Advice on designing and building an induction furnace for small-scale iron melting

dgfoster · May 20, 2019

Right up front I will confess that my knowledge concerning electronics, switching transistors and induction furnace design and operation is woefully inadequate. That said, my ignorance may be partially curable.

I am interested in building an induction furnace capable of melting 60 pounds of iron over a few hours time. I am very aware that their are any number of commercial units with capacities ranging in size from a pound or two of iron to a few tons. Units available in whatever size are usually designed to melt the charge in under one hour. While that is convenient and desirable, that short melt time is not necessarily essential. To achieve short melt times very large power inputs are required. I believe the required energy input to melt 1 ton of iron is roughly 500 KWH. So, to melt 50 pounds would require 25 KWH. If an ideal furnace could be designed (I think I have a fair handle on refractory insulation and general furnace design as I am currently using diesel to melt iron in a shop-made furnace and have been doing this for about 1 year.), I will assume that for the sake of discussion that a 5KW induction heater run over 5 hours might melt iron in 5 hours. (I know this is not a real world assumption but want to keep the question focused on the induction side of things.)

So, my question is how difficult and expensive would it be to design and build a 5KW 1000Hz power supply and heater coil? Is this potentially feasible and reasonably economical or am I tilting at windmills? Economical might mean 1000 dollars or less. Are there simple options I am overlooking for power supplies that might be re-purposed for my intended application? Is the obvious answer to just buy one of these and hook it up to an appropriately sized---9" diameter roughly---coil housed in a well-insulated furnace?

New 5KW 100~250KHz High Frequency Induction Heater Furnace Fast Shipping | eBay

Trying to reduce my ignorance,
Denis

PS The iron is used to cast camel back straight edges and prisms of my own design for sale for use in reconditioning precision machinery and equipment. Those castings range in weight from 20 pounds to close to 60 pounds.

The real Leigh · May 20, 2019

dgfoster said:
how difficult and expensive would it be to design and build a 5KW 1000Hz power supply and heater coil?

1000Hz is in the middle of the audio range. You can hear it.

The inductor size required to use that as a furnace would be huge.

The furnace you linked on evilpay operates at 100kHz, which is 100,000Hz.

- Leigh

SAF · May 20, 2019

I would be asking what is the input power requirements. 5KW output, likely means the input is higher.
Does it require 1 or 3Φ. Can it operate properly from 60Hz?
What would be required to remote the operating coil from the front of the unit?

SAF Ω

TDegenhart · May 20, 2019

I think you would be better off to find a reliable iron foundry.

There is more to this than meets the eye. I have been away from the foundry industry long enough that I can no longer advise accept to discuss this with people and companies that deal with induction melting equipment.

By the way, the 1000 Hz range is common in induction melting equipment.

Tom

9100 · May 20, 2019

My RF heater running at 490,000 cycles, one KW, will not melt iron because it quits when the iron gets over the Curie temperature.

Why not just get a pottery kiln?

Bill

dgfoster · May 20, 2019

TDegenhart said:
I think you would be better off to find a reliable iron foundry.

There is more to this than meets the eye. I have been away from the foundry industry long enough that I can no longer advise accept to discuss this with people and companies that deal with induction melting equipment.

By the way, the 1000 Hz range is common in induction melting equipment.

Tom

You understand that I migrated from a reliable foundry to casting my own due issues with quality control and pattern damage with a good local commercial foundry. My castings are of excellent quality (better than the commercial foundry by the owners own assessment) after perfecting my system for 18 months. Not bragging or BSing. It has been damn hard work, but it has paid off.

9100 said:
My RF heater running at 490,000 cycles, one KW, will not melt iron because it quits when the iron gets over the Curie temperature.

Why not just get a pottery kiln?

Bill

Iron needs to be poured at about 2500F. That is not in the cards for a pottery kiln. Aluminum, sure, but not iron.

Denis

TDegenhart · May 20, 2019

What is your current method of melting the iron? What way do you have for control of the chemistry and properties?

Tom

dgfoster · May 20, 2019

TDegenhart said:
What is your current method of melting the iron? What way do you have for control of the chemistry and properties?

Tom

I am using diesel---see first post. I have access to a spectrometer with which I can and have checked the composition of the iron I pour. But, let's stay with the original question so we don't go way off into the toolies. If you search PM you will find a number of posts regarding this endeavor and you have seen some of my work in the past and commented (positvely) on it.

Denis

Denis

TDegenhart · May 20, 2019

Yes I have. I didn't remember that the method of melting was brought up. Went back to post 1 and yes I see that.

I would be dubious of being about to do this on a $1000 ticket.

First order of business is to spend time researching the subject. While trying to get a cost estimate of a small melter, I ran across an article in Foundrymagazine how the frequency affect the melt. Other issues to study are the coupling of energy into the charge, type of crucible to used or if even used during the melt, nature of the waveform i.e. can it be switched or need it be a sine wave. Construction of the furnace itself. Can carbon steel be used and if so, are there circulating currents to be dealt with. Most likely other issues I haven't thought of. You might want to start with American Foundrymen's Society, magazine articles, manufacturers like Inductotherm.

Good luck, keep us up to date. If I find something I will pass it along.

Tom

bryan_machine · May 21, 2019

@dgfoster - I have a unit at least vaguely similar to the one in your post (don't recall specs would have to go and look) which I bought from the late Grant Sarver (importer of blacksmith tools among other things) (for a while Larry Langdon, also sadly passed, was selling them too I think.) [Both were in WA which is why I mention them.]

Some things to note - it requires a tig cooler or the like to run (to keep from cooking itself) - so in addition to refractory for your cupola, you will need cooling for the device - bigger unit likely requires bigger cooler. Total power required may be much larger than you'd think because the power to cool is so high.

Also, the class blacksmiths use (which again look similar to what you linked to) can make things like 1" steel forging hot very quickly - and melt through 1" steel stock not much slower - but only over a short length. And then they're off. You are talking about a continuous duty application.

Their effectiveness depends on having the steel being melted in a shape that the induction coils match well. What sort of shape will your iron stock be? (Billets? Broken up scap?) Getting that to align with the induction coil may be quite a trick. And in any case, you are talking about a pretty long continuous run rather than short "heats" as a smith would use.

I never got around to settting up my own unit in my own shop, I only saw it run in Grant's shop. But I do think all of the above applies to your application design.

Mine is single phase, and if memory serves takes at least 30 amps. So for the scale you are talking about, you'd likely need pretty big power service for the induction heater, a surprizing amount of power for its cooling.

memphisjed · May 21, 2019

The only advantage I see to induction furnace is atmospheric control, like total control with mixed gas or vacuum atmosphere for exotics. Coke is cheaper and self controls the atmosphere for cast iron, malleable, and hi-nickel iron alloys. coke is easy to store too.
The mini mills for steel (Nucor) all use arc because it is more efficient if you like the electron route. A cone 13 kiln can work to, 60# will take ages to bring up to temp.

bryan_machine · May 21, 2019

There are videos of people melting aluminum to poor into a mold using a tig torch (but nothing like 60#) I wonder if laying your source material in a refractory bed of some kind and then blasting it with an oxy-torch would at least start the process more quickly? Also, a fairly common sort of forge is a multi-burner propane or natural gas forge - wouldn't seem to be so far from that to outright melting iron - would that have an great advantage over your diesel arrangement?

dgfoster · May 21, 2019

bryan_machine said:
There are videos of people melting aluminum to poor into a mold using a tig torch (but nothing like 60#) I wonder if laying your source material in a refractory bed of some kind and then blasting it with an oxy-torch would at least start the process more quickly? Also, a fairly common sort of forge is a multi-burner propane or natural gas forge - wouldn't seem to be so far from that to outright melting iron - would that have an great advantage over your diesel arrangement?

No, diesel tends to be a fuel of choice though propane could be used. Diesel is easier to handle and maybe safer than propane.

memphisjed said:
The only advantage I see to induction furnace is atmospheric control, like total control with mixed gas or vacuum atmosphere for exotics. Coke is cheaper and self controls the atmosphere for cast iron, malleable, and hi-nickel iron alloys. coke is easy to store too.
The mini mills for steel (Nucor) all use arc because it is more efficient if you like the electron route. A cone 13 kiln can work to, 60# will take ages to bring up to temp.

Coke is good for a cupola, of course. Generally cupolas require several people to operate them they generate 100’s of pounds per firing typically. But I am running a diesel fueled device. Diesel is also easy to store. But, induction has many advantages. Those advantages may be out of reach unless I can figure out a way to give up speed of melt in favor of clean, quiet, cheap operating costs associated with electricity.

Tell me me about an electric cone 13 kiln. Most seem to top out at 10

Denis

TDegenhart · May 22, 2019

Having worked with cupolas, they are not suited for this application. Small cupolas are difficult to operate because of uneven airflow leading to bridging, they require at lease two or three people operate, they are hard to produce consistent metallurgy and require the lining to be repaired after each run. Unless they are run continuously, slag is a problem. Small amounts of coke may be difficult to obtain.

We had a #3 cupola at school that I helped run. When I worked at the Caterpillar foundry in East Peoria, they had cupolas. When Mossville was build, Cat moved to induction furnaces.

Tom

Robert R · May 22, 2019

dgfoster said:
So, my question is how difficult and expensive would it be to design and build a 5KW 1000Hz power supply and heater coil? Is this potentially feasible and reasonably economical or am I tilting at windmills? Economical might mean 1000 dollars or less. Are there simple options I am overlooking for power supplies that might be re-purposed for my intended application? Is the obvious answer to just buy one of these and hook it up to an appropriately sized---9" diameter roughly---coil housed in a well-insulated furnace?

/QUOTE]

Answer -Windmill Tilting

The penetration of a electromagnetic field in steel ranges from .05" at 400,000 hz to a 6. inch depth at 60 hz.

Older coreless induction furnaces were designed for operation at 60 hz. These require a liquid metal charge at startup and a tap changing system for the matching transformer for the coil and possibly a tap changer for the power factor correction capacitors Tap changing is required to adjust for crucible wall thickness and melt weight.

Newer systems, as of 40 years ago, use higher frequencies. Maybe up to 1 khz. (A guess) The higher frequencies allow the furnace to start with a cold charge, allow higher power densities in the melt, and allow power control by changing the frequency of the sine wave going into the L-C resonant circuit that forms the crucible induction coil- capacitor circuit..

One of the primary advantages of induction heating is that the melt is self stirring. The induction coil magnetic field produces a electric current in the melt which provides the heat. The induced electric current in turn produces a secondary magnetic field. The two magnetic fields force the melt to circulate.

Induction furnaces include a magnetic yoke surrounding the induction coil to retain the coil magnetic field. Otherwise there would be heating of the support frame as well as lower efficiency. The design of the coil requires some experience or the ability to copy an existing successful design.

The crucible geometry, wall thickness, melt weight, and required power density will determine the design of the coil and matching power supply.

A low cost improvised design might be put together using a government surplus auction 400 Hz aircraft standby motor generator set. I suppose one could adjust the speed of the motor part of the generator for the variable frequency and change the generator rotor excitation to adjust the voltage output. You will still need to understand how to design a induction coil resonant circuit, how to design the magnetic yoke, and how to work with a matching transformer.

There is also some work involved in the coil cooling system and the crucible tilt/pour mechanism. There are ready made small melt induction furnaces for the type of work that you are doing. These are available with a gate valve at the bottom of the crucible which allows for the direct filling of a mold without disturbing the dross at the top of the melt.

Reference: Metals Handbook, Ninth edition, Volume 15, Casting , Page 36 and pages 368 to 374

TDegenhart · May 22, 2019

Found a little information in the Electrical Engineers Handbook,10th ed, Fink and Carroll. Typical energy values for melting steel is 600 Kw-Hr per ton. Small furnaces a bit more. So a 200 lb charge would take roughly 60 Kw-Hr. Figure a two hr melt time, that's a 30 Kw power system. The time-Kw value is not linear, the longer the melt time the more the losses. The next piece of info is that starting from a cold charge requires either large pieces or high frequency. Reason being is that large pieces use eddy currents for the heating. Small pieces would depend upon hysteresis losses which vary as freq^1.6 (Steinmetz exponent). I haven't found a definitive value for frequency, but a 30 Kw power amplifier is nothing to sneeze at.

Tom

TDegenhart · May 22, 2019

dgf - send me your email address. I have a couple of Pdf files that may be of interest.

Tom

dgfoster · May 22, 2019

Done.and thanks

Street · May 23, 2019

See this crowd for a price, it may be easier to buy on in rather than build as you have no design, no research and development they have done the hard yards.

http://furnace.com.au/heat-processing-product.asp?Heat+Processing=Induction&FurnaceID={2579C329-C136-49B4-88AB-817A561D7182}

oughtsix · May 23, 2019

Look at all the sub $2K induction heaters on ebay. Look at the size of the coils. How are you going to shove 60lbs of iron in a coil that size and heat it?

You have to tune the circuit for the coil you are using. You can probably spend a year trying to calculate the parameters for the coil that will be big enough to melt 60lbs of iron with any sort of efficiency. Like 9100 - Bill said above the characteristics of the coil change as the metal goes from below the curie point to above curie. Curie is the temperature at which a metal looses its magnetic properties. So to melt a metal with an induction heater you need a coil that is a compromise between the below curie and above curie efficiencies. This is what proper induction heater design is about, designing a circuit and coil that are efficient enough to get the job done without having to use a massive power supply to over come the inefficiencies of your design.

Your table napkin math is assuming that you can come up with a design that is 100% efficient. Real life is far from 100% efficient... like I pointed out above you might be able to design a coil that has decent efficiency below curie but as soon as your metal reaches the curie temperature its power transfer efficiency is going to be so low that the heating will all but stop. So you are going to need a MUCH large power supply than your simple calculations suggest.

This is my Induction heater. It is computer controlled and monitors the intersection of the area under the voltage wave form and the area under the current wave form which is the actual power going into the metal. The closer the temperature gets to curie the lower you have to adjust your frequency to compensate for the loss in magnetic strength of the metal. At curie you have to make a big jump in frequency to continue to induce energy into the metal.

Bottom line... you are tilting at windmills!

Advice on designing and building an induction furnace for small-scale iron melting

Diamond

Diamond

Stainless

Diamond

Diamond

Diamond

Diamond

Diamond

Diamond

Diamond

Titanium

Diamond

Diamond

Diamond

Hot Rolled

Diamond

Diamond

Diamond

Hot Rolled

Plastic

Similar threads