Novice-Built Induction Heater (In 14 days?) / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Novice-Built Induction Heater (In 14 days?)

previous 1 2 3 next

Move Thread

LAN_403

AwesomeMatt

Fri Aug 22 2014, 02:24PM

Registered Member #4454 Joined: Sun Feb 26 2012, 12:47AM
Location: Western Canada
Posts: 74

Steve - You've got quad 6666s post count. Woo.

Good News - I fixed the water pump I had. 170 gallons/hour (700 L/hr). Should be enough.

Bad News - The Ebay hillbillies in Ohio I wanted to buy my caps on won't sell to dangerous foreigners in Canada. No exceptions. I could re-ship but that turns already expensive shipping into a $100 shipping cost.

Kinda bummed and scrambling for alternatives.

Ash Small

Fri Aug 22 2014, 02:32PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

AwesomeMatt wrote ...

I don't for example, understand why 60hz (or whatever) isn't just fine.

You can use 60Hz, but you'd need HUGE capacitors to get the resonant frequency of the work coil plus tank down to 60Hz.

Modern commercial induction furnaces run on 50-60Hz, but resonant frequency is lowered by the amount of metal to be melted, which contributes to the inductance.

The only way of tuning these systems is by increasing/decreasing the capacitance in the tank.

The induction heaters commonly built here have much higher resonant frequencies, and tuning is normally accomplished by adjusting the driving frequency.

AwesomeMatt

Fri Aug 22 2014, 06:00PM

Registered Member #4454 Joined: Sun Feb 26 2012, 12:47AM
Location: Western Canada
Posts: 74

wrote ...
You can use 60Hz, but you'd need HUGE capacitors to get the resonant frequency of the work coil plus tank down to 60Hz.

I thought there would be more to it than that.

wrote ...
The induction heaters commonly built here have much higher resonant frequencies, and tuning is normally accomplished by adjusting the driving frequency.

It seems like commercial ones use either really low frequency for mass melting, or really high frequency (low Mhz) for specialty hardening. And amateurs reside in the middle where switching losses aren't too bad and cap costs are low.

...

I could have caps delivered cheaply, but not in time for me to use them.

Maybe I'll try that, then order some smaller, lower powered caps for the 120V operation (our Makerfaire in 2 weeks)... then even split it off into two separate heaters later. Which means all I need are some quick 'n dirty lower power caps for the 1.5kW version. Probably an MMC array is easiest.

Anyone spotted any suitable deals recently?

AwesomeMatt

Sat Aug 23 2014, 05:42AM

Registered Member #4454 Joined: Sun Feb 26 2012, 12:47AM
Location: Western Canada
Posts: 74

Shopping for tank caps, trying to find suppliers on this side of the border (not easy) or at least this coast. I don't really know what I'm looking for. Some things look impossibly cheap but I can't figure out why I wouldn't want them.

Can anyone give me suggestions on the types of capacitors that are appropriate? I know what voltage and capacitance I'm looking for, but beyond that I'm kinda stumped.

Well, I don't even know that. I think ~10-15uF is a reasonable target. And, voltage less sure. 300VAC? 600VAC? I'm not sure how high it will reach. I've been presuming 300VAC but I might be wrong.

So far I've been looking at Polypropylene Film capacitors because most people have pointed me in that direction.

Sometimes they only have ratings listed in DC, so I'm ignoring those.

Most often they do not have an amp rating or a frequency they can handle so I don't know how to guess on that.

Here's a sampling of my short list:

#1 -

<-- WIMA MKP10. 6.8uF, 180VAC, audio caps. They're $15 for 2, I would need a string of 2 series, 4 parallel, for 13.6uF @ 360VAC. $60 total + shipping.

#2 -

<-- Polypropylene Film. 0.56uF, 275VAC. 100 pack for $15. Datasheet. I would need strings of 2 in series, and 20-25 in parallel. 14uF @ 550VAC.

#3 -

- (Local Supplier search/scroll for CDE 930C6W1K) - PolyP 1uF, 250VAC. $5 each. I would need 2x in series, 10 strings in parallel for 10uF @ 500Vac. $100 total.

There's lots more of the WIMA MKP caps, but I'm not sure if they're even suitable.

When I'm less busy, I'd like to write a guide for this. It seems everyone who does this stuff already knows what they're looking for, and everyone else keeps asking the same questions. If there's already a "Here's how to shop for caps and what to look for" guide, link away 'cause I haven't been able to find one.

Sigurthr

Sat Aug 23 2014, 06:06AM

Registered Member #4463 Joined: Wed Apr 18 2012, 08:08AM
Location: MI's Upper Peninsula
Posts: 597

I can vouch for the WIMA MKP caps, they're very good. The FKP are even better though.

If you can't find an AC rating you can always take 0.707 x the DC rating to get something similar, remember that you want some overhead anyway.

Since you're pursuing series resonant design you have to have a very high voltage rating, as resonant rise will cause a very high voltage to be built up across the tank. Likewise, Irms rating isn't nearly as critical since the current flowing through the capacitors will be much, much smaller than in a parallel resonant arrangement.

Steve Conner

Sat Aug 23 2014, 02:09PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

I wouldn't waste money on anything besides conduction-cooled capacitors designed for induction heating. Just keep an eye on Ebay and eventually some will turn up surplus.

Both series resonant and LCLR induction heaters end up with a tank voltage of a few hundred volts RMS. This is determined by the requirement to use copper pipe for the work coil for water cooling. You end up with a thick conductor that can only accommodate a few turns in the space available for amateur-sized workpieces, so the impedance ends up low, and you have low voltage, high current and large capacitance.

If you tried to increase the tank voltage, you would end up with more turns of thin wire that couldn't be cooled, or a coil so big that it wouldn't couple much energy into the workpiece. You could make the workpiece bigger, but then you would need more power than a domestic supply could deliver to melt it. To have a chance of melting steel you need at least 2kW of generator power per square inch of surface area.

In the series resonant design you use a step-down transformer to transform the output of the inverter to low voltage and high current. The one I built was something like 20:1. 160V at 10A in the primary gave 8V at 200A in the (single-turn, water cooled) secondary. Resonant rise in the work circuit increased that 8V to a few hundred again.

There are some books on this out there. One I found helpful was "Induction Heating" by Frank W. Curtis, available from Lindsay Books. It's a reprint of a 1940s text so don't expect any MOSFET circuits in there.

AwesomeMatt

Wed Aug 27 2014, 12:48AM

Registered Member #4454 Joined: Sun Feb 26 2012, 12:47AM
Location: Western Canada
Posts: 74

Well I have a friend who'll ship and loan/sell me some suitable caps, at least until after the Faire.

Running out of time, 10 days.

<-- Uzzor's new driver, I'll be cloning this.

<-- IAMSMOOTH's big power side circuit. I'll be cloning most of this.

I would ideally like to copy IAMSMOOTH's whole circuit with the arduino and all, but his code isn't posted anywhere and I don't trust myself to write it from scratch, hence, going with Uzzors' pre-driver.

Since I'm only building a 6KW unit, not a 10KW unit, I was hoping to save some money on the MOSFETs, not copy IAMSMOOTH's circuit entirely... but I don't know the specs I'd need.

He uses higher voltage (has a doubler) to feed his inverter.. I'm using 300VDC. I'm not sure if the MOSFETs need extra, (400, 500, or 600V instead of 300?) or if you can just rate them for what you give them. I'm also not sure about the amps. The breaker is 30A, but can I just choose MOSFETs that are 30A, or should I include safety margin (say, 50 or 60A?).

I also need:

- 1x LM393 (op amp?)
- 2x LM358 (op amp? why different than 393?)
- 1x TC4421 (gate driver)
- 1x TC4422 (gate driver, inverting)
- Maybe some superfast diodes, if I can't salvage from computer PSUs.
- 4x 18v zeners (how many amps?)

I presume I can just copy these components outright, no need to tweak them for different MOSFETs or whatever else in my design is slightly different.

I started taking pics of stuff:

2001

Transformers. Might be able to repurpose these for GDTs, if needed.

2002

Ferrite cores. 1.4" ID, 2.4" OD, and the total stack of 5 is 2.5" tall. Shoudl be ballpark good enough, I'll calculate that soon, but I have much bigger backups if need be.

2003

COils/Chokes. Someone suggested these would make good GDTs. They suggested to use the ones with thin wire, not the heavier 16g wire ones because they'll saturate as transfomers.

2004

Baby water pump. 0.04A.

2005

Double-sided circuitboard blank. A few square feet.

2006

7.5A variac. I have a 12.5A laying around somewhere too. Good enough to skip building a giant Buck regulator for the 120V demo.

2007

Big water pump. Good for 170 gallons/hour (2.8 gallons/minute).

It's no induction heater, but it's on it's way.

IamSmooth

Wed Aug 27 2014, 02:41AM

Registered Member #190 Joined: Fri Feb 17 2006, 12:00AM
Location:
Posts: 1567

You can levitate aluminum using my 2-3kw circuit and manually tune with the assistance of a PLL. It will run off of 120vac. You can build it in a day. You can get away with just air cooling for your mosfets. You can just copy this entire circuit and accomplish your task.

Here is a video of the 3kw unit doing what you want using manual tuning (no arduino):

It is the first schematic on the tutorial.

3kw unit here:

Levitation section of tutorial here:

The key to levitation is copying the coil topology on my website tutorial. This appears at the very end under the levitation section. Copper and steel levitation require close to 10kw. Sustained levitation during a completely molten state requires over 12-15kw. You better have some good mosfet cooling for this.

AwesomeMatt

Wed Aug 27 2014, 08:43AM

Registered Member #4454 Joined: Sun Feb 26 2012, 12:47AM
Location: Western Canada
Posts: 74

Smooth - I've read your tutorials and seen your vids 10x each. Very helpful. Some of it is over my head.

I leaned towards Uzzor's driver because it does more automatically and needs less setup. Or appears to. I'd ideally add an arduino just like you did eventually, when I'm not under such a time crunch.

AwesomeMatt

Thu Aug 28 2014, 03:15AM

Registered Member #4454 Joined: Sun Feb 26 2012, 12:47AM
Location: Western Canada
Posts: 74

TL;DR - Boring philosophy on novices with electronics.

So I went to order parts today. Even with a schematic I was copying, this was a lot more work than I anticipated, with a lot less certainty that I did it right.

For example:

- I need a simple op-amp. LM393.
- I punch it into Digikey... out comes 138 different options. O.o
- Being a novice, I'll just use big clumsy through-hole parts. 138 --> 12.
- Wanting cheap and quick shipping, local (Canadian) stock only. 12 --> 7.
- Okay, 7 to choose from. Main difference is "Differential" vs. "General Purpose". Uh oh.
- I wonder: "Does any of this matter?", "If I picked randomly, would it still work?", "Would any op-amp at all (LM393 just being a generic) work just as well?
- Inputs range from either 2mV or 5mV, and end at either 30V or 5V. I don't know what voltage range they're being used for in this circuit.
- Higher voltage range seems better as a guess, but, if it was designed for the lower range I'd lose... precision? Would my range be off? Would it matter at all?
- Max input bias is either 0.1uA or 0.25uA.. hrm.
- Output current is 16, 18, or 20mA. I know what this means, but not what is required.
- I sort by # in stock and pick the one that seems most popular.
- I have no clue if that was the correct decision.

Suddenly I went from "Great, I'll just grab this part" to "There are 20 contextual questions I don't know how to answer."

This is repeated for every part.

The trouble of blindly following instructions without context is the inability to solve my own problems. So if instructions aren't perfect ("Use these exact parts"), or my use differs slighly, I've got nothing to go on.

This is one of the things that's always intimidated me about electronics. I have no interest in basic LED blinky stuff, but the stuff I would use I don't understand.

I'm sure I have the ability to understand more complicated circuits, but my brain works differently than most electronics people's. I get nothing out of a static schematic. But if I can see how the circuit evolves from both ends "Input is around here, Output needs to be this"... and how and why parts fall into place in sequence... electronics is very simple and building-block like.

I see a schematic as a narrative. "In order to get E we need to have D, in order to set D we need to control C. In order to control C we need to supply it with B, and B can be had from A." But when I see the end result without those choices, it makes no sense to me.

But by the point I could answer all these questions, I wouldn't need to copy someone else's circuit. I'd just know how to create one myself.

previous 1 2 3 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob