My Fusor Progress
- Mark Rowley
- Posts: 909
- Joined: Sun Dec 29, 2002 12:20 am
- Real name: Mark Rowley
- Location: Sacramento California
- Contact:
Re: My Fusor Progress
Nicolas,
You may be interested in my earlier efforts regarding a ZVS driven PSU. The end result was a decent and reliable supply which provided enough power to activate a respectable array of different metals.
Keep in mind this supply was designed to operate for my specific fusor. It is not suitable for a large cavernous system.
viewtopic.php?t=13907
I'd also recommend your read the FAQ's on Precipitatior PSU operation. The do's and dont's tend to parallel ZVS driven systems. Learning the pitfalls will save you a ton of time and money.
Mark Rowley
You may be interested in my earlier efforts regarding a ZVS driven PSU. The end result was a decent and reliable supply which provided enough power to activate a respectable array of different metals.
Keep in mind this supply was designed to operate for my specific fusor. It is not suitable for a large cavernous system.
viewtopic.php?t=13907
I'd also recommend your read the FAQ's on Precipitatior PSU operation. The do's and dont's tend to parallel ZVS driven systems. Learning the pitfalls will save you a ton of time and money.
Mark Rowley
-
- Posts: 501
- Joined: Thu Sep 23, 2021 10:36 am
- Real name: Matt Gibson
Re: My Fusor Progress
Nicolas,
The capacitors are capable of supplying peak current during shorting that can kill your diodes. Seeing that you’re using a ballast resistor, you’re probably okay for now. Just don’t keep doing it as it’s still stressing them.
Another potential issue will be your 100khz operating frequency. From what I understand, ceramic capacitors, like the ones you’re using, don’t like high frequency. Your diodes are probably also getting pretty stressed at 100khz. Finn Hammer did some testing and found his diodes weren’t happy beyond 50kHz and I think he settled around 30khz. His power supply is pushing 3kW I believe.
-Matt
The capacitors are capable of supplying peak current during shorting that can kill your diodes. Seeing that you’re using a ballast resistor, you’re probably okay for now. Just don’t keep doing it as it’s still stressing them.
Another potential issue will be your 100khz operating frequency. From what I understand, ceramic capacitors, like the ones you’re using, don’t like high frequency. Your diodes are probably also getting pretty stressed at 100khz. Finn Hammer did some testing and found his diodes weren’t happy beyond 50kHz and I think he settled around 30khz. His power supply is pushing 3kW I believe.
-Matt
- Finn Hammer
- Posts: 298
- Joined: Sat Mar 05, 2016 7:21 am
- Real name: Finn Hammer
- Contact:
Re: My Fusor Progress
The issue of drive frequency is the result of design decisions, which arise from the "can of worms situation" which a power supply design process inevitably ends up as.
It is worth noting, that as an amateur designer, one does not always make the wise decisions that the more seasoned industrial designers would chose, based on decades of company experiences.
In my case, I started out wanting to use 100kHz drive frequency, simply because I had had a Spellman supply that used 100kHz drive frequency.
At that time, I had little experience with rectifying diodes, but I soon found out, that even fast 20nS reverse recovery types buckeled under at that frequency. During a long iteration of experiments, I found out that 50kHz was the upper limit of the 2CL2FP diodes I had on hand.
But that was not the end of it, because even at that frequency, the ferrite core would heat to uncomfortable levels, no matter how low a flux level I chose. (Remember that ferrites are saturation limited below 20kHz, above that, the limiting factor becomes the core loss, and the core loss goes up with frequency and flux density.
The final compromise landed my design at 35kHz, where all components were happy, and funny enough, that is around the values where most commercial high voltage switchers operate.
Once the magnetics are sorted, the power throughput is only limited by heating of the copper due to resistance, instead the limit occours in the driver electronics. In my case, 96A SIC mosfets warranted resilience at up to at least 2.5kW, but I think it could well go as far as 5kW.
Cheers, Finn Hammer
It is worth noting, that as an amateur designer, one does not always make the wise decisions that the more seasoned industrial designers would chose, based on decades of company experiences.
In my case, I started out wanting to use 100kHz drive frequency, simply because I had had a Spellman supply that used 100kHz drive frequency.
At that time, I had little experience with rectifying diodes, but I soon found out, that even fast 20nS reverse recovery types buckeled under at that frequency. During a long iteration of experiments, I found out that 50kHz was the upper limit of the 2CL2FP diodes I had on hand.
But that was not the end of it, because even at that frequency, the ferrite core would heat to uncomfortable levels, no matter how low a flux level I chose. (Remember that ferrites are saturation limited below 20kHz, above that, the limiting factor becomes the core loss, and the core loss goes up with frequency and flux density.
The final compromise landed my design at 35kHz, where all components were happy, and funny enough, that is around the values where most commercial high voltage switchers operate.
Once the magnetics are sorted, the power throughput is only limited by heating of the copper due to resistance, instead the limit occours in the driver electronics. In my case, 96A SIC mosfets warranted resilience at up to at least 2.5kW, but I think it could well go as far as 5kW.
Cheers, Finn Hammer
-
- Posts: 501
- Joined: Thu Sep 23, 2021 10:36 am
- Real name: Matt Gibson
Re: My Fusor Progress
Hi Finn,
How’s that transformer FAQ coming? Magnetics wasn’t my strong suit in college…Ferrite material, volts/turn, and frequency all get me turned backwards. Seeing as this seems to be the future for fusor power supply, I bet many would find themselves just as turned around![Smile :-)](./images/smilies/icon_e_smile.gif)
-Matt
How’s that transformer FAQ coming? Magnetics wasn’t my strong suit in college…Ferrite material, volts/turn, and frequency all get me turned backwards. Seeing as this seems to be the future for fusor power supply, I bet many would find themselves just as turned around
![Smile :-)](./images/smilies/icon_e_smile.gif)
-Matt
-
- Posts: 17
- Joined: Wed Jun 16, 2021 11:38 pm
- Real name: Nicolas Malo
Re: My Fusor Progress
Hi everyone, I can’t stress enough how much I appreciate all your replies. I did some little testing over the last couple of days and got some interesting results. Firstly, 1 of the diodes did die after arcing. I had gotten 8 for less than 10$CAD so it isn’t much of a trouble. I was using some microwave diode because I wasn’t aware of the concept of recovery time. I figured 100kHz wasn’t going to impact their functionality. I recently saw an old Richard’s post on HV diode where he did mention that they had a slow recovery. I examined them and found that their cutoff was as low as 1kHz. I do have 1 diode from a microwave switcher supply. It’s a UX-F5B rated for 8kV, 150mA, and 150ns recovery time. I tested it and it was fine at 100kHz. They’re inexpensive and would be a perfect fit if it wasn’t for their 8kV limit. In my CW multiplier, each diode would have to be able to withstand a maximum of 25kV. I’m pretty sure I already know the answer, but would I be able to push that 8kV limit to 25 if I were to put their connections in epoxy and under oil?
Secondly, I did some testing of the primary inductance of my ferrite transformers and found it to be around 75uH. With 2 in parallel connected to my ZVS driver, the total inductance will be 37.5uH and the capacitance of the ZVS is 2uF giving a theoretical resonant frequency of around 18kHz which matched my recent experimental result. This meant that I was way off with my initial 100kHz estimation.
Thirdly, in further testing, I noticed that my secondary coil was arcing to my core and thus my primary even under oil (See first video link). In the next couple of days, I will look at buying a new high voltage/high-frequency diode, rewinding a new secondary coil in epoxy, providing more isolation between secondary/core/primary and might look at buying a more resilient capacitor as they are known to lose considerable value over time.
Finally, while ramping up my voltage, I noticed that my variac was humming a considerable amount until the ZVS input voltage reached its 12V minimum voltage requirement. Does anyone know why exactly? I was thinking of adding a 12V/25-30A relay to the input of the ZVS If it were the source of the problem. I added an exact schematic of my circuit for reference.
Here’s a link to the arc to the core:
https://www.youtube.com/shorts/nTbfXjm2Grs
Here’s a link to some fooling around I did with some extra transformer I had made:
https://www.youtube.com/watch?v=9rOxToG ... e=youtu.be
Sorry for the bad quality of the videos
Secondly, I did some testing of the primary inductance of my ferrite transformers and found it to be around 75uH. With 2 in parallel connected to my ZVS driver, the total inductance will be 37.5uH and the capacitance of the ZVS is 2uF giving a theoretical resonant frequency of around 18kHz which matched my recent experimental result. This meant that I was way off with my initial 100kHz estimation.
Thirdly, in further testing, I noticed that my secondary coil was arcing to my core and thus my primary even under oil (See first video link). In the next couple of days, I will look at buying a new high voltage/high-frequency diode, rewinding a new secondary coil in epoxy, providing more isolation between secondary/core/primary and might look at buying a more resilient capacitor as they are known to lose considerable value over time.
Finally, while ramping up my voltage, I noticed that my variac was humming a considerable amount until the ZVS input voltage reached its 12V minimum voltage requirement. Does anyone know why exactly? I was thinking of adding a 12V/25-30A relay to the input of the ZVS If it were the source of the problem. I added an exact schematic of my circuit for reference.
Here’s a link to the arc to the core:
https://www.youtube.com/shorts/nTbfXjm2Grs
Here’s a link to some fooling around I did with some extra transformer I had made:
https://www.youtube.com/watch?v=9rOxToG ... e=youtu.be
Sorry for the bad quality of the videos
- Mark Rowley
- Posts: 909
- Joined: Sun Dec 29, 2002 12:20 am
- Real name: Mark Rowley
- Location: Sacramento California
- Contact:
Re: My Fusor Progress
ZVS drivers MUST be initiated at the starting voltage. Slowly ramping them up to the starting voltage causes overload issues with the fets and will surely cause them to fail (catastrophically sometimes) in short order. In other words, never start a ZVS driver at zero volts and ramp it up. That humming you hear is a ton of amperage being absorbed by the two mosfets fighting each other.
To properly start a ZVS driver, identify the startring voltage and build a psu that can give it the full voltage/amperage rating at an instant. Meaning, if starting voltage is 12 volts, turn it on with a 12 volt starting voltage. From that point you can slowly ramp up the voltage to the drivers maximum rating. Just make sure your PSU can supply the needed amperage at the higher voltages, if not, then your going to detonate your mosfets.
Best advice I can give you is dont overcomplicate this. It's really basic if you follow some essential guidelines. I know some folks here disagree, but minutia can bog you down for months, in some cases years.
Mark Rowley
To properly start a ZVS driver, identify the startring voltage and build a psu that can give it the full voltage/amperage rating at an instant. Meaning, if starting voltage is 12 volts, turn it on with a 12 volt starting voltage. From that point you can slowly ramp up the voltage to the drivers maximum rating. Just make sure your PSU can supply the needed amperage at the higher voltages, if not, then your going to detonate your mosfets.
Best advice I can give you is dont overcomplicate this. It's really basic if you follow some essential guidelines. I know some folks here disagree, but minutia can bog you down for months, in some cases years.
Mark Rowley
- Dennis P Brown
- Posts: 3190
- Joined: Sun May 20, 2012 10:46 am
- Real name: Dennis Brown
Re: My Fusor Progress
No, you can't raise a diodes operating voltage by placing them under oil or in epoxy! The maximum operating voltage of any diode is an innate electronic property of the diode material/design and cannot be changed. Never exceed its max. voltage - a huge failure will occur. One has to chain the diodes in series in order to achieve higher operating voltages (i.e. two 8 kV diodes in series can handle 16 kV; but do note that does NOT increase their current capabilities) - of course, placing in epoxy or oil reduces corona issues for a final design and covers up any bad soldering issues.
Aside: do be careful of RMS voltage issues for AC systems
Aside: do be careful of RMS voltage issues for AC systems
- Finn Hammer
- Posts: 298
- Joined: Sat Mar 05, 2016 7:21 am
- Real name: Finn Hammer
- Contact:
Re: My Fusor Progress
Matt, all
The transformer FAQ is up for preview in the high voltage FAQ's:
viewforum.php?f=29
Cheers, Finn Hammer
The transformer FAQ is up for preview in the high voltage FAQ's:
viewforum.php?f=29
Cheers, Finn Hammer
-
- Posts: 501
- Joined: Thu Sep 23, 2021 10:36 am
- Real name: Matt Gibson
Re: My Fusor Progress
Thanks, Finn. Looks like I found a problem with my design already: I built a 6 stage full wave multiplier but I’m only driving it with a single coil. I see that I need a second coil, so that may explain why I’m not getting what I expected! Looks like I have a new rabbit hole to go down…
-Matt
-Matt