Hello!

Im thinking of building a induction furnace of about 2,5 kW! The electronics seem to be very simple to build! With a vacuum pump and a good chamber it should be possible to cheaply and cleanly melt just about every kind of metal or to make perfectly homogenuos alloys :D

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

How would you solve the pouring in the closed chamber? I was thinking of using a solenoid to "pull the plug" in the bottom of the crucible. It would take much space to tilt the crucible and harder to get the pour in the right place :)

What kind of refractory should i use to be able to melt higher temperature metals? How high vacuum would i need? Will i need flux? What material should i use for crucible?

/Jay

I'm sure there is a reason why commercial furnaces don't use "pull the plug" on their crucibles....could be differences in thermal expansion.

I was thinking about a cone as the plug with a spring who push it down in the hole with the same draft angle.. Something like this! The spring should keep it tight :)

/Jay

Just out of curiosity, what would you make the 'plug and rod' out of to keep it from melting?

EDIT: Nevermind, read better and answered my own question. :idea:

Well the same material as the crucible i guess... Perhaps graphite or some kind of ceramic :) I dont want it to react with the metal! Would graphite change the properties?

/Jay

Well the same material as the crucible i guess... Perhaps graphite or some kind of ceramic :) I dont want it to react with the metal! Would graphite change the properties?

/Jay

So you would be planning on having someone custom make the crucible?

I looked at the cost of graphite crucibles once, they ain't cheap!

I guess I don't see why the hold/plug custom crucible would even be needed? Judging from the page you linked - it would seem that you would just make your coil large enough to put the crucible inside and use hooks/tongs to get the crucible out of the center off the coil (after you shut it off, I would assume).

Now that I think about it, you should be able to embed the coil in the wall of a self-made furnace... :idea:

I don't know, I just like to keep things ASAP - As Simple As Possible. :o

I'm planning on doing all melting and casting under vacuum in a closed chamber to avoid oxidation, so i have to be able to do the casting by "remote" :D

A ceramic crucible (and plug) should be very easy to make... Just build a form for it and then burn it in a kiln.

Simple solutions is working... But takes away the fun in building and designing ;) And is often not optimal :) This is what i'm thinking of... I guess it could be a problem with sealing. :P I guess it would be nice to be able to measure temperature! Can you use a IR thermometer trough glass?

There are lots of simple ways to protect the melt from oxidization other than a vacuum. As far as bottom tapping goes many have tried different methods of plugging and sealing without much success. Even the smallest leak that can't be stopped can ruin you furnace. Depending on what you are melting home made crucibles are fraught with danger, a bit like juggling eggs.
If you want to see a very successful home made induction furnace check out –http://home.iprimus.com.au/cmckeown/induction_furnace.htm
This furnace primarily melts cast iron. The results are an extremely high grade of metal.

Yep i have seen that one :) Very cool! Well if i build the electronics its easy to try different coils and configurations... I guess i would start with something simpler like that! Just to get started anyway :D

I've been aprofessional metalcaster for close to 25 years witha degree in metallurgy. Oxidation is least of your concerns when processing steel. Depending upon the composition, nitrogen is the biggest culprit.
You also use different molding materials for steel. Petrobond won't do. Kiln dried ceramic makes great mold and casting, all of that with it's own process restraints and problems.
You don't want to use graphite crucibles with steel, the molten metal dissolves the graphite in the crucible and you end up with a stainer.
You have to have steel in the 3050 degree F range to pour it and it looses heat really fast!!!! All the prep work and you end up with a mis run casting.
Commercial vacuum induction furnaces, especially with aerospace castigns and alloys, have the preheated ceramic mold on top of the induction furnace, at the proper time teh whole unit is inverted and the mold filled.

Were there any schematics posted somewhere for the one shown in the link?
It's OK I did'nt look hard enough (chair)
Al.

Have there been any development in this field lateley??
I am in the process of gathering information on a furnace that could be used for melting Al and possibly for making some Al/Mg alloys.
Ideally, I want to use electricity as the energy source, and have already made a ill-conceived attempt at making a resistive heater, but I can't find a suitable castable refractory, so I resorted to fireplace-brics and leca, loosely fitted in a large steel bowl, and using nicrothal wires coiled up as springs, and fitted in slots carved in the bricks.
The result was a big spike in the electric bill, and alot of red-hot leca and fire bricks...
The Al I managed to melt, ate it's way through the capped steel pipe crucible, and generally made a big mess..

I'm pondering a new design, using a red clay crusible and furnace liner/heat wire holder and using mineral whool for insulation.

While browsing the net, I found out about induction heating, and started to seriously consider this as an alternative route.

Al is a very conductive material and also it's not magnetic, and I would guess that it would be pretty hard to develop enough heat to melt it.

How would clay stand up to the thermal fluctuations that one could expect from a furnace like this?
Just hours ago, I exploded a small square of hardened glass by ignoring the stresses caused by uneven heating.
I made a small coil from my heating wire, and placed it on the glass, and placed a piece of Al inside...
The al melted, and solidified into a blob on the glass...
I removed the glob, and put the glass aside to cool down.
After maybe 5 minutes, it exploded, and sprayed small glass cubes all over the place...
Would clay behave in a similar manner, or is it more flexible?

Hello All,

I am new to posting to this forum. However, I have been a scientific lampworker for nearly ten years now. If you wish to make an electric furnace for the purpose of melting metals then you should contact these companies (they are both owned by Amaco): Amaco and the Brickyard.
This link is for the electric elements. The ones you all probably will want to use are the metal enameling replacement elements. I have one of these kilns, produced in 1948, and it reaches temperatures in excess of 2500 degrees C on 120VAC! The kiln has two elements of 500 watts each, connected in series for the high temp setting and in parallel for low (or the other way round, I cannot remember). A pyrometer and thermocouple show absolute temperature.
http://www.amaco.com/jsps/grouphome1.jsp?catID=30&GROUP_ID=10&CATEGORY_ID=30&TITLE_NAME=Elements%20-%20Amaco%20Kilns
The elements can simply be stapled to kiln brick, which is fairly cheep when not precut, in a zig-zag or cyclic fashion.

For kiln brick for any style of refractory, contact the Brickyard http://www.brickyardceramics.com/

For the making of high temperature crucibles as described to me by a dear friend:

1. Choose a suitable clay (some experimentation may be necessary for your exact application, e.g., type of metal).
2. Form some of the clay into thin pieces (about 4mm thick), dry, and bisk fire in a kiln, or your furnace, to a cone of less than that of the final firing, see later steps. Large underground kilns have been used for centuries for this purpose, heated by wood or charcoal.
3. Break the bisk into small particles, fine grind with a hammer, or some other means.
4. Combine up to about 50% bisk back into the original clay, this should be as homogeneous as possible.
5. Form into a crucible shape, either with a wheel, or core-formed. I would not recommend hand-built techniques, as it should be a single piece of material, not joined parts.
6. Dry, and fire the second time at a higher temperature.

Notes: After the firing process is heated to the desired temperature, always cool the stoneware inside whatever it was fired in, and as slowly as possible till completely cool. The addition of bisk clays provides for more strength and higher temperature capabilities compared to that without.

The crucibles can be made into any size and thickness. As small comercially available crucibles are relatively cheap, however, the larger ones are a bit expensive.

Hope this helps...

Just throwing out an idea here.....

An Argon 'flood' would probably be alot easier than vacuum. Argon is heavier than air, and tends to pool in low places (this is a serious safety concern in welding, use extreme caution when TIG welding in a confined space :drowning: ).

All equipment and refills are readily available at welding supply places. You could rig a system to keep a slow steady stream flowing into the furnace and simply displace everything but the inert Argon. Figuring out a way to pour might be tricky, but the flipping method mention earlier sounds like a winner. :cheers:

Um, since AL is non-ferrous, I don't think you can use inductive heating on it, since inductive heating relies on eddy currents from the magnetic field..since AL is not magnetic, it won't work - only for iron and steel...

-niko

Well steel and iron loose the magnetic properties after the curie temperature.. about 800 if i remember correctly... You can heat all conductive materials with induction! :)

Yes you are right, I was thinking about that too - when iron and steel get beyond a certain heat they lose their magnetic properties..so thanks you have answered it..

It works partially on eddy currents... but also on skin effect (http://en.wikipedia.org/wiki/Skin_effect)!

I have done some more research! I belive you could just bypass the output rectifier on a welding inverter and use it for induction heating! I don't know if the switching frequency is optimal for melting metal! But it should work!?

Actually it works on any conductive material. The steel and other magnetic materials heat *faster* initially due to being constantly remagnatised, but loose that effect over ~750C. somewhere around there.

Your basically relying on a huge field of electron charge, concentrated in the center of the coil, and running down the inside (or up)..> It sort of looks like a stretched out donut, with electrons running down the middle and around, up the outside.

Thats all I can remember from physics :p

this subject is interesting - from googling, here's a description of power supply and frequency requirements - what's the frequency of the welder?

http://www.inductionatmospheres.com/powersupplies.html

would the basis of the circuit be a timer switching on and off some big power transistors? kind of like a pwm but more heavy duty?

You would want much lower frequencies for meliting metal 50 Hz to 20 kHz or so! :) Higher is good for hardening and heat treatment as i understand it! I don't know what switching frequency a welder use.. But the principle is:

AC > Rectifier > H-Bridge(PWM) > Transformer > Work coil

If you have 10A DC supply... 50% Dutycycle... 10:1 Output transformer and a 15 turn work coil you should get ~700A, including losses, running trough the workpiece :)

Built a H-Bridge for my induction heater project today :) It's built for 350VDC and up to 30A (To bad i only got 16A). The heatsink i will mount it to is 42cm long (16,5")...

Have begun designing the driver board with a microcontroller, with lcd and rotary encoder for settings! :) Cant wait to try it out :D

There are two quite different mechanisms at work in induction heating. The first is hysteresis heating of any magnetic material. The higher the operating frequency, the greater the heat generated. But this only works for ferrous metals below the curie point. For non magnetic metals (or above the curie temperature for magnetic metals), the heating effect is created entirely through circulating eddy currents.

In effect, the metal being heated becomes a single shorted turn of a high frequency transformer. The difficult part is coupling enough energy into the work to be useful. In order for this to work reasonably well, the induction coil absolutely must be made part of a resonant tuned circuit.

The idea is that a resonant build up of energy vastly increases the circulating current in the work coil, way beyond what the power source would be able to create all by itself. Many hundreds or even thousands of amps are required to pass through the induction work coil in order to couple enough energy into the work.

So there is a lot more to this than just coupling up a few turns of copper to a welder and hoping for the best.

The first requirement is to build a tuned circuit that can carry all this power, possibly ten to twenty times the basic input power of the induction heater. Several hundred thousand VAs (reactive watts) of circulating energy may be involved, and a very special and expensive water cooled tuning capacitor will be required that can carry several hundreds of continuous amps without burning up.

So the starting point of your home made induction heater will be getting your hands on one of these special induction heater tuning capacitors to tune your work coil to the desired operating frequency.

The next problem is driving this tuned circuit in such a way that the driver will not self destruct if either the driver or tuned circuit get out of step with each other. This is a lot more difficult to do than many people realise.

An internet search will turn up many people that have tried to do this and failed. It just ain't that easy. Heating a few pennies to red heat, or the end of your pliers with a couple of hundred watts is dead easy. Feeding several (or many) kilowatts of power reliably into several pounds of metal is going to be a far greater challenge.

My interest in induction heating has just now been aroused, and I am prepared to have a go at this myself. It will take me a while to track down and obtain some of the rather difficult to find (and expensive) special parts that will be required to try out some ideas of my own.

You make it sound like you could get over 100% efficiency by using a tank circuit, all know thats impossible! What efficiency do you expect (not including losses from heat radiation of the workpiece)?

Ah yes indeed, over 100% efficiency is obviously not possible, very true, but perhaps I did not explain this very well.

Think about how a pendulum or a flywheel stores energy. Keep giving it a little kick at the right spot, and the internal stored energy can rapidly build up to fearsome levels. Resonance is a very powerful means of storing or building up stored energy, but obviously you can never draw more time averaged energy out of it than you put into it.

The problem with induction heating is to get the electrical current to flow in the work. It relies on a magnetic field around the work coil, and magnetic fields do not travel very well through air.

So what happens is a certain energy is fed into the tuned circuit, and that energy quickly builds up, and keeps building up, until the power draw from the circuit exactly equals the input power.

So you might have (say) a thousand watts going in, fifty thousand circulating watts, and a thousand watts going into the work. This means that the coupling into the work from the coil is not very strong, only 2% of the circulating energy actually makes it from the coil into the work.

Without a resonant tuned tank circuit, your thousand original potential watts could only transfer maybe twenty watts into the work. Your thousand watt induction heater would remain essentially unloaded, because of the very poor output energy transfer.

But by tuning the circuit to resonance, a massive potential energy reservoir becomes available, and that stored energy becomes sufficiently high to be able to leak away into the work at a sufficiently high rate to be useful.

Seems that you just took this thread to the next level with your analysis and explanation.

Thanks.....

In order to build a workable induction heating tank circuit, you will first need to get your hands on a suitably rated tank tuning capacitor.

I have been investigating the tank capacitor procurement problem for quite a while, and none of the big capacitor manufacturers would even bother to reply to my e-mails. But finally *bingo* I hit the jackpot. I discovered a small company in Israel called Celem that makes nothing else but induction heating tank tuning capacitors. Because it is their only product, they will never tell a small customer to just piss off, like the sales reps of the big companies do.

http://www.celem.com/

Take a look at the ratings of these things, they are truly amazing. An example might be something like 400 volts rms, 900 amps rms, and up to 360,000 maximum circulating watts (360 Kvar) if pushed to the ratings limit. They are designed to operate most efficiently around specific frequencies, so just choose your capacitor, and wind your tank coil to tune it to the desired operating frequency, and that solves the tuned tank problem rather neatly.

There is far more to building a working induction heater, but it does overcome one very huge problem, and it will get you started.

Nice... Did you get any prices?

Yes, I have just now ordered and paid for one of these capacitors, and it should hopefully arrive some time next week. What I ordered was:

C200T 21uF, 400v rms, 900 amps rms, 250 Kvar, best operating range 12-25 Khz

Capacitor $121.50 (US dollars)
Insurence fifty cents
Packing box $2.00
Freight UPS, Israel to Australia $44.00

Total $168.00 US dollars.

I hope the moderators here don't mind this, I am not touting for this company, I have no connection with them apart from being a one time only customer.

Just trying to help out some fellow CNC zone induction heating freaks.

Thats not bad! If i would go for a transformer instead it would cost at least that! :) And get perhaps at best get 60% efficiency (i have seen indications of that number). How big heater will you build?

That is a decision I have yet to reach.

It will depend on what sort of power transformer I can get. Once I have a suitable transformer, whatever voltage and current it is capable of will pretty well fix my final power level. But at least 5Kw to begin with. My three phase power here is limited to 25 Kw, so it can be no higher than that.

At this stage I really have nothing, finding that tank capacitor is the big breakthrough I was hoping for. Without that capacitor this project would not even have begun. The next major obstacle is sourcing a suitable power transformer.

When I know how many rectified dc volts and amps will be available, then I can start thinking about the rest of the electronics.

Okay, a bit better than here! We only have 16A on the 3-phase, so max is about 10 kW... 5 kW seems like a resonable start, then you should be able to melt about 5 kg of steel in one hour or so.

I found these numbers on a forum for a induction heating company:
Steel: 3 lbs per kWh for melts over 100 lbs.
2 lbs per kWh for melts under 100 lbs.
Brass: 7 to 10 lbs per kWh
Aluminium: 4 to 5 lbs per kWh

If i get everything working good i might get some BIG 1200V 600A IGBTs and a old 200 kVA military generator ;)

Yes, I agree, anything less than 5KW is really not a practical usable induction furnace, and 10Kw would be far better. Very many people have reached the several hundreds of watts level, and then given up because of repeated problems with transistors blowing up. If I can get 5KW working reliably as a very first attempt, I will be more than happy. Nobody else that I can find on the internet seems to have yet achieved anything like this in a completely from scratch home built project. But I am game to try.

I am convinced that using a well thought out and methodical design approach, and correctly specified components, it should be possible.

The way I see all this, the first design problem to overcome is deciding on the operating frequency and building a suitably rated water cooled tuned tank circuit. I have finally decided on 20 Khz, but that will vary up and down somewhat in operation, depending on the metal load and type of metal (ferrous or non ferrous).

The second design problem to overcome is getting suitable dc power at a convenient operating voltage to run some IGBTs. I have decided to use 600 volt rated devices, simply because there is a much wider choice of high performance IGBTs and diodes available at a reasonable cost, at that rated voltage. Later I may go up to a higher operating voltage to increase power, but to begin with for this initial prototype, I plan to start off with two dc supply rails of +220v and -220v with respect to mains neutral.

I was outbid yesterday on e-bay for an ideal three phase transformer, so will now go back to using something less wonderful that I already have here. This has three 240v primaries connected in star, and nine individual isolated 28 volt secondaries, each rated at 15 amps rms. The transformer rating is about 3.8 kva, and I believe it was originally salvaged out of some sort of battery charger.

What I plan to do is place the three secondaries on each transformer leg in series giving 28 x 3 = 84 volts and use this as a three phase bucking transformer. I have a 240v phase to neutral incoming supply, subtracting 84 volts from that, will give me 156 volts phase to neutral, at 15 amps per phase.

A six diode full wave bridge will then produce 156 x 1.414 = 220v peak output voltage with a very low 300Hz ripple. Only a comparatively small filter capacitance is required. That should reduce any inrush problems and give me a nice solid low stress reasonably well regulated dc power supply, 440v total at 20 amps or around 8.8kw continuously rated. Those values of voltage and current should be well within the capability of some fairly low cost high performance individual IGBTs and fast diodes.

Strictly speaking I do not really need any filter capacitance at all on the output of my three phase rectifier bridge. The ripple voltage is quite low, but that would leave my IGBTs susceptible to mains voltage spikes. Some bulk filter capacitance, and some transient supression across the transformer windings should help eliminate any really destructive voltage spikes or surges originating from the mains supply.

I can get much more dc output power than rated transformer va's because a bucking transformer does not have to supply the total power, only the proportion of the output power required to reduce the voltage.

This has yet to be assembled and tested, that is the next step.


.

There are two things that immediately strikes me:
1: The tank cap might be a bit overkill. Remember the current flowing through the workpiece is the current flowing through the induction coil multiplied by coil windings so a 10 turn coil having 20 amps going through it would induce 200 amps into the workpiece.
2: Why not go straight for the mains supply, bypassing the transformer entirely? Sure, there's a major safety issue here, but as long as you a) know about it and b) got some proper fuses in place, it would save you the transformer entirely as long as you're looking for a "proof of consept" design and not a "production" unit.

Isn't the bucking configuration already hot with respect to ground? If so, there is little point in using the transformer.

Ken

Here is a link with some good info on induction furnaces

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

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.

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/tmoranwms/Elec_IndHeat1.html

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.

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.

Here are some pics:

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

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.

Mr Awsome Mawson, GOLLY !

And you class yourself as a total amateur, hehehe.

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.

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:

http://i144.photobucket.com/albums/r166/Warpspeed_photos/2007-04-10-1049-52_edited.jpg
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.

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.

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.

...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.

A valid concern. I have wondered if you simply surrounded the induction furnace with a steel mesh screen whether that would shield it. A kind of inside-out Faraday cage?

EDIT

Its not a Faraday cage is it? That is for electric and magnetic fields? Or maybe it is...electromagnetic theory is not my strong point. (Are not my strong points????...obviously I have problems with grammar also.)

I had planned to sit the whole prototype on a trolley and wheel it well away from everything else before running it at anything like full power. But that is all probably still a very long way off into the future. At this stage I just wish to test independent parts of the system in isolation. Never having even seen a commercial induction heater before, this is all rather an adventure.

My calculations of electrical losses suggest I might need about roughly 300 watts of cooling capacity for the electronic heat sinks, and around 1,500 watts cooling capacity for the tank system. Can some of you old hands tell me if that sounds realistic ?

My most recent cooling system configuration should do that with a measured 10C water temperature rise. (water into cooler 21C, water out of cooler 11C (above ambient) with a 1600 watt immersion heating element as a test load). The water/air heat exchanger is reasonably large with about 800 CFM of measured airflow going through it.

Would that sort of cooling capacity be reasonably comparable to a commercial unit in the 10KW to 15Kw class ? I have absolutely no idea about this.

I.....Would that sort of cooling capacity be reasonably comparable to a commercial unit in the 10KW to 15Kw class ? I have absolutely no idea about this.

I have no idea either. The system I ran simply had tap water coming in and then running to drain. I don't even know what the volume flow was.

When I think back on my experience it was really Keystone Cops. After I gave notice of quitting they hired a couple of other guys to replace me. On their first run on the furnace they did something wrong, panicked and dumped 500 lbs of molten copper onto the mezzanine floor. In the process encasing a couple of rubber propane gas lines in molten copper. I had to walk across the red hot copper to turn off the gas valves meanwhile screaming for fire extinguishers to put out my shoes that had caught fire. I guess I was lucky to survive that job.

That is exactly why I prefer to run a fully closed water cooling system. If the incoming power fails, the source of heat goes away as well.

It sounds like you had a really "interesting" job there Geof. Smoking shoes just add to the excitement. I am very surprised that system did not have suitable automatic safety interlocks and shutdown features. If something can go wrong, it usually does, sooner or later.

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.

To confirm the stray magnetic field issue. When I rebuilt one of my furnace bodies, I marginally re-routed the copper feed pipes to make things more convenient. ERROR! I ended up with a jubilee clip on one of the reinforced nylon water pipes effectively in the centre of a 'half turn' of the current carrying pipes. It took me a while to realise why this clip kept overheating and melting the nylon. Ended up having to source some nylon clips to substitute for the jubilee clip !

Here is a picture before I changed the clips

Warpspeed, Your design is looking good.
Have you considered how you will deal with a shutdown?
There is a lot of stored energy in a system like this and the energy will find a way to dissipate. (flame2)

As far as stray magnetic fields, you should keep all steel components and electronics a couple of feet away from the coil. Any fasteners on or near the coil should be made of non-ferrous materials. Brass or a high grade Stainless work well.
There is a discussion on myinduction.com concerning distance magnetic fields will travel. http://myinduction.com/forums/view_posts.php/forum-32/topic-40

What size cap are you looking for? Around 440V @ 20Khz I see, but I need one more datapoint. Either amps, uf, VA - of capacitor. Or if you know the approximate uhy and Q of your coil/load.

That is exactly why I prefer to run a fully closed water cooling system. If the incoming power fails, the source of heat goes away as well.

It sounds like you had a really "interesting" job there Geof. Smoking shoes just add to the excitement. I am very surprised that system did not have suitable automatic safety interlocks and shutdown features. If something can go wrong, it usually does, sooner or later.

As Shawn D points out you have stored energy so just turning something off may not be the way to go.

My excited state when my system shutdown was motivated by the fact that I had molten copper separated from a coil of water filled copper pipe by about 1-1/2 inches of refractory material. The auto shutdown closed the incoming water valve and I had no idea if the outlet was now also closed. Water expands when heated and the picture that went through my mind was a crack occuring in the coil squirting water into the bottom of the crucible. Whee!!!! My own mini-Krakatoa.

I never looked into it further because I moved off that furnace but I have often thought there should be a Panic Mode option which would blow compressed air into the coil to displace all the water if cooling circulation failed with a melt in place. You might have to rebuild the furnace but at least you would not need to rebuild the where it was situated.

As Shawn D points out you have stored energy so just turning something off may not be the way to go.

My excited state when my system shutdown was motivated by the fact that I had molten copper separated from a coil of water filled copper pipe by about 1-1/2 inches of refractory material. The auto shutdown closed the incoming water valve and I had no idea if the outlet was now also closed. Water expands when heated and the picture that went through my mind was a crack occuring in the coil squirting water into the bottom of the crucible. Whee!!!! My own mini-Krakatoa.

I never looked into it further because I moved off that furnace but I have often thought there should be a Panic Mode option which would blow compressed air into the coil to displace all the water if cooling circulation failed with a melt in place. You might have to rebuild the furnace but at least you would not need to rebuild the where it was situated.

Many professional furnace bodies have conductive wires embedded in the 'drypack' that the crucibles are supported by, and have electronics to detect the first signs of electrical leakage between these and earth and shut down. Mine doesn't have that system, but the water pressure and temperature are monitored and the system shuts down fast if they go beyond safe limits. The hypothisis is that a coil failure will start small and grow rather than imediately pour forth all its fluid. The conductivity of the water is also an issue that you must consider.

...but the water pressure and temperature are monitored and the system shuts down fast if they go beyond safe limits. The hypothisis is that a coil failure will start small and grow rather than imediately pour forth all its fluid. The conductivity of the water is also an issue that you must consider.

This does not address what concerned me. You can shut down the energy input if something starts to go wrong but shutting down the cooling when you have a hot system may not be a good idea. Especially if shutting the cooling down means closing valves both on the in and out lines. That is what would be undesirable; a liquid system closed at both ends and getting hotter.

Many professional furnace bodies have conductive wires embedded in the 'drypack' that the crucibles are supported by, and have electronics to detect the first signs of electrical leakage between these and earth and shut down.
Not that I am trying to sell the kit, but my company has a Ground/Short kit, that connects to the heating coil and monitors the resistance path and voltage potential from coil to earth. The kit also has two contact outputs, one basically to be used as a warning, and the other to cause an emergency shutdown. I think other manufacturers have equivalent detection circuits.

When I stated that Warpspeed should consider a shutdown mode, I was actually speaking of the energy in the power supply, not necessarily in the coil. I was speaking of energy in the choke that would be in the Buck converter and any capacitors in the system. Many parallel tuned power supplies create a short (crowbar) across the tank by firing all of the inverter switching components at once. But, before you create the crowbar, you would need to disconnect from the utility. This can be done with the Buck transistor or using SCR's for the bridge rectifier instead of diodes.

This does not address what concerned me. You can shut down the energy input if something starts to go wrong but shutting down the cooling when you have a hot system may not be a good idea. Especially if shutting the cooling down means closing valves both on the in and out lines. That is what would be undesirable; a liquid system closed at both ends and getting hotter.

I don't know why anyone would want to shut down the cooling medium (water) in the system. Unless, the cooling is close enough to the heated load, in your case molten copper, and there is a concern that the cooling would solidify the copper.

I don't know why anyone would want to shut down the cooling medium (water) in the system. Unless, the cooling is close enough to the heated load, in your case molten copper, and there is a concern that the cooling would solidify the copper.


How about explosive release of steam as a good reason to shut down the cooling water fast in the case of a leak getting near the hot zone ?

How about explosive release of steam as a good reason to shut down the cooling water fast in the case of a leak getting near the hot zone ?

This sounds like an Emergency situation. In this situation, I would shutdown all power in the induction system AND close a solenoid on the supply and return side. I would still allow water to circulate in the induction power supply, just not the coil.

But, I thought that the previous comments were talking about shutting the water off to the coil in a non-emergency shutdown.

So, how do you monitor for a leak near the hot zone?

This sounds like an Emergency situation. In this situation, I would shutdown all power in the induction system AND close a solenoid on the supply and return side. I would still allow water to circulate in the induction power supply, just not the coil.

But, I thought that the previous comments were talking about shutting the water off to the coil in a non-emergency shutdown.

So, how do you monitor for a leak near the hot zone?


Electrical leakage to earth. My coolant is at something like 40 psi so a small leak goes a long way !

I keep my coolant circulating for quite a long time after shutting down the driver electronics as if I stopped the pump too soon the now static water in the coils would be in grave danger of boiling

I don't know why anyone would want to shut down the cooling medium (water) in the system. ....

I concur with this opinion. However, read all my posts; I was faced with a situation were the cooling water was shut off externally...it was not my choice. I wasn't the slightest bit worried about cooling the molten copper I was worried about the water getting up to that temperature.

This is why I postulated my Panic Mode a few posts up to get water out of the hot zone when things have gone seriously wrong and you don't want them to get wronger.

Thanks guys, this is exactly the sort of discussion I wish to have, it is bringing up possibilities that need to be thought through very carefully.

O/k, the bare tank circuit will be designed to resonate at 20Khz (with no metal loaded into the coil). The tank capacitor value will be 21uF, and that has a reactance of around 0.4 ohms. The design tank amplitude will be 300v rms, so the circulating current should end up in the region of 750 amps. Tank power works out to 225 KVAR. The capacitor maximum ratings are 400v rms, 900 amps and 250 KVAR, so it will be running just nicely within it's ratings. The recommended optimum frequency range for this particular capacitor is 12 to 25 Khz so that is o/k as well. Fully loaded the tank Q will be 225KVA/15Kw = 15, so I would expect it should couple fairly well into the load.

Both startup and shutdown are very critical times for any high stored energy system, and this has all been thought through very carefully indeed. The fact that the tank circuit and H bridge are current driven means that startup cannot overload anything. The H bridge will work at continuous full power into a dead short or into a very highly reactive load, either leading or lagging without any problem at all. The peak voltage across both tank and H bridge will be clamped to operate between the 440v dc suppy rails.

The H bridge will have a commutation diode in series with each IGBT, so the IGBTs will automatically block in the reverse direction, just as SCRs do. In fact, with this particular topology, all four IGBTs can be turned on simultaneously, or in any conduction order with respect to the stored energy in the tank, without any danger at all. That is the greatest advantage of the current driven converter topology, it is very immune from electrical mishap if the tank and drive system become out of step with each other.

I have not yet given much thought to thermal shutdown, but agree that the cooling system should continue to run after tank power is removed. As I have never seen a real induction heater, I am still learning. But those types of details can be sorted out later on once the beast is running. Electrical robustness, and possible electrical failure modes are a far more important design considerations at this stage. Thermal design at this point, is just to ensure that there is enough total cooling capacity.

While I still know zip about induction heating, my previous power electronics design experience is fairly extensive.

Inductive coupling both into and from any circuit relies on loop area. That means that any circulating currents will radiate in proportion to the size of the current loop as well as the strength of the current. So the trick is to keep all wiring as short and direct as possible, and keep wires that flow current in opposite directions bunched close together. That also applies to the drive and control system layout to make it less susceptible to inductive pickup.

The physical layout of the parts is just as important as the circuit diagram. I have had plenty of design experience with similar sorts of thing, so feel confident I can succeed with this. I have been designing and building switching power supplies as a professional design engineer for years, and have had some radio transmitter experience both with high power commercial broadcast transmitters, as well as being a radio amateur. So although induction heating is a completely new field to me, I do have a lot of quite relevant experience.

One other thought is the problem of the cooling water being electrically conductive. I have absolutely no experience with this, but my first attempt will be to use automotive engine coolant. This has a corrosion inhibitor and PH buffer, but the electrical characteristics of this brew are completely unknown.

If the whole thing looks like being a problem, I can always use oil as the cooling medium, and increase the flow with a more powerful pump. But it seems everyone else just uses straight tap water, so the problem cannot be too difficult. We shall see....

One other thought is the problem of the cooling water being electrically conductive. I have absolutely no experience with this, but my first attempt will be to use automotive engine coolant. This has a corrosion inhibitor and PH buffer, but the electrical characteristics of this brew are completely unknown.

If the whole thing looks like being a problem, I can always use oil as the cooling medium, and increase the flow with a more powerful pump. But it seems everyone else just uses straight tap water, so the problem cannot be too difficult. We shall see....

Even if you use automotive coolant it will still have some water, yes? Using anything other than water will greatly reduce your heat transfer capacity; nothing has the same specific heat as water. Using oil could introduce its own hazards; water is non-combustible oil is.

I have a GE capacitor catalog handy, so you need GE part # 19L1049WH* The star position relates to mounting style.
This is a 1095KVAR 600V 20Khz 1659Amp with 8 studs at 2.5uf each (20uf total).

Use distilled water if you are concerned about the conductivity of the water. Don't use oil, it always has a flash point! (flame2)

Some really interesting ideas there guys, please keep them coming. My main concern with water is the constant high dc voltage potential between various water cooled parts. This is likely going to cause very rapid erosion through ion migration (as in metal plating). I am hoping that distilled water mixed 50/50 with automotive engine coolant will limit that process.

Reduced specific heat is not really a problem if the cooling system is made large enough. What I have here at the moment can dissipate 1600 watts with an average measured temperature rise of only 16C above ambient. Even with something only half as good as water, the temperature rise would then be 32C above ambient at that same power level. That could be offset somewhat with increased coolant flow.

Almost any coolant can be dangerous in some respect. I would certainly prefer water, or at leased an aqueous based solution of something or other. But oil is used to cool all sorts of things and the total volume of coolant in this system is quite small.

I have a GE capacitor catalog handy, so you need GE part # 19L1049WH* The star position relates to mounting style.
This is a 1095KVAR 600V 20Khz 1659Amp with 8 studs at 2.5uf each (20uf total).


That would be an interesting alternative, but probably larger than I really need. This is what I am trying to get my hands on at the moment:
http://www.celem.com/datasheets/C200T.pdf
x
http://www.celem.com/images/photos/C200T.jpg

One other thought is the problem of the cooling water being electrically conductive. I have absolutely no experience with this, but my first attempt will be to use automotive engine coolant. This has a corrosion inhibitor and PH buffer, but the electrical characteristics of this brew are completely unknown.

If the whole thing looks like being a problem, I can always use oil as the cooling medium, and increase the flow with a more powerful pump. But it seems everyone else just uses straight tap water, so the problem cannot be too difficult. We shall see....

I have been advised NEVER to use normal car antifreeze coolant but to use pure ethylene glycol as antifreeze as a 20% solution in water. Apparently there are silicates in car antifreeze that are in solution at the high temperatures of a car system, but settle out of solution at the much lower (under 35 deg C in my case) temperatures of the furnace coolant. These silicates coat the surfaces of the heat exchangers and reduce their efficiency very markedly

I am looking at some automotive coolant that I have here in front of me right now, and it says "mono ethylene glycol" at 1080 grams per litre. This is what I intended to try first.

Looking up the specific heat of various types of oil:

http://www.engineeringtoolbox.com/specific-heat-fluids-d_151.html

Most types of oil seem to have about roughly half the specific heat of water. I can probably live with that if I use a more powerful circulating pump. At the moment I am pushing only about roughly 2.6 litres per minute through about sixty feet of 3/8 inch pipe with a Holley electric fuel pump. It is really only a trickle, but it does match the mass airflow I have rather well.

Oil will definitely solve both the corrosion and electrical conductivity problem. If I boost my flow up to around five litres a minute, it should work pretty well. I am not too concerned about the fire risk. If my coolant ever gets hot enough to burn, it will be the very least of my problems.

I took my advice from a firm here in the UK that services induction furnaces, and they said it was a regular problem that they have to deal with when cooling systems have been contaminated with car antifreeze. Although the main constituant of vehicle antifreeze is indeed mono ethyl glycol, the additions that are made are the issue.

See paragraph 4 of this url for confirmation:

http://www.ethyleneglycol.co.uk/antifreezeguide.html

Thank you for that information Andrew, very helpful indeed.

What concerns me most about all this, is what happens if you have one piece of immersed copper tubing at +220 volts dc, and another piece of immersed copper at -220 volts dc, with some electrical leakage current flowing through the water based coolant. I am sure that something really bad is going to happen.

Oil being a pretty good electrical insulator, as well as being fairly chemically inert, may well be the best solution to all this. I am thinking here of very high voltage oil filled substation transformers, and similar applications.

I would really like the main liquid cooling system to cool all my IGBT and diode heatsinks, as well as the tuned tank system. There will be some fairly drastic voltage differentials between different parts of the cooling system, and that aspect has me a bit worried.

The manual for my CFEI furnace driver states that the coolant (Caracteristiques de l'eau de refroidissement !!!) should have a resistivity of a higher value than 2500 ohm cms. The coolant is pumped not only round the furnace tank coil, but also through the tank capacitor, and the main thyristor copper heatsinks. The busbars (which are massive) have copper pipes brazed to them, again with water cooling. Each item to be cooled is connected via a neoprene rubber 1/2" hose and it would seem that the hose lengths have been selected to ensure a long path to reduce conduction.

Here is a picture showing the 415v three phase bus bars (vertical and 95 sq mm area) coming down to the thyristor rectifiers to produce the DC bus (horizontal and having water cooling pipes brazed to the rear)

That looks to be pretty much exactly how I planned to cool my system. Solid copper heatsinks with a length of copper pipe silver soldered onto it. I have already tested this, and it worked surprisingly well. I fed a hundred watts from a bolted on metal clad resistor into a 2" by 2" slab of copper 1/4 inch thick, with a quarter inch copper pipe six inches total length silver soldered to it. It would have only just been sufficient. My final design will probably be more like 3" by 3" by quarter inch copper slab with about nine inches of 3/8 pipe. There will be six of these heatsinks. Two for the buck regulator, and four for the bridge.

Something similar but a bit smaller bolted to each side of the tank capacitor should conduction cool that. The expected power loss from the capacitor (according to the capacitor manufacturer), may be around fifty watts. The tank coil is where all the heat is. I am anticipating around a kilowatt to get rid of, but I really do not know for sure. Difficult to estimate the skin effect in the coil, and how much heat soak there will be from the work. But the tank coil will be the last heat load in the whole series flow system.

If I remember correctly, you are planning to run at 20Khz, which seems rather high to me. At the 3 khz that I run at the skin effect means that very little conduction happens below 1.25mm, so at 20khz it will be much shallower. What diameter coil are you aiming for?

My tank capacitors have integral cooling pipes immersed in the oil. When I was doing my low power experiments I was amazed what a variation in dialectric losses there was from cap to cap even from the same batch - I was using polypropolene motor start caps in parallel series combination to get the votage rating and capacity.

Here's some information:
Cooling medium conductivity requirements from the Bone Frontier Company manual = 200uS/cm.
I think you will find that distilled water will be the best solution. If you mix glycol the conductivity will be raised, therefore worsening the situation.
Electrolysis is a problem when dealing with the voltage potentials that are used in a typical induction system. Here is a discussion that took place on myinduction concerning electrolysis. Electrolysis discussion. (http://myinduction.com/forums/view_posts.php/forum-20/topic-20)
Here is also an article that Bone Frontier wrote concerning electrolysis and a simple device that we call a water target. Electrolysis article. (http://myinduction.com/articledetail.php/id-49)

Here is a link to some formulas when dealing with cooling flow requirements (http://myinduction.com/UserFiles/Cooling.pdf).

And finally, here is a link to a skin depth chart (http://myinduction.com/UserFiles/SkinDepth.pdf).

Sorry for all the links guys. :drowning: Just thought the information might be useful.
Like I mentioned before, the myinduction site is a bit clumsy. However, it does contain a lot of useful information.

Warpspeed, can you put some information up on what you are attempting to heat, or are you just building the induction system for the fun of it and to tinker with various heating applications?

Great stuff guys. I originally chose 20 Khz more from long established habit more than anything else. It is a good compromise design frequency for designing larger high powered switching power supplies. I can move the operating frequency easily enough later on if I choose to do so. But for better or worse, I have decided to begin at 20 Khz.

Way back about a month ago when all this began, I did calculate some suitable tank coil dimensions using Wheelers formula, and standard 3/8 copper pipe. The wall thickness of standard commercial pipe in this size is 0.9mm, and I estimated the straight ohmic dc copper power loss at around roughly 500 watts. More of a stab in the dark than anything else, using skin depth curves from a reference book, I figured only about half the outer pipe wall thickness would be seeing some serious action, probably doubling the actual real power loss to about one Kw. Accuracy was not the objective. I just wanted a ballpark figure with which to very roughly estimate the total system cooling requirement. Silver plating the coil should reduce the losses, but it hardly seems worthwhile.

So much has happened in this last month, with so many different things to think about, the original tank loss calculations for this have been lost and now forgotten. I came up with a final estimate, but cannot now remember all the exact details of how it was derived. Once I have my tank capacitor I will build a real tank circuit and go through the whole process again with a bit more care.

One question I have. Is it normal practice to surround the work with the total required tank inductance ? I can see that it may be fairly impractical at lower frequencies to do that, and theoretically it is not required. It should be possible to design a suitable work coil to closely couple into the job, and tune the tank with some additional uncoupled inductance. Is that actually done ? I have no idea being totally unfamiliar with all this.

Wow, thanks for that skin depth chart Shawn, I have been looking for something like that for YEARS. Skin depth curves for copper can be found in many of the reference books, but nowhere else have I been able to find information on skin depth in other metals.

Shawn, I did begin to build an iron melting natural gas fired furnace, but then became seriously interested in induction heating. What has really spurred me on, is that the more I search the internet, the more people I discover that have attempted to home design an induction heater, and not been able to get anything working reliably beyond a few hundred watts. A kilowatt seems to be something like breaking the sound barrier to most hobbyists. Looking closely at what others have done, and how they have gone about it, I can see all sorts of reasons for their failures. I may fail too, but I am going to have a very serious try. I will also share my success and failure with you guys, and maybe others can follow in my footsteps. Even if I do fail in this, it may advance the art just a little bit and increase the pool of available public knowledge. Any help and encouragement from you guys will be very gratefully accepted.

So yes, Shawn, I am just building this for the fun and the challenge of it.

Warpspeed,
Stray tuning inductance has been used in the industry. However, it does have a major disadvantage. Any additional stray inductance will effect the amount of energy that is put into the heated part. Stray inductance will have a voltage drop. Model your parallel tuned tank, only slip an inductor in series with your heating coil. Now, you will see that the voltage is divided among the two inductors. The larger the stray inductance, the less voltage that will be applied to the heating coil.
For this reason, most manufacturers minimize the stray inductance in a system. For instance, when making transmission leads (wire or rigid copper buss), the leads are kept adjacent to each other so as to minimize the inductance loop.

Normally, any tuning required for frequency is achieved through capacitance changes. That is why more tank capacitors have selectable studs.

Another option is to put taps on your heating coil.

Thank you for that great explanation, it is all now starting to become very much clearer.

I can understand the theory, it is just that never having actually seen any of this equipment, it leaves me wondering about a few things.

Another question. Every amateur home built induction heating circuit that I have so far discovered on the internet, none of them have attempted to control the tank amplitude in any way. There is always just a large power amplifier driving the tank circuit flat out all the time through a series inductor.

To my way of thinking, the tank amplitude must be controlled in such a way as to be maintained fairly constant. Sufficient additional drive power should then be fed into the tank to make up for the energy consumed by the heating load.

From the very little I have been able to discover about commercial heaters, the big low frequency SCR ones like Andrew's CEFI work in a rather similar way to the one I am attempting to build. They use a six SCR phase controlled rectifier bridge to generate a variable dc current, and that goes to four more SCRs in a resonant mode bridge chopper circuit to switch the regulated current at the tank operating frequency. As the SCRs are physically constructed in pairs, it looks like only three SCRs plus two SCRs, but there are actually six plus four..

Do all the commercial heaters use some sort of tank amplitude control system? as I would think that feature absolutely necessary, yet none of the amateur home built circuits have that. I rather suspect it is one of several rather fundamental reasons why some hobbyists are consistently blowing up parts at higher power levels.

In the commercial ones I've been able to see they all seem to work like mine. As you say they produce a DC rail by rectifying the three phase mains then chop an 'H' bridge of thyristors with the tank circuit as the bar of the H. As the tank is driven into resonance the tank voltage and current will increase enormously, which it is why it is important to actually achieve resonance to get reasonable efficiency.

With my CEFI, the microprocessor monitors tank input current and tank voltage, displaying it on a control panel. There is a 'power knob' that you can tweak to vary from 15KW up to 120 KW and it does this by altering the triggering phase angle of the phase controlled rectifier bridge across the mains. The micro has limits set for current and voltage - current to protect the semiconductors, and voltage to protect the tank capacitors. Infact it has a warning level, and a 'no you don't go there' level !

Here is a picture of the control panel after switch on showing 'ready'. The next action is to switch it to 'test frequency' mode and ping the tank. This lets the micro know what it's starting point is. Then switch back to heating mode and press 'go'. You will see on the bottom right hand corner there is a big black knob that sets the power level by phase shifting the rectifer trigger points.

Awemason has it right.

Basically there are two topologies when dealing with induction power supplies.

#1) Variable DC Buss, with the inverter running at (or very close to) the tank's natural resonant frequency. - This is the on that Awemason has described.
Basically in this mode, the power (or voltage applied to the tank) is controlled by the amount of DC Buss voltage achieved.

#2) Fixed DC Buss and "Swept Frequency Inverter". This style usually has diodes in the rectifier (unless the design engineer wanted to use SCR's for the shutdown mode). The inverter will start at a frequency far different that the natural tank frequency. Then as power is dialed up, the frequency of the inverter is moved closer to the tank frequency, thus producing power more of the time in the tank, and thereby increasing the average power produced in the load.

Some manufacturers have used a combination of the two topologies.

Some induction units are series tuned tank circuits, some are parallel.

Then there is always the famed "Crowbar" induction units. Which usually work until the day that any one component fails, then ALL components fail at once. This type of machine is easy to troubleshoot. Step 1, replace all components, Step 2 turn the machine back on. Step 3, if Step 2 failed repeat Step 1, then replace the cabinet, then proceed to Step 2. :D

That is pretty much how I thought it would work. I am planning to do pretty much the same thing, at least in overall concept. I preferred to use a high frequency buck regulator instead of using SCRs to control the main dc supply. It is far easier to control one IGBT than six SCRs !! But for really high power, the SCRS make a lot more sense. For relatively low power, the buck regulator is a lot simpler.

I will run six diodes in an ordinary bridge rectifier, and then PWM the dc output of that at 40 KHz with an IGBT. That will make my dc feed choke much smaller. Your feed choke is forced to operate at only 300 Hz, and it must be an enormous iron cored monster. This one will need to have a powdered iron core, but in physical size it will be fairly small for the power.

Using SCRs in the H bridge unfortunately limits the operating frequency to below 10Khz. My H bridge will use a commutation diode in series with an IGBT in each leg. The operating mode is really identical, except IGBTs allow a much higher operating frequency. The very efficient zero voltage resonant mode switching function is exactly the same in either circuit.

I feel I am on the right track with this, just getting my hands on some of the rather special parts is proving to be a hurdle, but the electronics design for it should be fairly straightforward. At least I hope so.

Awemason has it right.

Basically there are two topologies when dealing with induction power supplies.

#1) Variable DC Buss, with the inverter running at (or very close to) the tank's natural resonant frequency. - This is the on that Awemason has described.
Basically in this mode, the power (or voltage applied to the tank) is controlled by the amount of DC Buss voltage achieved.


That is what I figured. I see the variable dc bus topology as having far fewer problems, being much easier to control, and generally being more robust.

While sweeping the frequency off resonance may be very ingenious, I can see that building a control system from scratch without any prior knowledge or experience to fall back on, would be a lot more of a challenge. There are quite enough problems to solve as it is, without seeking more.

I'm not understanding the IGBT and Diode series setup. Do you have a schematic of your inverter?

Yes I do have a proposed schematic. But I am very reluctant to post completely untried ideas that others may be tempted to copy.

The basic idea is based on the conventional well known self commutating SCR full bridge converter, where the tank voltage reverses and turns off the pair of SCRs that are conducting.

The tank voltage itself turns off the SCRs, so the SCRs can never become a dead short across the tank, even if the gate triggering goes completely nuts.

IGBTs have a few problems, in that the collector cannot swing negative of the emitter more than a very few volts without destroying the device. Also under fault conditions, it may be possible for a pair of upper, or a pair of lower IGBTs to conduct together shorting out across the tank, with disastrous results.

The cure is to place a diode in series with the collector of each IGBT. That makes each IGBT/diode combination reverse blocking, (just like an SCR). If the tank tries to force the collector negative, the diode just turns off.

In fact it would be quite possible with this circuit to turn on all four IGBTs in the bridge simultaneously. No damage will occur, because two of the diodes will ensure that the particular pair IGBTs will be protected from reverse polarity from the tank.

This topology of four IGBTs plus four diodes in H bridge, should be hazard free from cross conduction caused by noise or any fault in the gate drive system. In fact the circuit will also work without any allowance for dead time. The commutation diodes ensure smooth turn off as the tank polarity swings through zero voltage, regardless of how the IGBTs are being driven.

I'm not understanding the IGBT and Diode series setup. Do you have a schematic of your inverter?

O/k here is a link to my PROPOSED driver and phase locked loop. Parts of it have already been built and tested in isolation, but the unmarked components are as yet undetermined.

I will leave this link up until the final tested working schematic is available, and I will post that in all it's glory when it finally becomes available.

http://i144.photobucket.com/albums/r166/Warpspeed_photos/2007-04-13-0831-43_edited.jpg

I think in part of that discussion above you refer to a dead short across the tank whereas you mean across the DC rails perhaps? The problem with series diodes is the energy loss, which is more heat to get rid of. Even with high efficiency avalanche type you are almost doubling the losses already encountered in the switching device itself, be it SCR or IGBT as the forward voltage drop is comparable.

On the swept frequency issue and tuning, it is important to realise that the resonant frequency of the tank circuit varies as the melt proceeds. This is for two reasons: a/ As more charge is added the inductance of the coil changes and b/ As the charge warms up and changes its magnetic properties again the inductance of the coil changes. It is very easy to detect resonance by either having an oscilloscope across the tank while tuning (best method) or a current meter in series with the DC feed to the H bridge which will peak on resonance.

If you observe the oscilloscope waveform while sweeping the drive frequency over a wide range you can easily differentiate between harmonics and sub-harmonics and the true resonance, whereas the simple current meter can easily confuse you unless you start very close to the correct frequency. I spent a considerable time doing this and working out control strategies and mechanisms when I was developing my low power pilot unit.

When I stopped my developments due to buying the 'real thing', rather than throw out my prototype or put it in a cardboard box never to be seen again I put it on ebay thinking that the power semiconductors (IGBTs) and heatsinks would be useful to somebody. Amusingly it was bought by a company making induction heaters to see if I had any ideas that they hadn't thought of, which I very much doubt!!! The pulse width modulation circuitry however was rejigged to form the basis of the motor contoller for a friends electric bicycle and is still in operation to this day <G>

There are two fairly major issues with any H bridge design that could lead to device failures.

The first is turning on either two "top" devices or two "bottom" devices simultaneously. That could potentially place a dead short directly across the tank as described in my previous post. The commutation diodes are required anyway if either IGBTs or MOSFETs are to be used to prevent the possibility of sudden catastrophic device failure. Neither of these devices have any inherent reverse blocking ability by themselves. SCRs act like diodes within themselves and block reverse voltages naturally, so this can never a problem with SCRs.

If circuit operation above 10Khz is a requirement, SCRs are really out of the question, and the extra conduction losses of IGBTs or MOSFETs, plus commutation doides is really unavoidable.

The second problem is what happens if one "top" device, and one "bottom" device on the same side of the H bridge conduct simultaneously. The dreaded evil cross conduction, which effectively places a dead short across the two dc supply rails.

But remember, in this topology the power is coming from a constant current supply, not from a stiff constant voltage source. The dc current feed choke with its comparatively enormous series inductance, prevents any sudden instantaneous rise in current, even with an instantaneous dead shorted load. The control system up stream that sets the constant current will take care of any slow changes, while the feed choke takes care of very sudden changes.

The constant current supply powering the output bridge circuit is an extremely important feature of this design topology. Destructive current spikes are simply impossible in a constant current high impedance circuit. Voltage spikes can certainly occur, but they are far more manageable with some simple clamping diodes.

What actually happens is that the H bridge is always switched precisely at the voltage zero crossings. At resonance the current is exactly ninety degrees out of phase, so at the voltage minimums, the tank current has just reached it's peak and is about to reverse. The H bridge switches just at the current reversal point, so the incoming constant current is always steered through the tank at the correct time, and in the correct direction to add to circulating internal tank energy.

Realize too that zero voltage, means not only zero volts directly across the tank, but effectively it pulls the voltage down across the whole bridge to zero as well. The constant current source is pulled down to almost nothing by the tank circuit and two conducting IGBTs. If you looked at the supply voltage to the bridge circuit it looks like the voltage coming out of a full wave rectifier. The current is constant, but the bridge supply voltage is always changing.

Zero voltage switching means exactly that. When the IGBTs switch, there is almost no voltage anywhere around either the tank or bridge. Switching losses are almost zero, and fairly slow devices can be used.

Off resonance, the tank still switches at the voltage minimums, but now the circulating tank current will have moved fairly far from its ideal ninety degrees phase. This will be particularly true with such a high Q tank circuit. The phase change in tank current either side of resonance will be fairly dramatic.

What happens is that the H bridge still switches at the voltage crossover through zero, but now the constant supply current will partly aid, and partly oppose the circulating tank current throughout the ac tank cycle.

This is an extremely important feature of this particular circuit topology. It can work safely and continuously into an extremely reactive off resonance load, with either leading or lagging power factor without any additional extra circuit stress. It can also work continuously into a dead short, as it must when the circuit starts up initially.

The phase locked loop will always track the tank resonant frequency, or at least attempt to do so. But for safe and reliable operation, the power circuitry absolutely must be able to take care of itself under worst case transient conditions.

Yes indeed, tank resonance will certainly be strongly effected by the metal load, and especially ferrous materials below the curie temperature. From my reading I believe +/- 10% frequency change is not unrealistic. But I have yet to experience any of this this for myself.

Circuit losses are unavoidable, but the extra voltage drop of one diode is not that significant in a high voltage circuit. Assuming 440v supply and 34 amps (15KW) if the diode drops one volt at 35 amps that is only 35 watts per diode additional loss. 70 watts total for two diodes (only two are working at any instant), just under 0.5%. That is not terribly serious. One solution to this problem is to run the whole circuit at a much higher dc supply voltage. While quite possible, it is not something I really wish to explore at this stage.

Warpspeed,
I like it. I hope you will let us know how things progress.

Determining tank frequency can be easily achieved with a 9V battery and an oscilloscope. Just make sure that you disconnect the tank from the inverter.
Momentarily connect the 9V battery across the tank. You will notice a sinusoidal diminishing waveform across the tank. This will be at Resonance.
I have found this to be very useful when traveling around the country, since a 9V battery is much smaller and lighter than carrying a function generator. :rainfro: This method also gives you a feel for the Q of the tank. The rate of diminishing is directly related to the Q, the higher the Q the faster the waveform amplitude will diminish.

Shawn, that is absolutely brilliant!

I have never before worked on anything of such low impedance level that had such a high Q. By my rough calculation it should have an unloaded Q of around maybe 260 which is absolutely huge. Still waiting patiently to get my 21uF tank capacitor, until that arrives I feel the whole project is really going nowhere.

Actually obtaining the parts for this has been a real emotional roller coaster.

I bid for a beautiful 10KVA 240v/120v three phase autotransformer on e-bay. I put in a bid about a minute before the close, but was beaten in the last five seconds. I finally recovered from the disappointment of that, to be contacted by the seller about a week later. He said he had three more identical transformers, and was I still interested ? I picked one up for LESS than my e-bay bid. He just wanted to get rid of them and planned to sell the remaining two for scrap copper!!!

I will tell you about the long and tortuous saga of the tank tuning capacitor later, if and when it finally arrives. So far I have paid twice for the capacitor, and three times for the UPS freight costs half way around the world. It is starting to become a very expensive component. But without it, I am sunk.

Until I actually have all the large major parts here in front of me, I cannot really begin to plan a mechanical layout and start to actually assemble things. That is why there are yet no pictures, there is really nothing interesting to show. The design really revolves around the parts I can get, particularly the mains transformer and tank capacitor.

Operating voltage, operating current, power level, and frequency are determined more by the parts I can actually obtain, rather than what I may prefer to have.

Anyhow, although progress may be rather slow, I will certainly keep everyone fully informed as to progress.

I'd like to use an induction heater to solder some small brass parts we use in our products, so I'm following this closely.

I would guess that this design will scale down quite nicely.

One thing I'm missing though is that I don't see the computer. I would probably put a micro in the system with some LED or LCD displays and some buttons. I would want to set the operating parameters that way.

Also, I would probably close the phase lock loop with the microprocessor. But, hey, I'm mostly a digital guy.

Ken

Ken, it is the concept, and the circuit topology that is important, and understanding exactly how it is all supposed to work. Scaling it down should be no problem.

My aim is to design a very simple, robust, no frills circuit that others can copy, or at least gain some useful ideas from. That is if I can actually get this working reliably.

Why a microprocessor ? All I plan to have is a big round knob to control the tank voltage and a mechanical amp meter to show the output power level.

Realize that a microprocessor and all it's I/O will be susceptible to electrical noise pickup, and any unexpected malfunction of either software or hardware could cause sudden catastrophic failure of the power devices.

If the current is rising somewhere at 200 Amps per microsecond, servicing an interrupt is simply not going to save you in time !!!!!!!!!!!

I'd like to use an induction heater to solder some small brass parts we use in our products, so I'm following this closely.

I would guess that this design will scale down quite nicely.

One thing I'm missing though is that I don't see the computer. I would probably put a micro in the system with some LED or LCD displays and some buttons. I would want to set the operating parameters that way.

Also, I would probably close the phase lock loop with the microprocessor. But, hey, I'm mostly a digital guy.

Ken


Ken,

Unless you are running a dedicated micro that is not running an operating system I think you will cripple the phase locked loop in terms of response time. Even then I think it adds complications. I reckon that having a micro doing supervisory and display functions should be ok though.

AWEM

O/k guys., latest update on progress with the induction heater.

I have finally received my tank tuning capacitor, in fact two of them.

This has all been a rather interesting if not a stressful exercise. It is a rather long and involved story, but Celem of Israel initially sent me the wrong value capacitor, which I promptly sent back. Only it was lost somewhere in transit. After much frantic worry they agreed to send me another one, trusting that I was not in fact some con man trying to pull a swifty. They were really nice about the whole thing, and I decided to order and pay for a SECOND tank capacitor. I have now received both, so the whole project can now proceed in earnest.

A second capacitor allows me the possibility to explore a lower operating frequency in the future, have a spare, or try various other things. So it is not really an extravagant waste.

I have also now received six pulse transformers, another fairly important component of the design. These pulse transformers are rated at 400 volt/microseconds with a 100nS rise time and 2.2mH inductance, with a 1:1 ratio . They allow me to drop fairly low in operating frequency if required, (either by design or mishap), as well as having a particularly fast edged square wave response. These were supplied by Schurter, part number ITRA-0239-D502.

http://www.schurterinc.com/pdf/english/typ_IT.pdf

I have finally bitten the bullet, and decided to switch to oil cooling instead of water cooling, and completely avoid any potential ionic corrosion problems. First I though I would use an oil pump salvaged from a Mazda rotary engine that I already had here in my junk pile. But it pissed oil everywhere, which does not matter when it is completely enclosed inside an engine sump. To seal it properly would be a lot more work than I really want to do at this stage. My next attempt will be with a power steering pump off something or other, which may also conveniently have it's own small oil reservoir. I am guessing it might handle the required gallon a minute or so, of flow, powered by something small like a sewing machine motor, which I also do not yet have. I have my evil eye on a candidate pump on e-bay, so we shall see how it goes.

Meanwhile I have been doing some design work on the constant current supply and control system that will feed power to the the H bridge. All the proprietary switching power supply control chips that I have looked at so far, (and tested), have fairly significant problems and/or limitations of one sort or another for this application. So I have decided to use discrete "normal" chips to do the job instead. More on this later.

You might want to take a look at zone pumps for hydronic heating systems.

Grainger has some 1/25 HP pumps for under a hundred dollars. Some advantages of these pumps include:

* Designed to be used at over 120F.
* Designed for long life with continuous operation.
* Designed to be leak free.

They are NOT designed for use with oil.

Ken

You may find a power steering pump has internal valves which restrict the flow and make it too difficult to drive with a small motor. Look for a regular hydraulic pump; you should be able to get by with a used one that leaks too much for hydraulic use but still has good flow at low pressure.

I have several different pumps here already that I have looked at, thought about, and some I have tested. Originally I had planned to use a Holley "blue" electric fuel pump. This is a small positive displacement vane pump, with an internal pressure relief valve. Depending on voltage, it pumped a measured half gallon per minute of water through my heat exchanger. It would be ideal in many respects for a water based induction heater coolant system. But it does use exposed steel internal parts, so the coolant would need to be of the corrosion inhibiting type.

http://www.holley.com/12-802-1.asp

Other pumps I have here were considered, included a hydronic hot water circulating pump, that I had used previously in a solar domestic hot water heating project. It simply did not develop sufficient pressure to pump enough flow through my heat exchanger which has roughly sixty 3/8" tubes, each just over a foot long all connected in series. Ditto with a marine fresh water pump and a washing machine pump. I even tested the coolant pump from my lathe. These centrifugal pumps have vastly more open flow than I need, but insufficient available pressure unless driven really fast with a grossly overpowered motor.

One possible candidate for the job might be an industrial spray pump. These use a wobble plate driving multi diaphragms, can develop at least 40psi and have roughly about the right flow. But I don't have one here to test, and they may not like oil. But may be worth a passing thought for for a water based system in a moderate to medium sized induction heater. Temperature is not a problem, my radiator cools the water down to only a few degrees above ambient. So a plastic bodied pump should be perfectly o/k.

Oil has roughly the same specific heat as glycol, (about half that of water), and much higher viscosity than water. So I need around twice the flow at a greater developed pressure. Oil really requires a gear or vane pump of some type. A small hydraulic pump, engine oil pump, or power steering pump would seem ideal. I ran an engine oil pump in my lathe experimentally, and it needed about 700 rpm to do the required job, but it leaked horribly. Thinking about that a bit more, the pump could be enclosed inside a sump or oil reservoir where leakage does not matter, and driven from a vertical shaft.

I don't have any hydraulic pumps here to try, it is not something I have ever played around with. Looking on e-bay, hydraulic pumps seem to be rather expensive, but would certainly otherwise be ideal.

I have previously dismantled a power steering pump, and know pretty much what to expect inside. They usually use a multi vane positive displacement rotor, considerably larger than the Holley fuel pump, so should easily develop enough flow at probably only a very few hundred Rpm. Steering pumps also normally run quite happily at several hundred psi of pressure, limited by an internal pressure relief valve. I can easily change the spring, and see no reason why it cannot work well at a much lower outlet pressure with comparatively unrestricted open flow. Being pulley driven makes driving and speed changing easy, and having it's own built in small oil reservoir is a bonus. And it will be designed not to leak.

Don't forget when comparing oil to glycol filled systems, that the glycol is actually an aqueous solution with about 20% glycol & 80% water.

Oil will obviously be inflamable so the slightest leakage in a hot spot may cause problems fast ! (But so does water !!!!)

AWEM

I realize that a greater flow will be required with oil to carry away the heat, but I don't see that as being a very major issue.

Yes Andrew, I can well imagine what might happen if a bit of molten iron gets splashed onto the very thin copper tank coil. Oil or water cooling, the situation would be rather serious.

I have thought all this through, and plan to keep the oil volume to an absolute minimum beyond that required to fill the system. One little sqirt somewhere, and with luck the pump will then run dry fairly quickly. While an oil leak somewhere will create a horrible mess, it would be nowhere near as destructive (or dangerous) to the electrical circuitry as a water leak would be.

Another idea might be to encase the tank coil in crust of castable refractory, allowing for some slight movement of the copper coil within the refractory. That should provide both electrical and thermal insulation, and some mechanical protection of the tank coil. Is this normal practice, or is there some reason why this may not be a good idea ?

....Another idea might be to encase the tank coil in crust of castable refractory, allowing for some slight movement of the copper coil within the refractory....

This prompts the question:

How do you intend to protect your coil if it is not encased in a refractory covering?

On the commercial equipment I used the coil was embedded in refractory material and there was no way molten material could come into contact with it, splashing or otherwise.

Having replaced the crucibles recently in both my furnace bodies I am perhaps TOO intimately aware of their construction which is:

A/ Each turn of the coil has 4 brass bolts whose heads are brazed on leaving the threads pointing outwards in the four compass points

B/ There are four vertical insulators running outside and parallel to the axis of the coils through which the bolts pass to support the coils very rigidly

C/ The four insulators are bolted to the bottom and top plates of the furnace body sandwich fashion.

D/ The inner surface of the turns of the coil is 'grouted' with a plastic thermal insulator or capram

E/ A layer of refractory paper is laid on top of the grout forming a tube whose outer wall is stuck to the grouting

F/ The base of the furnace body is built up with refractory cement within the bottom of this tube to give a platform for the crucible to sit on.

G/ A layer of 'drypack' (unbonded refractory compound) is put on this platform and a crucible 'bedded in' by twisting to make sure it sits level and centred.

H/ Layers of dry pack are poured between the crucible and the tube and rammed very tightly with a slender iron bar until up to within 1/2" of the crucible top, which should align with the top of the top plate of the furnace body.

I/ 'Capram' which is a plastic refractory putty is then formed round the top of the crucible and taking up the last 1/2" of the crucible/tube space.

J/ Small (1/16") holes are made in the capram to ventilate any moisture from the drypack space.

H/ The furnace is then fired at a low long heat to ensure that the drypack is indeed dry and the capram set.


The drypack forms a protective layer that will effectively stop molten metal reaching the coils in the case of a cracked crucible (yes it does work, I've had it happen !!!)

So you see, the crucible becomes an intimate part of the coil assembly

Having replaced the crucibles recently in both my furnace bodies I am perhaps TOO intimately aware of their construction which is:...

Sounds familiar :) .

Warpspeed,

I'll check back here once in awhile and put in a tip or two if I can help you out. I work in the field too, and I even met Shawn a few times, hehe.

On the cooling thing -- go ahead and use water. Even tap water will be all right, the only thing you have to do is use more than two feet of non-conductive hose per 1,000 volts of potential. Coil the hose up into loops and tie-wrap them together. Suspend the hose away from grounded surfaces where it's really close to a high potential. Try to route the water cooling to similar potentials if possible, and in the places you can't, just add a foot or two of extra hose if you're worried about it and forget it. If the hose is non-conductive (important no matter what you cool with) and it's long enough, and the water is tap water quality or better (not too hard), then it will be many years before you have a problem. We cool 15,000 volts DC with superimposed 400 KHz on it this way. I wouldn't bother with glycol, unless the thing is in danger of freezing.

And you can make your own sacrificial anode if you're really worried about it, but it's probably not necessary.

In order for you to develop the power you want, you might find yourself needing a matching transformer at the output unless you can alter the output coil turns to match the voltage. For that reason I recommend you start with more turns than you think you need and put taps on the turns. You will definitely need to be able to tune around with your capacitors to get the load into the range you want. Every time you alter the capacitance or inductance, it changes the voltage/current. And you can't develop full power until you get full current at full voltage. If you end up going too far in either direction with the inductance in order to get the voltage in range, then the Q of the circuit suffers. For multi-turn coils like used in melting, it's common to use an autotransformer that can be kept upstream of the tank.

You might need more filter capacitor than you are anticipating, not to smooth the DC, but to provide the high amplitude, short duration current pulses to your inverter section (appropriate source impedance of the DC supply at the desired frequency). And the series inductor should take into account instantaneous fault current that will be seen at your chopper and upstream when the thing fires out of sequence. Just mentioning it in case you haven't already taken it into consideration.

You might want to learn how to silver-solder if you aren't already set up to do it. Copper is easy to work with, and most if not all of your conductors can be easily made with common refrigeration tubing, with tabs brazed on. You bolt thru the tab, and connect a cooling hose to the end. You can make low impedance buswork by brazing a length of tubing to a strip of copper -- make two of them and separate them with a layer of teflon. It makes a difference in the feed to the tank circuit, and in the tank circuit itself.

There is an induction heating company with a facility in Australia, Inductotherm, which also owns Inductoheat. If you call them up and try to make friends with the service manager or some kindred soul, you might be able to talk them out of some obsolete or cast-off parts from used equipment. Heck, they might even offer you a job once they find out what you are up to! Too bad you don't live near me, I have lots of stuff around.

On another note, the frequency you are using will be just fine for what you are doing, you don't really need lower frequency until you get into larger loads. Generally, higher frequency couples better with non-magnetic materials (if you keep the coil close enough to the load).

What are you planning to use for your IGBT drivers?

--97T--

Thanks 97T, excellent advice exactly what I am seeking. While I have an electronics background, my practical working knowledge of induction heating is just about zero.

Taking your points in order:

I plan to liquid cool not only the tank components but the various IGBTs and diodes as well. Heat dissipation and semiconductor junction temperatures becomes a limiting problem long before these devices reach their full rated maximum current. Cool running devices are also faster with lower losses. Thermal design of the power electronics is a very important consideration. In short, I believe air cooled heat sinks are just not really up to the job. Liquid cooling of the power electronics also allows a very compact assembly which reduces stray parasitic inductance between the switching devices. Anyhow, as you suggest the various voltage differentials can be quite high, and massive coils of insulated flexible pipe are not going to be exactly convenient.

The way I see this is that oil offers superb electrical insulation properties, and completely eliminates any possibility of corrosion. I know I am probably alone in this, but if there prove to be unforeseen problems with oil, I can always go back to water later on. Immersion in oil is very common in very high voltage equipment for both insulation and cooling. While I suspect oil cooling may be relatively unknown in the induction heating industry, it is in common use elsewhere. So I would like to at least try the idea first and see how it goes.

Matching power into the tank is always going to be less of a problem with a semiconductor driver than with a high voltage vacuum tube driver. There are fewer volts and many more amps being switched in a semiconductor driver. I don't anticipate any matching problems directly driving the tank circuit with a fairly high switched constant current.

In my circuit the tank itself forms part of a self resonant oscillator, so it is just not really possible for the tank to go off resonance. The operating frequency can wander all over the place (and it will), but the power driving the tank will always be exactly at the tanks own self resonant frequency. The constant current driven converter topology is extremely tolerant of high reactive power and overload. It is also tolerant of random switching events and cross conduction in the H bridge driver.

The issue of sufficient filter capacitance is extremely important, and I am glad you have raised it. The constant current buck regulator will be pulse width modulated, switching on and off. When it is off, there is no power drawn at all, when it is on, there is an extremely heavy power load from the rectifier. Any series inductance in the up stream rectifier and filter circuit will create massive voltage spikes that would be death to my switching devices.

The trick is to use a low self inductance filter capacitor located right at the switching devices with almost no lead length. It must charge and discharge every PWM cycle and not allow any inductive voltage spikes to appear. It is not just a case of sufficient capacitance, but low enough source impedance at the PWM switching frequency. That capacitor also needs a high rms current rating to prevent overheating or failure. I plan to use a sufficiently rated GTO MKP snubber capacitor that is designed for that type of application. Something like this perhaps:

http://www.schusterusa.com/IMAGES/wimagto.pdf

I am very familiar with silver solder, it is wonderful stuff to work with. My semiconductor heatsinks will be slabs of copper busbar with a copper cooling pipe silver soldered onto it exactly as you describe. I have already run some thermal tests, and the heat sinking capacity of this is just amazing.

This whole exercise fascinates me, and even if I could buy a fully working commercial induction