:
22DFFEC8-E9AB-4E5B-904D-D546CEA437CF.jpeg
The last few mods to the power supply were as follows:

PVC insulators around the flyback primaries
Acrylic shields between flybacks and multipliers
Silicone Rubber on secondary of Flyback #4
Replacement of 3 faulty multiplier caps
Installation of 1gig ohm resistor for voltage meter

It’s not as aesthetically pleasing as before but it’s much more functional with minimal signs of oil churning and no arcing.
054E6B0A-5A88-4A95-B4D0-931CCDA02F32.jpeg
Next few mods will be to shorten the leads from the ZVS driver to the flyback by 75%. If workable, the 12awg leads will be replaced with two pieces of 3/8 copper tubing.

So far I’m not overly impressed with the 1nF/20kV caps. I’ve had 3 fail including another in an unrelated 7 stage voltage multiplier. Other HV enthusiasts have noted similar issues with these caps so it may be worthwhile to upgrade. I’ve got some Soviet doorknob caps I may try at a later date. But FWIW, none of them failed during today’s run. Perhaps the arcing from last week damaged them... time will tell.

Once the water cooling is in place I’ll conduct a sustained 30 minute operational test.

Mark Rowley

User avatar
Richard Hull
Moderator
Posts: 13305
Joined: Fri Jun 15, 2001 1:44 pm
Real name: Richard Hull

Re: Custom Fusor Power Supply

Post by Richard Hull » Sat Mar 13, 2021 5:54 am

Mark, as a 10 year veteran of tesla coiling with a great deal of experience with ceramic HV capacitors. I have found that 100% of this type of capacitor working at high frequencies and high powers (high peak pulse energy) will lose value by shedding the silver vapor deposit (plates) on the ultra high dielectric constant Barium Titanate slugs. Too much voltage will naturally have internal arcing in spite of the epoxy encasement.
Ceramic Capacitors in parallel lead to uneven stresses as one seems to always lose more value than its mate in parallel. Disruptive spark gap coils often placed 50 amp peaks through the caps in even smaller coils. Ceramics in bigger coils rarely performed more than an hour's worth of runs before needing retuning to keep the spark length up due to loss of plate area. The solution was polypropylene foil energy storage caps

If you caps heat after a run, then they are shedding plate area and dropping in value. The weakest cap in the serial discharge multiplier will control the output. I assume when you say they failed, it was due to internal arcing?

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
Retired now...Doing only what I want and not what I should...every day is a saturday.

User avatar
Mark Rowley
Posts: 878
Joined: Sun Dec 29, 2002 5:20 am
Real name: Mark Rowley
Location: Sacramento California
Contact:

Re: Custom Fusor Power Supply

Post by Mark Rowley » Sat Mar 13, 2021 6:17 am

I can’t say for sure why they failed. No physical damage. Just a total loss of value which appeared sudden as opposed to gradually. Almost exactly how others have stated.

Only arcing I was specifically referring to was the earlier breakdown of the mineral oil between the primary and secondary. Very small arcing took place in that region which coincided with the cap failure.

This is the tiny arc (now remedied) that resulted with the cap failure:

https://youtu.be/q8KR4qslSYU

Maybe a spike somehow found its way from that arcing to the three bad caps.

Btw, I got my fusion star videos accidentally mixed up. This one is of the higher neutron count rate, 4500-5000cpm on the Ludlum 2221.

https://youtu.be/yRAZMjBiNaY

Once I get cooling set up and run this for it’s make-or-break 30 minute test, I’ll check the capacitance on each ceramic and report back. It’s not big deal if the whole idea fails. Other than the $40 ZVS and some fasteners it hasn’t really cost anything other than my junk pile raids.

I suppose I could add two more flybacks and forgo the multipliers... ;)

Mark Rowley

User avatar
Mark Rowley
Posts: 878
Joined: Sun Dec 29, 2002 5:20 am
Real name: Mark Rowley
Location: Sacramento California
Contact:

Re: Custom Fusor Power Supply

Post by Mark Rowley » Sat Mar 13, 2021 8:59 pm

Hooked up the cooling and let it run for 30 minutes at about 35% input power to the ZVS driver. Neutron output to the Ludlum 2221 / CHM11 at ~17cm distance averaged 7500cpm for the duration. Removing the moderator dropped the neutron count to background. No hiccups in power supply performance....yet.

This evening I’ll conduct a higher powered 30 minute run. If this pans out ok I’ll get the gamma spec operational and try to activate some Manganese.

MR

User avatar
Richard Hull
Moderator
Posts: 13305
Joined: Fri Jun 15, 2001 1:44 pm
Real name: Richard Hull

Re: Custom Fusor Power Supply

Post by Richard Hull » Sun Mar 14, 2021 4:57 am

I'm glad someone has got a fusor doing great. Good work Mark! Mine sits in the cold, awaiting my continued efforts.

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
Retired now...Doing only what I want and not what I should...every day is a saturday.

Post Reply

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