Author Topic: 500Khz Induction Heater - Need help with Driver  (Read 943 times)

Offline Mysteriousduck

  • High Voltage Enthusiast
  • *
  • Posts: 5
  • Karma: +0/-0
    • View Profile
500Khz Induction Heater - Need help with Driver
« on: January 17, 2020, 11:17:33 PM »
Hello all,

As part of a research group I have been tasked with building an induction heater capable of 500khz switching frequency. The purpose of the heater is to heat and eventually melt basaltic regolith, and the higher frequency is required to overcome the low permeability of the material. Here's the rub - I'm an architecture student, not an EE. Back in high school I built an (extremely dangerous) spark gap tesla coil (with homemade alumnium foil capacitors!) but that's about the extent of my electronics knowledge. Op-amps, PLL circuits, slew rates, gate isolation - this is all new to me.

I got started in induction heating using one of the $40 amazon 1000w units to case-harden exhaust valves and head bolts, and so my first instinct was to pull out all but two of the 0.33uf caps and turn a smaller coil to see if I could get a higher frequency. This worked (325kHz on my el Cheapo scope) but the coil heats up fast (like too hot to touch in <10s) Would it make more sense to further decrease capacitance and increase inductance to keep the coil from heating up so much? Im using a 1kw 48v smps, btw.

Once I've confirmed the material can be heated using induction with a decent efficiency, our directive is for my division to design and build a 15kw heater running on US domestic 240V split phase. I understand how difficult, potentially dangerous, and batspit crazy this is, which is why I am reaching out to as many sources as I can to gain advice. If we had the funding, I would just buy a commercial unit and be done with it, but we don't. so here I am. ;D

I've researched a lot of driver topologies, from Mazilli ZVS drivers to ATMega-controlled PLL circuits. While I would love to stick with the Mazilli driver because of its simplicity, I hear so many stories about people blowing things up that I am wary. (change my mind?)

Uzzor2k's PLL circuit using a 4046 IC seems pretty legit, as it operates in the voltage neighborhood I need, has control for tank voltage, and has a non-explosive current limiting function.  I've been working on an adaptation of it for the last few days and my desk is covered in data sheets for op-amps. Can the LM324 / LM393 keep up at 500kHz+? Should I be looking at something like the OPA828 even though it requires a 2V input bias?

I scored some SCTW90N65G2V samples, and I plan on running two of them in parallel for each bank of the inverter. For the inverter design, I'm looking at something similar to Jonathan Kraidin's '10kw High Power Inverter'

I realize this has been a long and rambling post, but any help yall could provide would be most welcome. Hopefully this can eventually morph into a build thread!

Cheers,

Mike


Offline T3sl4co1l

  • High Voltage Technician
  • ***
  • Posts: 153
  • Karma: +9/-0
    • View Profile
Re: 500Khz Induction Heater - Need help with Driver
« Reply #1 on: January 18, 2020, 12:30:19 AM »
Regolith?  Like, Moon architecture then??

Why not solar?  Power's gotta come from somewhere, anyway.  Hard to argue with a solar vacuum furnace.

What do you expect to heat?  Lunar regolith is mostly stone (there is a small metal content, but it is too diffuse to be conductive AFAIK, and it would take a very long time molten to coalesce), so it's not conductive at all when cold, and only moderately so when hot (where ionic conductivity picks up; also where sintering and melting begin).

The high resistivity actually suggests a far higher frequency, perhaps 10s of MHz for heating the metal particles present, or for driving dielectric losses, at least initially.

Induction works fine with a susceptor though.  A graphite nozzle for instance.  It can also be direct heated (in which case a resistive coil and an insulating e.g. zirconia nozzle can also be used), or, again, focused light can be used, when available.

If this is for a research project, you should be able to convince someone to budget some materials; a Chinese 5kW power supply would be good enough.  It'll be medium frequency, 20-100kHz or whatever, but that will do fine with a susceptor.

Or you can go completely the other way, see if microwaves work.  Near-1kW power supplies at 2.45GHz are conspicuously cheap and available. ;D

(FYI -- you may recognize me from having built a 5kW supply back in 2010.  After that, I designed a family of 400kHz power supplies up to 50kW.  In a small team, we took the better part of 6 months to get a working demo, and the better part of 2 years to commercialize the product.  Even if you don't need as many features as an industrial unit, you will still need a few minimum features, like tuning or load control, and fault protection.  This may take an average EE, heh, well... really, EE or not, it will take many more years to develop the knowledge necessary to design and operate something like this.)

Tim

Offline Mysteriousduck

  • High Voltage Enthusiast
  • *
  • Posts: 5
  • Karma: +0/-0
    • View Profile
Re: 500Khz Induction Heater - Need help with Driver
« Reply #2 on: January 18, 2020, 04:41:22 AM »
Regolith?  Like, Moon architecture then??

Yes sir, that's what my superiors are after.

Why not solar?  Power's gotta come from somewhere, anyway.  Hard to argue with a solar vacuum furnace.

True, but the stuff is wicked nasty when its melted. even eats through PtRh. The goal is to try and touch the molten stuff as little as possible.

What do you expect to heat?  Lunar regolith is mostly stone (there is a small metal content, but it is too diffuse to be conductive AFAIK, and it would take a very long time molten to coalesce), so it's not conductive at all when cold, and only moderately so when hot (where ionic conductivity picks up; also where sintering and melting begin).

There is a fair amount of metal in the stuff, especially in the coarse-grained fraction. Its certainly not ore, but its in there. http://meteorites.wustl.edu/lunar/lunar_soil/lunar_soil_composions.htm

The high resistivity actually suggests a far higher frequency, perhaps 10s of MHz for heating the metal particles present, or for driving dielectric losses, at least initially.

Yes it will be difficult to achieve initial heating without some sort of reusable flux.

Induction works fine with a susceptor though.  A graphite nozzle for instance.  It can also be direct heated (in which case a resistive coil and an insulating e.g. zirconia nozzle can also be used), or, again, focused light can be used, when available.

See corrosive comment. nozzles get et up. Ukrainians have the same issue in the manufacture of rockwool, and their sources are way purer that what you find on the moon. I like the solar concentration angle, I could see it working kinda like powder deposition in an additive manufacturing workflow, helluva mess though.

If this is for a research project, you should be able to convince someone to budget some materials; a Chinese 5kW power supply would be good enough.  It'll be medium frequency, 20-100kHz or whatever, but that will do fine with a susceptor.
To be honest, I haven't asked. I figured building it for $400 would be the cost of my education.

Or you can go completely the other way, see if microwaves work.  Near-1kW power supplies at 2.45GHz are conspicuously cheap and available. ;D

Never underestimate the usefulness of a microwave oven. 8) Hadn't thought of this, I'm definitely going to throw a sample into a spare I found to see what happens.

(FYI -- you may recognize me from having built a 5kW supply back in 2010.  After that, I designed a family of 400kHz power supplies up to 50kW.  In a small team, we took the better part of 6 months to get a working demo, and the better part of 2 years to commercialize the product.  Even if you don't need as many features as an industrial unit, you will still need a few minimum features, like tuning or load control, and fault protection.  This may take an average EE, heh, well... really, EE or not, it will take many more years to develop the knowledge necessary to design and operate something like this.)

Tim

400kHz and 50kW! That'll get you halfway to the moon on its own! What were they used for?
Admittedly I don't have a lot of time in this field, but I'm always looking to expand my understanding. I don't plan on commercializing the power supply per se, but there are a few terrestrial applications of the tech I would like to explore further. Any chance you would be willing to send over a demonstration unit?  ;D

- Mike

Online klugesmith

  • High Voltage Engineer
  • ****
  • Posts: 271
  • Karma: +6/-0
    • View Profile
Re: 500Khz Induction Heater - Need help with Driver
« Reply #3 on: January 18, 2020, 06:53:02 AM »
The bit about microwave ovens reminds me of a demonstration now easy to find on youtube.
Which I might finally try at home this weekend, in a MWO remnant saved for that kind of thing.

You can melt part of a glass bottle in a MWO, by preheating with a torch before closing door and turning on the waves.
At low red heat, I think, the conductivity gets high enough to absorb 2450 MHz energy without depending on dielectric loss.

Am reminded of Nernst glower lamps of more than 100 years ago, whose incandescent ceramic rods needed to be preheated.
Are today's silicon carbide heating elements a direct descendant?   Perhaps now doped, like other semiconductor materials, to make the room temperature resistivity low enough to be useful?

-Rich


« Last Edit: January 18, 2020, 06:55:19 AM by klugesmith »

Offline T3sl4co1l

  • High Voltage Technician
  • ***
  • Posts: 153
  • Karma: +9/-0
    • View Profile
Re: 500Khz Induction Heater - Need help with Driver
« Reply #4 on: January 18, 2020, 08:51:26 PM »
True, but the stuff is wicked nasty when its melted. even eats through PtRh. The goal is to try and touch the molten stuff as little as possible.

Hmm.  Although, there are a lot of counterindications for Pt and alloys; it's inert to a lot of things, but hardly everything.  Alkaline fusions being one of the more common examples in the lab.

That's why I mentioned zirconia, it's very insoluble in silica slags.  To the point where glass melting furnaces made with the stuff can function for a decade before repairs are needed; not to mention the glass is clear, not filled with tiny specks as the brick erodes. :)

I wonder if carbides would be suitable as well.  Silicon carbide is normally inert in air due to a passivating layer of silica, which of course doesn't count for anything in a slag melt; if the slag is oxidizing, it should erode relatively quickly (and the occasional bubbles of CO or CO2 won't be great; also, obviously, graphite would be even more prone to this), but if it's reducing, that won't be a big deal.

Hm, probably regolith should be sintered or fused before final melting, to homogenize it -- the first melting would be non-equilibrium, so even if it's reducing on average, it could still wear SiC, y'know?

There are some other, more unusual compositions, for special purposes, like chromia brick and phosphates, but I don't know that they are relevant here.  They may be more for rough handling of molten metal, for example.  Some very tough materials out there.  Magnesia graphite brick is used for steelmaking (extremely refractory, tough, and the alkalinity absorbs sulfur and phosphorus impurities), but isn't so good on slag resistance (steel ladles and furnaces use different materials at the slag line).

I'm probably half remembering a lot about these things, but suffice it to say, there is a wealth of knowledge out there on refractories, and vacuum furnaces have to deal with this too of course, making for a very similar environment.  Guarantee there's a lab somewhere that's worked with stuff far more corrosive, and figured out which materials handle it.

There's probably more than a few NASA sponsored research programs that did this, too. :)


Quote
400kHz and 50kW! That'll get you halfway to the moon on its own! What were they used for?
Admittedly I don't have a lot of time in this field, but I'm always looking to expand my understanding. I don't plan on commercializing the power supply per se, but there are a few terrestrial applications of the tech I would like to explore further. Any chance you would be willing to send over a demonstration unit?  ;D

Heh, you'd have to ask Radyne.

Typical use is industrial processes, like metal joining and heat treating.  At 50kW, you can heat the surface, a mm or two, of a piece of steel about the size of your wrist, then spray it with water to quench-harden it.  It's a much faster, more repeatable and more controllable process than chemical case-hardening.  Controllable in the sense that, case hardening is very superficial, unless you're willing to cook it for days to diffuse the hardening elements deeper; here, you just dial in the heating duration and that's your depth.

The higher frequency is more suitable to conductive materials like aluminum and copper, and heating thinner surfaces of steel.

Lower frequencies deposit heat deeper into the material.  There was one job while I was there, some kind of oversized trailer hitch about the size of your head.  They needed a case depth of 6mm or something like that.  They used a 3kHz power supply (up in the offices, you could definitely tell when they were running it), delivering a few hundred kW.  What's neat here is, the magnetic field is so strong, the steel saturates magnetically.  Skin depth depends on permeability of the material, being shallower in permeable materials like steel; but when it's saturated, the magnetic field goes much deeper.  Magnetic steel also absorbs a lot more power (hysteresis loss).  As a consequence, the temperature beneath the surface can actually rise higher than the surface temp, following the boundary where the steel is going nonmagnetic (above the Curie temperature, 750°C or so).  This combination of physics, with the right choice of power level and duration, makes for an unusually consistent hardening process.

Also, it's convenient that steel undergoes its important transformation, at almost exactly the same temperature it becomes nonmagnetic.  Quenching this high-temperature phase, causes it to become very hard; cooling it slowly leads to a soft (normalized or annealed) condition.  So you can control the hardness through local heating and controlled cooling rate.  (Compare to chemical case hardening, where you start with a soft alloy that does not harden when quenched, then introduce alloying elements, principally carbon and nitrogen, into the surface, which allow it to quench-harden.)

Tim

Offline Mysteriousduck

  • High Voltage Enthusiast
  • *
  • Posts: 5
  • Karma: +0/-0
    • View Profile
Re: 500Khz Induction Heater - Need help with Driver
« Reply #5 on: January 19, 2020, 04:06:36 AM »
That's why I mentioned zirconia, it's very insoluble in silica slags.  To the point where glass melting furnaces made with the stuff can function for a decade before repairs are needed; not to mention the glass is clear, not filled with tiny specks as the brick erodes. :)

I wonder if carbides would be suitable as well.  Silicon carbide is normally inert in air due to a passivating layer of silica, which of course doesn't count for anything in a slag melt; if the slag is oxidizing, it should erode relatively quickly (and the occasional bubbles of CO or CO2 won't be great; also, obviously, graphite would be even more prone to this), but if it's reducing, that won't be a big deal.

Hm, probably regolith should be sintered or fused before final melting, to homogenize it -- the first melting would be non-equilibrium, so even if it's reducing on average, it could still wear SiC, y'know?

[...]

I'm probably half remembering a lot about these things, but suffice it to say, there is a wealth of knowledge out there on refractories, and vacuum furnaces have to deal with this too of course, making for a very similar environment.  Guarantee there's a lab somewhere that's worked with stuff far more corrosive, and figured out which materials handle it.

There's probably more than a few NASA sponsored research programs that did this, too. :)


[...]
Tim

Thanks for the tip on zirconia, I'll have to look a little further into that. I've got a second edition of 'Metallurgy and Metallurgical Problems' that I've paged through a few times, but most of it has to do with wwII era steel industry practices. (though the section on thermochemstry and thermophysics is relevant) The only NASA literature I've been able to find are guys in Florida using joule heating to run molten salt electrolysis (a very nifty trick that I was unaware of that's been around for a century) and studies done to hypothetically melt the stuff for subterranean energy storage. I'll look harder and see what I can find.

If joule heating is the way to go, I may have to pivot... however again I think induction would prove to be a more feasible tech in actual lunar conditions as it doesnt require direct contact with reactive molten material. (lots of oxides in there) Maybe joule heating or solar concentration to get it hot enough to pellitize, then melt further for refinement and casting? I'm going to run some trials next week, I'll keep the thread updated on any progress made. Thank you all again for the input! :)

OT: Is Uzzor2k still on here? I have some questions regarding the op-amps used in his design. http://uzzors2k.4hv.org/index.php?page=ihpll1

-Mike

High Voltage Forum

Re: 500Khz Induction Heater - Need help with Driver
« Reply #5 on: January 19, 2020, 04:06:36 AM »

 


* Recent Topics and Posts

post Re: Rogowski current probe
[Laboratories, Equipment and Tools]
Hydron
September 29, 2020, 09:50:02 PM
post Re: Half bridge sstc help with troubleshoot
[Solid State Tesla Coils (SSTC)]
Zipdox
September 29, 2020, 09:24:24 PM
post Re: Rogowski current probe
[Laboratories, Equipment and Tools]
flyglas
September 29, 2020, 09:12:52 PM
post Re: Voltage tripler for my medium DRSSTC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
octopus1
September 29, 2020, 08:52:12 PM
post Re: Rogowski current probe
[Laboratories, Equipment and Tools]
Hydron
September 29, 2020, 08:13:43 PM
post Re: Rogowski current probe
[Laboratories, Equipment and Tools]
klugesmith
September 29, 2020, 08:08:22 PM
post Re: Rogowski current probe
[Laboratories, Equipment and Tools]
flyglas
September 29, 2020, 07:45:22 PM
post Re: Voltage tripler for my medium DRSSTC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
klugesmith
September 29, 2020, 07:23:27 PM
post Re: Half bridge sstc help with troubleshoot
[Solid State Tesla Coils (SSTC)]
costas_p
September 29, 2020, 06:46:33 PM
post Re: Voltage tripler for my medium DRSSTC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Hydron
September 29, 2020, 05:23:17 PM
post Re: Welcome new members, come say hello and tell a little about yourself :)
[General Chat]
KenM_Whitearc04
September 29, 2020, 03:15:57 PM
post Voltage tripler for my medium DRSSTC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
octopus1
September 29, 2020, 02:40:28 PM
post Re: Half bridge sstc help with troubleshoot
[Solid State Tesla Coils (SSTC)]
davekni
September 29, 2020, 05:44:51 AM
post Re: Half bridge sstc help with troubleshoot
[Solid State Tesla Coils (SSTC)]
plasma
September 29, 2020, 04:13:43 AM
post Re: Polyphonic Interrupter with STM32F103C8T6 (Bluepill)
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
Max
September 28, 2020, 11:48:25 PM
post Re: Half bridge sstc help with troubleshoot
[Solid State Tesla Coils (SSTC)]
costas_p
September 28, 2020, 09:56:15 PM
post Re: Easy to build Polyphonic MIDI Interrupter
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
futurist
September 28, 2020, 09:52:47 PM
post Re: Half bridge sstc help with troubleshoot
[Solid State Tesla Coils (SSTC)]
davekni
September 28, 2020, 09:50:44 PM
post Re: Half bridge sstc help with troubleshoot
[Solid State Tesla Coils (SSTC)]
costas_p
September 28, 2020, 09:16:53 PM
post Re: Half bridge sstc help with troubleshoot
[Solid State Tesla Coils (SSTC)]
Zipdox
September 28, 2020, 09:09:46 PM
post Re: Half bridge sstc help with troubleshoot
[Solid State Tesla Coils (SSTC)]
costas_p
September 28, 2020, 09:03:23 PM
post Re: Half bridge sstc help with troubleshoot
[Solid State Tesla Coils (SSTC)]
Zipdox
September 28, 2020, 08:50:33 PM
post Re: Stroboscopic photography & motors
[Digital Compact Camera]
Zipdox
September 28, 2020, 08:49:09 PM
post Re: Half bridge sstc help with troubleshoot
[Solid State Tesla Coils (SSTC)]
costas_p
September 28, 2020, 08:40:11 PM
post Re: Stroboscopic photography & motors
[Digital Compact Camera]
klugesmith
September 28, 2020, 07:36:45 PM
post Re: DRSSTC Tuning for Music vs Big Sparks?
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Uspring
September 28, 2020, 07:07:29 PM
post Re: Stroboscopic photography & motors
[Digital Compact Camera]
Zipdox
September 28, 2020, 06:41:56 PM
post Re: Polyphonic Interrupter with STM32F103C8T6 (Bluepill)
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
Zipdox
September 28, 2020, 06:28:50 PM
post Re: Stroboscopic photography & motors
[Digital Compact Camera]
klugesmith
September 28, 2020, 05:14:51 PM
post Re: Polyphonic Interrupter with STM32F103C8T6 (Bluepill)
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
Max
September 28, 2020, 04:44:56 PM
post Re: Stroboscopic photography & motors
[Digital Compact Camera]
TMaxElectronics
September 28, 2020, 10:50:31 AM
post Re: Stroboscopic photography & motors
[Digital Compact Camera]
plasma
September 28, 2020, 09:15:09 AM
post Re: Stroboscopic photography & motors
[Digital Compact Camera]
Zipdox
September 28, 2020, 09:09:45 AM
post Re: type of driver
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Mads Barnkob
September 28, 2020, 08:39:30 AM
post Re: Stroboscopic photography & motors
[Digital Compact Camera]
klugesmith
September 28, 2020, 08:16:46 AM
post Re: Polyphonic Interrupter with STM32F103C8T6 (Bluepill)
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
Zipdox
September 28, 2020, 01:27:07 AM
post Re: Polyphonic Interrupter with STM32F103C8T6 (Bluepill)
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
TMaxElectronics
September 28, 2020, 01:14:46 AM
post Re: Rogowski current probe
[Laboratories, Equipment and Tools]
flyglas
September 28, 2020, 12:30:58 AM
post Re: Rogowski current probe
[Laboratories, Equipment and Tools]
Hydron
September 27, 2020, 11:48:35 PM
post Polyphonic Interrupter with STM32F103C8T6 (Bluepill)
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
Zipdox
September 27, 2020, 11:43:14 PM
post Re: Rogowski current probe
[Laboratories, Equipment and Tools]
flyglas
September 27, 2020, 11:01:13 PM
post type of driver
[Dual Resonant Solid State Tesla coils (DRSSTC)]
octopus1
September 27, 2020, 10:51:03 PM
post Re: Easy to build Polyphonic MIDI Interrupter
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
Max
September 27, 2020, 10:44:43 PM
post Re: MIDI videos from my DRSSTC1/2/3 demonstrations/shows
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Mads Barnkob
September 27, 2020, 09:44:58 PM
post Re: DRSSTC Tuning for Music vs Big Sparks?
[Dual Resonant Solid State Tesla coils (DRSSTC)]
davekni
September 27, 2020, 09:20:32 PM
post Re: Easy to build Polyphonic MIDI Interrupter
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
futurist
September 27, 2020, 07:49:49 PM
post Stroboscopic photography & motors
[Digital Compact Camera]
klugesmith
September 27, 2020, 07:08:06 PM
post Re: Led Closeups with Mobile phone and lens .
[Smart Phones]
klugesmith
September 27, 2020, 06:43:46 PM
post Re: DRSSTC Tuning for Music vs Big Sparks?
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Mads Barnkob
September 27, 2020, 06:42:46 PM
post Re: Finding accelerometer location inside a phone
[Smart Phones]
klugesmith
September 27, 2020, 06:36:41 PM

Sitemap 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 
SimplePortal 2.3.6 © 2008-2014, SimplePortal