Induction furnace

[AustinT]

Here is a link with some good info on induction furnaces

http://www.dansworkshop.com/Induction%20Heating.shtml

[Warpspeed]

DukerX,

What you say would be very true if the work coil and the job were both tightly surrounded within an efficient magnetic core structure. In other words they were were effectively both windings that were part of a properly constructed and efficient high frequency transformer.

But with no really efficient magnetic coupling, and a very open magnetic air path, only a very small portion of the magnetic flux passes through the work. In fact, it is even worse than you may imagine ! The large eddy currents flowing through the work actually sets up it's own counter magnetic field that opposes and repels the field from the tank coil.

Technically this is called leakage inductance. Any magnetic lines generated by the primary (tank) that do not pass through the secondary (work) therefore cannot create any useful heating output. It is this weak and inefficient coupling of energy into the work that requires such an enormous energy buildup in the tank to create a useful energy flow into the work.

Lerman, you are quite right. Here in Australia the neutral wire is at ground potential (almost), and there are three 120 degree phases, each at 240 volts with respect to neutral. Without this bucking transformer to reduce the voltage, I would have ended up with +340 and -340 volts dc. That is 680 volts total, and far higher than I feel comfortable with, at least initially to begin with in this prototype.

The alternative would have been a bridge rectifier running between one phase and neutral. That has many disadvantages, including having a large ac component with respect to ground on both sides of the rectified dc. That makes looking at waveforms with an oscilloscope rather difficult, and it is somewhat more dangerous. Another limitation is only being able to get 340 volts dc that way, which is rather lower than ideal. And lastly, it would mean pulling all the power from only one phase with a very high ripple current requiring an enormous and potentially dangerous electrolytic capacitor bank.

I really wish to use all three phases to spread the load, and have a low output ripple requiring minimal filter capacitance. My buck transformer also allows me to adjust the dc output voltage upwards in the future just by shifting some transformer tappings around.

A bucking or autotransformer is also going to be a lot smaller than a true full isolation transformer. A 10 kw full isolation transformer will weigh about 200 to 250 Lbs and would not be exactly cheap. As neutral here is grounded, not having full mains isolation is not really a great disadvantage.

Getting reliable reasonably well regulated dc power at a suitable voltage for an induction heater project is not as simple as it may first appear. It requires a bit of thought and ingenuity no matter how you finally go about it. It is just one of several rather interesting problems to overcome in this type of project.

Photographs and schematic circuits will come later, once I have various sections built and working. At this stage I have plenty of ideas, and a plan of action mapped out, but no actual hardware assembled yet to show.

[Warpspeed]

Originally Posted by AustinT Here is a link with some good info on induction furnaces

Dan, here are a couple more that may be of interest:

http://www.richieburnett.co.uk/indheat.html

http://webpages.charter.net/dawill/t..._IndHeat1.html

[Tekko]

I´m currently working on an IH that actually does about 1kW at full power (what the transformers i use are capable of)

My biggest problem is the imp matching trafo heating up and the work coil getting hot so water cooling of atleast the work coil is a must, esp is heating stuff red hot.

[Warpspeed]

What are you actually running there Tekko ? Come on, a full and complete confession is in order <grin>.

It appears that all the larger commercial induction heaters use water cooled tank coils as normal practice. If you work out the heating effect of all that circulating current passing through a thin walled copper tank coil, the resistive power loss will be fairly high. Making the pipe wall thicker is not a solution, because electrical skin effect forces all that current to flow right near the surface anyway.

Once I get my dc power supply up and running, the next job to tackle will be the closed circuit water cooling system. After doing some initial speculative calculations, I may need to wind my feed choke out of small bore copper pipe as well, but am not yet certain if it will actually need to have water circulating through it. With all that cooling water being pumped around, I may as well cool my transistors with some home made water cooled heat sinks as well. That should simplify the whole thermal design considerably, and solve many potential problems.

[Tekko]

Here are some pics:

http://diymania.hv4all.com/PLL%20controlled%20IH/

[awemawson]

OK a bit of background. Total amateur who has had a propane furnace for years. A couple of years back built a prototype induction furnace good for a few hundred watts, and intended scaling up to 15KW as that was the max electrical power I have in my home foundry. Found a commercial unit on ebay which I am now running so stopped the development.

System so far: I have a 100KVA 3 phase 415v diesel generator powering a 120KW CFEI (French) induction furnace driver for which I have two furnace 'bodies' both made by Radyne. One is a 12Kg capacity (of stainless steel) inverting body intended for lost wax casting, the other is a 27kg tilting body intended for conventional pouring. The copper coils in the bodies are tubes with water pumped round at a great rate of knots with a 1HP Grundfos CR8 pump though a heat exchanger which I cool using a 35KW commercial water to air chiller unit driven off the mains.

The CFEI driver takes in 3 phase 415v and rectifies it to a nice lethal DC rail, which is then chopped by an 'H bridge' of thyristors at about 3Khz. The system is microprocessor driven. A second set of thyristors 'ping' the tuned circuit of tank capacitors and furnace coil to determin the initial frequency (which varies dependant on how much metal is in the pot and also how hot it is) and then the micro tracks the charge keeping it on resonance.

Connection from the electronics unit to the furnce body is by four (two out and two back) 35mm sq mm cross section welding cables threaded through 3/4" 'Brewers Hose' (no carbon content) and chilled water is pumped down the hoses. There are several tens of thousands of amps circulating in the tuned circuit.

It's been a long journey but I'm getting there ! Both bodies have had to be re-crucibled - they are perminent fixtures, I end up having to have bespoke ones made in the Cech republic, and the hydraulics and pneumatics have had to be totally rebuilt.

[Warpspeed]

Mr Awsome Mawson, GOLLY !

And you class yourself as a total amateur, hehehe.

[Shawn D]

Hello everyone,
I am an avid hobbyest in several different fields. However, I also work. And my job is at an Induction Heating manufacturing facility. (I do field service work on the heaters).
I may be able to answer some of the "myths" as related to Induction Heating.
While I didn't read the entire 4 pages of this thread yet, I have enjoyed seeing the Ideas and realisms that you have come to.

By the way, as some of you have already pointed out, ANY conductive material can be heated with induction. However, the non-ferrous materials do not heat as efficiently as do magnetic materials.
Most of the time when Al is being melted with induction, a crucible made of graphite, or other material that can be heated fairly efficiently is used. Then the Al is heated both by eddie currents and by thermal transfer from the crucible. Of course, the crucible also has to be made of a material that can withstand the melting temperature of the Al. And another nice side effect of using induction is that the Al is stirred by the magnetic fields.

There is a fairly new community at myinduction.com it appears to have both hobbyests and professionals. While the forum is not as elaborate as most, it is a great place to get some questions answered. The website also has valuable information in the "Tools" section. I am pretty new to that community also, but it looks to be a good start.

[Warpspeed]

Hello Shawn, welcome to the CNC Zone Forum. I am pretty new here myself. I am just about to register in at the myinduction.com Forum, so expect to see you over there as well.

My home built induction heater project is coming along very slowly, but it is certainly progressing. This is going to take more like several months to complete, not just a few days or weeks, so for those interested in this, please be very patient.

I now have all the important large parts to build a 15Kw three phase dc power supply of plus and minus 220v (440v total @ 35A), and after a lot of scrounging and testing now also have a suitable water pump, fan, and water to air heat exchanger for the cooling system. There have been some rather big dramas trying to obtain a proper commercial tank tuning capacitor, and after spending several hundred dollars and wasting a month, I still have nothing. A company messed me around, accepted my money, then sent me the wrong part. I sent it back to be told it never arrived. I will wait another week (it may be still lost in the mail) before forking out even more money and trying all over again.

If I cannot obtain a suitably rated tank tuning capacitor for this project, then the whole project is killed stone dead before it even starts. But at least all this has given me plenty of time to think in more detail about the electronics. Here is an overall block diagram of the general idea:


The tank circuit forms part of a classical "Royer Oscillator", or more modernly known as a "current fed converter". This is a very efficient and low stress means of turning square waves into pure sine waves.

There will be two control loops. The first measures the amplitude of the built up resonant tank voltage, and uses that to control the constant current dc supply which feeds the H bridge current steering section. The constant current supply being a simple buck regulator run in current mode.

The second control loop senses the tank zero voltage crossings to operate a phase locked loop. The PLL then drives four bridged IGBTs, and steers the constant current through the tank circuit to maintain the circulating tank energy.

The tank energy should rapidly build up to a set voltage under no load, and it should require fairly little power to do that. As soon as some metal object starts absorbing significant power from the tank, the tank amplitude will fall, and more drive current will be fed into the tank to return the tank voltage to it's set level.

Being able to adjust the target tank voltage means I can control the power level over a very wide range, right down to zero.

This is a lot different to how others have attempted to build a home induction heater, but it is a far lower stress way of switching large amounts of power through a resonant tank, than the usual hard voltage switching with a series inductor.

[Geof]

I think Awesome Mawson is correct.

I have peeked at this thread of and on for a while. Not planning on building anything but interested because nearly forty years ago I ran a 500KW, 20KHz induction furnace at a research facility near Montreal doing experiments in continuous casting of copper alloys. Yes you need enormous capacitor banks to tune the system and the furnace coils are copper pipe for cooling water. I remember the day I had the furnace up to heat full with 500 lbs of molten copper and the idiot plant maintenance engineer shut off the water supply to the building without any warning. My system was busy flashing red lights and making loud siren noises while I ran down five flights of stairs to tell the idiot if I didn't get water soon he was going to see one almighty steam explosion.

I also discovered that stray high frequency magnetic fields and wire reinforced hydraulic lines don't get on well together. Hot hydraulic oil spraying onto molten copper makes a hell of a stink. I should point out I was not involved in the design so I can't take credit for this discovery.

The whole installation was built on a steel mezzanine floor in the pilot plant of the research center...yes steel, that magnetic stuff that is also conductive. The floor used to get too hot to walk on around the furnace.

I lasted six months in that job before I decided my continued health and sanity required a change.

[Warpspeed]

Thanks Geof, that confirms something that has been worrying me for quite a while. The magnetic field around the tank coil is going to radiate significantly beyond the actual work being heated. Any nearby electrical equipment or instrumentation could suffer rather badly from induced voltages and currents.

I would not like to eventually fire this thing up in my home electronics lab, and then discover I have destroyed several thousand dollars worth of nearby electronic test equipment.