[siamon99]

Can you help me in building the electrical part of the induction melting furnace capacity
15KW running with input 440 volts 3 phase 60HZ for melting alumnium.

[chistyakov1971]

Melting Aluminium ,
further
Centrifugal moulding of pipes and engines for rockets,
further
Assemblage of rockets.
Further
To fire rockets of the American soldiers
Joke


Induction channel furnaces.
The book on Russian, is drawings, formulas for calculation.
It is necessary to download a file *pdf
http://window.edu.ru/window/catalog?p_rid=28488
From electronics - one transformer :-)

[chistyakov1971]

Layers of a thermal protection of my new melting furnace on 2-3 kg of a steel, to pig-iron.
The skeleton is not shown.

[mrcabdriver]

Hi,

It is this topic that lead me to become a member on this forum

I'm a newbie knifemaker and induction furnace seems the only way for me to go given that i have a limited space workshop. I've been sniffing around the net for some time on this topic but never been able to find a place to ask some questions

Basically, i'd need these functions:

1. Get the steel to red hot but not liquid (then pound on it)
2. Get the steel to 300-350 degrees celsius and keep it there (for annealing - got a weird idea to build an array of IR temperature sensors that would do the trick - switch the thing on and off to keep a steady temperature)

3. The heater coil must be at least 6" in diameter and 20" high (i like big knives!)

I'm good with microcontrollers and i'd be able to make something up at the driving side if this is feasible.

All this and it can't use up more than 1-2 kW of AC power

My conceptual idea is this:

Take AC (220V) voltage through full 1:1 transformer and then to a rectifier and pass it onto the block of transistors (my thought: 4 x BUZ91 MOSFETS - each can handle 600V DC @ 6 AMPs and handle 1MHz frequency). The transistors would be driven by an ATMEL microcontroller (with proper decoupling of the two systems so nothing low-voltage fries). So 4 AMPs * 300VDC equals 1.2 kw of power for the coil.

Then we have IR sensors that feed the microcontroller with current temperature of the steel blank so it doesn't melt.

My question(s):

1. Is this even possible with such small amount of power (due to the necessary radius and lenght of the coil)? Speed of getting something red hot isn't critical - it can take a minute if necessary, i can use that time to rest from pounding on the steel anyhoo.

2. Can i compensate lower power by rising frequency. I've seen most examples work from 2kHz up to some 100 kHz. What's the difference between 100 watts @ 2kHz and 100watts @ 500 kHz, so to speak :-) ?

I'm currently making a small prototype to see how all this stuff works anyway. I'm gonna try to hook up regular signal generator to my MOSFET block (currently i have only BUZ11 - they go up to 50VDC but they can handle the frequency and current) and try driving a small coil.

[chistyakov1971]

Small capacities - induction melting has EFFICIENCY of 25-50 %, for capacity of 2-3 kw, a steel.
capacities melting - of 3 kw on 1 kg of a steel, minimum
The law of conservation of energy - great force. The law of conservation of energy you will not deceive.
Depth of penetration a current in a steel, depends on frequency.
There are some kinds of induction heating and the recommended frequencies for a steel:
Superficial heating - high frequency, from 30 000 Hz and above
Through heating - low and average frequency
Melting - low and average frequency, depending on the sizes of pieces of a steel
Depth of penetration of a current from frequency, a steel/copper, mm
2400 Hz 12,02/1,45
10 000 Hz 5,89/0,71
20 000 Hz 4,16/0,5
66 000 Hz 2,29/0,28
100 000 Hz 1,86/0,22

Depth of penetration, is that layer in which electric energy is allocated.
The sizes and parities of the sizes an induktor/detail are very important

For heating of a knife to 350-400 degrees, capacity of 2-3 kw, ОК
Frequency, 66 000 Hz are desirable.
For uniform heating: length of the coil, it is desirable, 1,4-2 lengths of a knife.
Good thermal protection, the minimum backlash between the coil and a knife.
The coil form not round, concerning a knife axis
During heating, frequency of a resonance and a consumed current will change

Computer transfer

Preliminary studying of fastening of the coil

[T3sl4]

I'm going to go out on a limb and guess you haven't forged many knives... "red hot" is only 1400F or so, a temperature where steel is soft, but just not soft enough to work with. You don't get many swings before it's too cold. You need at least yellow hot (1800F) to do real work.

If your steel is actually this hot, then you need to calibrate your eye. "Red hot" is in the fairly narrow range of 1200-1400F. I know some metalworkers who are colorblind to the point that they can hardly tell "just barely glowing red" from "bright yellow-white" (a range of maybe 1000-2300F!). If this affects you, you may consider a non-contact IR thermometer (one with a wide temperature range!). The difference is very important: steel at 2300F is easily dissipating four times more power than at 1500F, and if it takes 1kW to reach just 1500F, you'll never get to 2300 under the same conditions.

With 2kW, you might get acceptable results heating small sections to forging temperature. You'll use a small coil for this, probably a double pancake shape which the blade fits inbetween. You will never get the entire blade at once hot enough, not in air at least. With insulation, you might be able to heat treat it. You will be able to temper it, but you'll get more even heating with a domestic oven.

As for design, I have a 1kW unit (posted earlier in the thread). I have now dropped the price of plans to $50, as seen on my website:
http://webpages.charter.net/dawill/t...eat8.html#kits
I do not recommend using a microcontroller. If you want to do this all yourself, it will take a long time. Valuable experience, but you won't make many knives.

Tim

[chistyakov1971]

I agree with T3sl4, for removal of internal pressure in a steel, after heating to 200-400 degrees, very slow cooling, is necessary till 12-24 o'clock.

Heating and removal in a steel conduct in furnaces with heater R and a good thermal protection

HRC50-60, heating 180-250 degrees

Management of frequency of the microcontroller - bad idea.
Classics a genre CD4046,
The comparator LM319

[mrcabdriver]

Thanks for your answers.

Yes, i actually meant yellow hot but i'm bad with english terminology in knife making. And i am colourblind but in the green spectrum (good thing there's not green-hot in knifemaking or i'd be in trouble :rainfro

My next question - why is microcontroller bad idea? I've driven a tesla coil with one, it works even today after 3 years.

I've seen your plans online and my experience from my past projects is that stuff you get online in kits - they usually never work "As seen on TV". And since i've not built it entirely, i'd not know what to do with it other than scrap the stuff.

[chistyakov1971]

CD4046 -Smooth adjustment for frequency,good PLL
The microcontroller provides the big step of step-type behaviour of frequency and cannot be good PLL, Small speed PLL

[orjan_nilsson]

It started with a busted high pressure washer.
A little aluminium peece worth < 1$, rendering the washer out of order.
How hard can it be to make such a detail?
Got a christmas present titling "How to make a metal lath out of scrap in your garage".
Studying charcoal based aluminum melters. One year later, 5KW charcoal aluminium furnace.
It melts 3 litres of aluminium in about 1 hour. Smoky and load(big vaccum cleaner in reverse).
Tired of standing in the coold and melting aluminium and
trying to make use of my PWM high power audio bass loudspeakers knowledge I set to understand
how to use induction heating to melt aluminium?
Now I got a prototype consuming about 2.4KW of power from 240AC/10A.
Estimate that about 2KW is delivered into the stainless-steel crucible.
The contraption is made of IGBT-half-bridge(FGL60N100BNTD)-inverter and transformer/LCL.
Now the question. Experimented with different diameter cruisables and
found that if the iron-metal crucible was larger than a certain diameter I could not get ZCS.
Looking at your induction furnace coil chistyakov1971 I believe that I'am using a to large
crusiable compared to other diameter of the coil.
Experimented with LTC design programs for power electronics I discovered that,
in LCL designs there is a limit of hard you can load the resonant LC-circuit and still achieve
ZCS(switch voltage when current pass throw zero, correct on this if I'am wrong,
feel this is confusing, as I believe that sometimes they talk about ZCS and really mean ZVS and VS).
As I'am understanding this, IGBT are good at switching high voltage at moderate currents.

---
And as long as you understand my jibberish I will not excuse me writings.
PS: My background is large-scale software systems with a foot in embeded systems.
It's OK to use micro-controller if you do not put it in the PLL-control-loop.
More like controlling impedance matching circumstances and displaying values.

Sorry for not being able to compress my thoughts even more this is a to long input.

[orjan_nilsson]

And now some cool(or not so cool) pictures. About 2KW of induction heating into a stainless steel crucible with vermiculite isolation.
http://www.cnczone.com/gallery/showg...0&ppuser=14974 and http://www.cnczone.com/gallery/data/...PC170508_s.JPG.

[chistyakov1971]

Excuse, I badly know English.
The difficult childhood, shirked English lessons, drank vodka and played a balalaika.
I use computer transfer

The parity of the sizes coil and a heated up body is a big problem.
1. It is necessary to provide the minimum backlash
2. It is necessary to provide good isolation.
It contradicts each other. It is necessary to search for the compromise. There are recommendations
For a steel one decision.
For fusion of nonferrous metals - other decision.
Correct choice of frequency of a current.
Variant for the aluminium fusion, special crucibles. They are issued by various firms
http://mammut-wetro.de/

Parametres of my coil, for fusion steel
Height of 100 mm, internal diameter of 120 mm, number of coils 8
The sizes of a liquid steel
Height 100-120мм
Average diameter of 40-45 mm
Stuffing of crucible FINMIX 0,5

It is laboratory installation for small experiences
===================

I use a consecutive resonant contour
And adjustment of capacity by frequency

[chistyakov1971]

And now some cool(or not so cool) pictures. About 2KW of induction heating into a stainless steel crucible with vermiculite isolation. [Just want to insert this cool gloving image] OK I'am used to Motif and Unix systems and not how to insert a URL into response. Please help!

There are intermediate services for storage of images.
For example the Russian-speaking http://savepic.org/77850m.htm

[orjan_nilsson]

This is really cool. Being able to communicate with fellow metal-caster nerds.
My dream: Being able to a low price, manufacture broken stuff in my home.
I think we are tooo dependent on large scale manufacturing plants.
Knowledge roles!

[chistyakov1971]

My power module for the inverter of 30-40 kw

Control system, the driver - are already made and adjusted.
There was an assemblage in the case.
I will be engaged in assemblage after start of the module for fusion of a steel of 1.5 kg.

[orjan_nilsson]

chistyakov1971, are you using parallel or series LC.
What do you think are better in view of melting metal?
When I look at commercial induction ovens they usably contain about 20-50 turns of copper tubes. I estimate that transfers to about 40uH to 200uH in the work-coil(correct me if I'am wrong). (I now use 11u/190mm-diam/8turns/150mm-hight work-coil and 8u capacitor= 17KHz resonance, crucible=90mm Diameter,2mm stainless steel), if I use a larger diameter crucible I can't reach ZCS=the switching occurs when current is large(=large switching losses).

[orjan_nilsson]

What do I see on the picture, chistyakov1971?
Heavy semiconductor assemblies on copper plate=large current/high voltage contraption?

[chistyakov1971]

for "crucible=90mm Diameter,2mm stainless steel"

coil d=130mm, h=aluminium

There is enough thermal protection of 2-2.5 sm

It is required the transformer on ferrite
Factor of transformation 1/10-1/20 at connection to a line in 220 volt

I recommend a consecutive resonance on a secondary winding

Thickness of a wall of a crucible 2mm - a little, look dependence of depth of penetration of a current on frequency in a steel
=============

OK, SEMIKRON
======
Enough copper tube in diameter of 8 mm.
Backlash between coils to 2 mm
Tube isolation is desirable.

The transformation factor steals up taking into account inverter loading
The core form is desirable III + III .

Specific capacity a minimum of 1 kw / kg

[orjan_nilsson]

At the moment I use a 2.5KG homemade ferrite core for transformer(now 1:1 coupled, it's made of crushed fly back transformers from scraped TV-sets... bound by silicon-rubber, gives about 0.13uH/n2, commercial IH-transformers gives about 3-6uH/n2) and a series inductance of estimated 50uH.
Is "consecutive resonance" series resonance? I use parallel LCL topology.
Because this gives small currents demands supplying the resonant large-current CL-circuit.

[orjan_nilsson]

If copper tubes, wall-thickness 1mm(water cooled in final assably) are isolated by teflon(ca 200 degrees Celsius) would that be enough?
Why is it necessary to transform to 1/10-1/20. Currently I use 1:1 transformer and achieve optimal power throughput. If I put a larger diameter crucible in(about 150mm-diam/150mm-height) I'am not able to keep ZCS. If I change the transformer ratio I'am able to keep ZCS. This at the cost of increased current to the LC-resonance circuit. )
What is the recommended thickness of iron-crucible when melting aluminium at 17KHz?
Should I protect the inner wall of the crucible with some "stuff"?
What happens with aluminium at 700 degrees Celsius in contact with iron(and for that matter nickel/chrome if stainless steel is used for crucible)?
A lot of questions from Mr Warrum.