Author Topic: Building electric furnace to melt aluminium  (Read 10210 times)

0 Members and 1 Guest are viewing this topic.

Offline ^_^

  • Regular Contributor
  • *
  • Posts: 82
  • Country: au
  • EE
Building electric furnace to melt aluminium
« on: March 27, 2017, 04:46:52 pm »
Hey  :-+

last summer I've made concrete furnace with a friend, powered by coal, to melt aluminium. It turned out pretty cool. We did some hand castings and stuff.

This time I wanted to step it up a bit and make it powered by electricity.
The main reason is electric furnace will also be able to heat proffesional casts (like Kerr Cast material) according to their profile with temperature control, which enables making good-looking castings.
All the other reasons, like coal-handling, weight, preparation time also prove electric furnace advantage.

I've got some questions regarding the "design", though.
This time, I've decided to make a truly engineering approach towards this build ;)

I'm assuming a temperature of 1000*C as this should easily melt alloys I have, plus it gives some heat capacity margin when pouring to a form.
Assuming I want to heat just 0.6L of aluminium (I assumed 0.45L of alu and 0.15L for it's crucible, very roughly).

Code: [Select]
903 J/(kg*K) * 2702 kg/m^3 * 0.6L ~= 1500 J/K
Which gives, for 1000*C:
1.5kJ/K * 1000*C ~= 2MJ
(the little fail here is, I took specific heat at room temperature, but should be OK...)

Now, providing 2kW just to aluminium:
2MJ / 2kW ~= 17 min

Seems doable.
So now I get to thermal insulation, which seems to be the most important part of the furnace.
I've used special insulating bricks before not to damage the lawn. I've looked at them again in a shop and they specify heat coefficient.
Basically there are 3 bricks available in my area (10cm, 7.5cm and 5cm thickness). They have 1.4W/(m^2*K), 1.9W/(m^2*K) and 2.8W/(m^2*K) coefficients respectively.
If I divide their coefficients by respective thickness I get 0.14W/(m*K) for each of them, which seems to be heat conductivity of the material.
Anyway, the coefficient is provided for room temperature. Somewhere, I've found it can change by 50% at 1000*C.

So now I calculate losses into the walls (assuming they are 25x25cm as well as height):
Code: [Select]
(1.4W/(m^2*K) + 60%) * (1000*C) * (25cm)^2 * 6 ~= 1070W

So I need a heating element of >3kW.
For example, at 14Amps 230VAC, I get 3.2kW.
Unfortunately I'm limited to 16A per phase and 25A in total. It's a house not a workshop :(

Now I need a heating wire,
Code: [Select]
R = (230V)^2/(3.2kW) = 16.5 Ohm

Or I could go with two wires at 5kW total:
Code: [Select]
R = (230V)^2/(2.5kW) = 21.2 Ohm

I've already bought 20 meters of Kanthal A1 type wire (operating temperature 1400*C, melting point 1500*C), 1mm diameter.
It has 1.85 Ohm/m, so 9 meters should be enough.
Or will it be?
I can't find any info about the max current, or power per meter for that wire even though the datasheet is pretty awesome.

What I did found though, is a "Maximum recommended surface loads" (W/cm^2) chart in Kanthal industrial type datasheet (http://www.kanthal.com/Global/Downloads/Furnace%20products%20and%20heating%20systems/Heating%20elements/Metallic%20heating%20elements/S-KA041-B-ENG_2011-09.pdf).
From the other Kanthal home appliances datasheet (http://www.kanthal.com/Global/Downloads/Materials%20in%20wire%20and%20strip%20form/Resistance%20heating%20wire%20and%20strip/S-KA026-B-ENG-2012-01.pdf):
I can read the surface area of my 1mm wire: 31.4cm^2/m.

9 meters of that wire will have 282.6cm^2, which at 3.2kW gives 11 W/cm^2.

Seems too much according to the datasheet, but I don't think this is the way to calculate it (take the surface of the wire itself). Or is it?

Anyway, let me know what you guys think,
is there any righteousness in this approach and my very rough calculations?
Thanks for reading "TL;DR" post  ;D ;D

 
The following users thanked this post: jonovid

Offline danadak

  • Super Contributor
  • ***
  • Posts: 1875
  • Country: us
  • Reactor Operator SSN-583, Retired EE
Re: Building electric furnace to melt aluminium
« Reply #1 on: March 27, 2017, 05:01:23 pm »
Take a look at Youtube, "induction heating", many DIY experiments
and solutions.


Regards, Dana.
Love Cypress PSOC, ATTiny, Bit Slice, OpAmps, Oscilloscopes, and Analog Gurus like Pease, Miller, Widlar, Dobkin, obsessed with being an engineer
 

Offline mycroft

  • Regular Contributor
  • *
  • Posts: 50
  • Country: br
Re: Building electric furnace to melt aluminium
« Reply #2 on: March 27, 2017, 05:07:42 pm »
Note you also need to heat the bricks mass.
 

Offline ^_^

  • Regular Contributor
  • *
  • Posts: 82
  • Country: au
  • EE
Re: Building electric furnace to melt aluminium
« Reply #3 on: March 27, 2017, 05:49:00 pm »
Induction heating is not using resistive wire, it's totally different approach, that I'm aware of ;)
This approach is much easier, almost no electronics except temperature control, which can be bought.

Note you also need to heat the bricks mass.

Good point.
Nevertheless, as the bricks get hotter, the temperature difference gets smaller, so I've calculated for the worst case scenario it seems.
 

Online jbb

  • Frequent Contributor
  • **
  • Posts: 891
  • Country: nz
Re: Building electric furnace to melt aluminium
« Reply #4 on: March 27, 2017, 07:06:51 pm »
The aluminium will expand a little when melted (I guess you've seen that already).  And there will be a little more heat load because of the latent heat of fusion (aka melting energy).

Have you chosen a suitable crucible? It seems like the energy is transmitted radiantly from the heating coils to the work (given high temperature), and you'll need to work out how hot the coils get to radiate how much energy into the crucible.  This actually sounds like a job for a mechanical or chemical+materials engineer...

Two electrical thoughts:
  • you may from time to time splash a little aluminium.  Think carefully about what happens if a splash hits a live heating coil.  Ground/earth the metal frame of the furnace so that you blow a fuse / rip a breaker rather than have the whole thing go live.
  • think about what happens if the wires fall off your thermocouple, or the thermocouple becomes dislodged.
 

Offline Gyro

  • Super Contributor
  • ***
  • Posts: 6017
  • Country: gb
Re: Building electric furnace to melt aluminium
« Reply #5 on: March 27, 2017, 07:20:47 pm »
Wouldn't you be better using gas? A lot more controllable than coal and a lot less complex than electricity.
Chris

"Victor Meldrew, the Crimson Avenger!"
 

Online Conrad Hoffman

  • Super Contributor
  • ***
  • Posts: 1304
  • Country: us
    • The Messy Basement
Re: Building electric furnace to melt aluminium
« Reply #6 on: March 27, 2017, 07:24:34 pm »
I'd look at commercial potters and glass-working kilns to see what type of heating elements they use and what temperatures they reach. IMO, losses due to air leakage and such may be higher than you think.
 

Offline ^_^

  • Regular Contributor
  • *
  • Posts: 82
  • Country: au
  • EE
Re: Building electric furnace to melt aluminium
« Reply #7 on: March 27, 2017, 07:32:05 pm »
Wouldn't you be better using gas? A lot more controllable than coal and a lot less complex than electricity.

To just melt, propan-butan is the way to go, but as I said I want "2in1" solution.
Proper cast has a temperature profile to follow, almost like the PCB-reflow profile ;)
I have already prepared a hand made of wax (attachement), that I want to cast in aluminium.
So I need to put it in cast, heat it somewhat following the profile, make the wax inside melt and escape, and the form shall be ready.

BTW this project is just for fun :)
« Last Edit: March 27, 2017, 07:37:13 pm by ^_^ »
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 8046
  • Country: de
Re: Building electric furnace to melt aluminium
« Reply #8 on: March 27, 2017, 07:36:10 pm »
5 kW sounds like not very much power to reach near 1000 C for that size.
The thermal conductivity usually goes up quite a lot at high temperature, as there is more and more radiative heat transfer. There can also be quite some heat loss due to air currents.

Thermal expansion of the wire can also be tricky, so the wire will need some support. Also consider what happens if the wire should break.

1 mm wire seems to be very thin for high temperature use.

It might still be easier to use the electric furnace to e.g.  600 C (for the cast) and gas for melting.
 

Offline ^_^

  • Regular Contributor
  • *
  • Posts: 82
  • Country: au
  • EE
Re: Building electric furnace to melt aluminium
« Reply #9 on: March 27, 2017, 07:36:40 pm »
It's done like that in:
https://www.amazon.co.uk/gp/product/B00KYN6T2Q/ref=as_li_qf_sp_asin_il_tl?ie=UTF8&camp=1634&creative=6738&creativeASIN=B00KYN6T2Q&linkCode=as2&tag=youtube071cf-21

And the guy that inspired me to build this:


He built it, but in a way that 'it just worked out'.
I just wanted to take a little more sophisticated approach ;)
Hence questions about the heating wire.

Quote
Have you chosen a suitable crucible?
Yes, I have a graphite crucible, the most pro element in the whole build ;)

Quote
Think carefully about what happens if a splash hits a live heating coil.
I think turning it off while operating with crucible seems enough.

Quote
Thermal expansion of the wire can also be tricky, so the wire will need some support.
True, fortunately it's not much for Kanthals.

Quote
1 mm wire seems to be very thin for high temperature use.
Idk, they say 1400*C and the wire they make starts at 1mm, so why not  ;D
« Last Edit: March 27, 2017, 07:43:50 pm by ^_^ »
 

Offline Etesla

  • Regular Contributor
  • *
  • Posts: 109
  • Country: us
Re: Building electric furnace to melt aluminium
« Reply #10 on: March 27, 2017, 07:50:51 pm »
Last summer I made a kiln for heat treating steel. It gets up to about 2000 F in about an hour, and gets to about 1400 (good enough for aluminum I think) in maybe 25 mins. The internal surface area of the kiln is about 2500 cm^2, which is large enough for everything I do, and can be built vertical to easily accommodate for a crucible. Total internal area is about 7000 cc (7 liters). It runs at about 2 kw, switched by a cheap solid state relay.

Here are the maths you will need to do.

Terms:
Power Density: The amount of power you want to pump into your kiln. units: (watts / cm^2). Typical value for power density for your application is anywhere from .7 to .9. equation: Power Density = watts / internal surface area of kiln

Surface Loading: A value to represent whether you are using your gauge of wire efficiently or not. High values mean you are stressing you element wire out and will probably have to replace it. Low values mean you kiln heats up a little slower, but your elements will last a long time: units (watts / cm^2) typical value for surface loading for your application is anywhere from 1 (pretty slow) to 4 (pretty fast), equation: Surface Loading = watts (same value you use in power density) / external surface area of wire (get your geometry on)

Example calculations:

Power Density:
My target power density is .8, and my design calls for an internal surface area of 2000 cm^2, so .8 = Watts / 2000cm^2.
Watts = 1600 watts


Surface Loading and wire length:
This depends on the gauge of wire you choose to use. Different gauges will yield different results for the surface loading coefficient. Try to use the largest reasonable wire diameter and achieve a reasonable surface loading coefficient. In this example I will use 13 gauge kanthal A1 wire. All specs are straight off the datasheet.

Wire Length:
Specs for 13 gauge kanthal A1:
diameter = .1829 cm
resistance = .174 ohms / foot (at highest temperature I expect to run it from, again, look at the datasheet)
Since I plan to run at 1600 watts, and I know my operating voltage (120 v residential), I can calculate my required resistance.

V = IR

Watts = VI
I = Watts / V

V = Watts / v * R
V^2 = Watts * R
120^2 = 1600 * R
R = 9 ohms

if the wire has a resistance of .174 ohms / foot, I would need roughly 51 feet of wire, or about 1550 cm.

Surface area of wire is calculated as though the wire were a very long cylinder.
Diameter = .1829 cm

SA(cylinder) = circumference * height
= 3.14(pie)*.1829(diameter) * 1550("height")

= 890 cm^2

SL = Watts / SA = 1600 / 890 = 1.7977

Surface Loading:

SL = W / cm^2
 = 1600 / 890
 = 1.8, a pretty low value, so a slightly thinner wire may better suit the purpose as it would heat up quicker (increase wire gauge = lower SL and longer element life, but slower to heat up)... At this point I would change wire gauge and try again until I get a surface loading coefficient of about 2.5. It looks like you already picked a gauge, but just go through the calculations and make sure it stays within my recommended values.

From here, after I have picked a wire gauge, I would do my best to estimate the length of the coiled element (which I assume you will be doing on your own), and make sure its not absurd. good rule of thumb I used for winding the element around a half inch mandrel will result in an final wound element length about 8 x smaller than the length of the wire before it is wound.

So far it looks as though you have tried to use chemistry knowledge and perhaps some research to come up with values for your power and stuff like that, but I will recommend that you use the practical knowledge gained by people in the field of kiln making to your advantage. Here's a great link if you have any more questions: http://www.euclids.com/Html%20pages/element-design.htm (not html) Also use type k 23 firebricks (easy to work and cheap if you don't have to pay for shipping. My local pottery supply store had them for I think 2 bucks a piece), and wrapping in some kaowool for extra insulation can't hurt and is cheap.

Good Luck!
 
The following users thanked this post: edavid, ^_^

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 16035
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Building electric furnace to melt aluminium
« Reply #11 on: March 27, 2017, 09:11:14 pm »
I've done this before.  The electric furnace I made was rather small (under 1kW), and used inappropriate wire (stainless steel, which burned out from time to time, and can't really be spliced in a reliable way once it's been installed).  It was a neat experiment, anyway.  I've also done induction heating (5kW) and got the most use out of propane fired furnaces (crucible and reverberatory types).

The numbers you give look pretty reasonable.  I'm not clear on what type of brick that is; just make sure it's refractory, not a gypsum or concrete based product (which is bonded with hydration -- guess what happens when you heat it up again?).  You'll have to cut slots to hold the wire coils, I guess; go carefully so it doesn't get chipped too badly.

Great thing about electric: you can also fit a PID controller and do pottery with it! :)

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 
The following users thanked this post: ^_^

Offline MagicSmoker

  • Super Contributor
  • ***
  • Posts: 1419
  • Country: us
Re: Building electric furnace to melt aluminium
« Reply #12 on: March 27, 2017, 10:27:26 pm »
...(stainless steel, which burned out from time to time, and can't really be spliced in a reliable way once it's been installed).

Did you try spot welding it? I built a 3000A load bank out of stainless steel sheets that I spot welded together* and it held up well enough even when glowing a dull red for several hours.


* - don't laugh until you've priced one of these things; we were quoted $50k.

 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 16035
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Building electric furnace to melt aluminium
« Reply #13 on: March 28, 2017, 02:25:25 pm »
Did you try spot welding it? I built a 3000A load bank out of stainless steel sheets that I spot welded together* and it held up well enough even when glowing a dull red for several hours.

Not sure there'd be much left to spot weld once it's been hot and crusty.

But no, didn't try.  Never did make a spot welder.

Quote
* - don't laugh until you've priced one of these things; we were quoted $50k.

I wasn't going to.

I've used steel strip as load resistance before.  The galvy stuff even turns yellow and gently releases magic smoke to tell you when it's ready......  :-DD

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline ^_^

  • Regular Contributor
  • *
  • Posts: 82
  • Country: au
  • EE
Re: Building electric furnace to melt aluminium
« Reply #14 on: April 08, 2017, 12:00:56 pm »
Thank you bros!!

I see what I've done wrong, you guys are right, I need special fire-resistant bricks. The ones I had will not insulate enough at high temperatures and might break.
There's an alternative: kaowool.
Right now I'm researching which one will be cheaper and will fit my needs better.

I have also done an Inventor model of the furnace in which I have used firebricks available in my location.
Here's the render:


The project: https://github.com/stawiski/electric-furnace

I'm gonna get back here when I finish the research, buy the stuff and build it.
See you guys then!  ;D
 

Offline calexanian

  • Super Contributor
  • ***
  • Posts: 1878
  • Country: us
    • Alex-Tronix
Re: Building electric furnace to melt aluminium
« Reply #15 on: April 09, 2017, 01:25:09 am »
I have built furnaces like this before. Kanthal is the most practical element to use. When you stretch it do the whole length at one time. I tied one end around a bolt in a bench and vice grips on the other end and stretched it out.

Yes it will expand during the first heat, but it will also contract a little when it cools back down.

Also do not do what the guy in the video did and make the connections on the inside of the furnace to the external wires. Make them outside. Also use a stainless gold or a LARGE brass one. Lots of thermal mass in the connection point. Otherwise the end of the element will become too brittle. Kanthal only survives the high temperatures because it forms an even oxide layer and that can embrittle the material.

About the maximum current being hard to find this is to be expected. That will depend of the thermal load in whatever it is heating. In a vacuum it will require far less current to reach its max temperature than if it were say imbedded in something that is a large thermal sink such as a concrete block as shown in your picture. Blocks like that and silicate fire bricks are very different things. Kaowool or ceramic blanket have a lower thermal transmission but have no physical structural strength. Spend the money on the silicate block. Avoid all mill boards or other non brick shaped high temperature boards. They have binders that need to be burned out and have very little structural strength afterwards.

Either get a very large variac or a SCR type oven control to regulate the temp. The controls are not very expensive and can usually be found at places that sell appliance parts. As for a Variac that may be a bit more difficult. They really do a great job though. The pre made kanthal elements I have bought were made to operate in free air at 240 volts max right at their max operating temp. In a furnace they would quickly exceed their max temp without a control do to heat gain under light load. the high current was not an issue as the current actually will drop as the temp increases.   
The variac would allow you to run at a lower temp and a lower current from the wall.

Otherwise the one in the video is actually a pretty decent design.
Charles Alexanian
Alex-Tronix Control Systems
 
The following users thanked this post: ^_^

Online Cerebus

  • Super Contributor
  • ***
  • Posts: 6218
  • Country: gb
Re: Building electric furnace to melt aluminium
« Reply #16 on: April 09, 2017, 02:36:21 am »
Avoid all mill boards or other non brick shaped high temperature boards. They have binders that need to be burned out and have very little structural strength afterwards.

Not quite all, there are boards available that are used for shelves in kilns that have to take quite substantial loads. Don't know offhand what the material's called, it's the other half who's the potter.

Either get a very large variac or a SCR type oven control to regulate the temp. The controls are not very expensive and can usually be found at places that sell appliance parts.

Aw, c'mon. Buy one? Where do you think you are? It's not often a bloke gets a good excuse to start a project that involves power semiconductors that you don't solder to but instead bolt thick cables to.  :)
Anybody got a syringe I can use to squeeze the magic smoke back into this?
 

Offline SeanB

  • Super Contributor
  • ***
  • Posts: 15434
  • Country: za
Re: Building electric furnace to melt aluminium
« Reply #17 on: April 09, 2017, 04:38:45 pm »
Remember with the bolt on SCR construction you will need snubbers, along with some kind of large chokes so you do not feed back radiated noise into the entire district power lines.  The snubbers are available with a bolt pattern similar to the SCR devices, and bolt onto the bus bars of them as well, but are almost as expensive as the SCR units themselves.

Do not forget you will need some power factor correction, or at least some harmonic filtering, using some industrial power factor or motor run capacitors directly across the mains. And of course a really beefy mains switch, along with some input circuit breakers as well preferably across active and neutral as well for single phase, or all 3 phases if you are doing 3 phase, and there you just do the load heaters as delta so you do not need a neutral connector.
 

Offline calexanian

  • Super Contributor
  • ***
  • Posts: 1878
  • Country: us
    • Alex-Tronix
Re: Building electric furnace to melt aluminium
« Reply #18 on: April 09, 2017, 07:46:51 pm »
It's funny to hear you say use motor run caps across the mains. Totally illegal here. You will get a very angry utility company if too many people do that. In actuality there are very little
PFC requirements here in the US. If you go capacitive as a load though, the utility will shut you off berry quickly.
Charles Alexanian
Alex-Tronix Control Systems
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 16035
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Building electric furnace to melt aluminium
« Reply #19 on: April 09, 2017, 09:11:19 pm »
It's funny to hear you say use motor run caps across the mains. Totally illegal here. You will get a very angry utility company if too many people do that. In actuality there are very little
PFC requirements here in the US. If you go capacitive as a load though, the utility will shut you off berry quickly.

Can you cite a utility source for that?

I'd be very interested to know about this.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline calexanian

  • Super Contributor
  • ***
  • Posts: 1878
  • Country: us
    • Alex-Tronix
Re: Building electric furnace to melt aluminium
« Reply #20 on: April 10, 2017, 02:37:58 am »
I cannot site any specific example at the moment as I learned all of this from our local utility many many years ago. Pre internet or digital source, but..  The problem is that on a 3 phase delivery system if one leg goes capacitive you get a phase shift in that particular leg and can cause all sorts of sag and surge in the other phases. Very problematic in neighborhoods and small industrial parks. The grid can easily handle inductive and resistive loads and even non linear loads, but capacitive is a big problem.
Charles Alexanian
Alex-Tronix Control Systems
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 14686
  • Country: gb
  • 0999
Re: Building electric furnace to melt aluminium
« Reply #21 on: April 10, 2017, 08:11:25 am »
Remember with the bolt on SCR construction you will need snubbers, along with some kind of large chokes so you do not feed back radiated noise into the entire district power lines.  The snubbers are available with a bolt pattern similar to the SCR devices, and bolt onto the bus bars of them as well, but are almost as expensive as the SCR units themselves.

Do not forget you will need some power factor correction, or at least some harmonic filtering, using some industrial power factor or motor run capacitors directly across the mains. And of course a really beefy mains switch, along with some input circuit breakers as well preferably across active and neutral as well for single phase, or all 3 phases if you are doing 3 phase, and there you just do the load heaters as delta so you do not need a neutral connector.

No need for filtering or PFC.

Use burst control with zero crossing and make sure you keep the on/off times to even numbers of mains cycles and the harmonic content will be minimal.
 

Offline CaptainNomihodai

  • Contributor
  • Posts: 37
  • Country: us
Re: Building electric furnace to melt aluminium
« Reply #22 on: April 10, 2017, 04:56:05 pm »
I've been experimenting with resistive heating furnaces for a while, and I finally built a ~5kW (240V at about 20A) furnace a while back that I'm mostly happy with. It can fairly easily melt not only aluminum but copper too (but that's scary hot, so don't do that very often). Hopefully some of the things I've learned can be helpful...
My design (I'll post pictures later maybe) is based largely on the fairly common firebrick hot face (most commonly a hexagon, but see below) surrounded by insulation/refractory all enclosed in a section of circular air duct.
1. Hot face: I use an octagon made of K26 (alumina/silica) insulating firebrick, with K26 on the bottom too.
I use an octagon for a couple of reasons:
-First, to increase the surface area of the hot face. As I think somebody mentioned, you need to pay attention to your power/surface area. The maximum will be specified in your heating element datasheet. On one of my earlier furnace attempts, I burned out a relatively expensive custom Kanthal element, even though I was well within the limits of the power per surface area of the wire (more on this below), because my surface loading on the hot face was too high. Damn did that heat up quickly, though. More surface area also means a longer groove, which allows you to stretch out your element more which lengthens its life.
-Second, to make the grooves more "circular." Using thick Kanthal, which you need at this power level, the element is going to naturally assume a circular shape and you're going to have to fight it to keep it in the corners. If I built another furnace I might actually go with a decagon or dodecagon. As it is now, I'll probably have my next heating element made with periodic uncoiled sections for the corners.
Regarding the firebrick: K26 is nice because it is super insulating and easy to cut (there are a bunch of blogs/videos/etc explaining how to use a table saw to get good grooves). However, it is incredibly fragile. If I built my furnace again I would use hard firebrick as the hotface. It's much less insulating but isn't going to break if you accidentally bump it with your crucible. I would then surround it with a layer of K26.
2. Heating element: Kanthal A1 is the way to go. I use 12 or 13 AWG with something like a .375" inner diameter (all my notes are at home so I'm going off memory). This stuff is about as stiff as steel, so winding your own element with anything thicker than 16AWG, unless you have a lathe, is probably out of the question. Even stretching a pre-wound element is tough, which is why I ordered mine pre-stretched. I had a pottery supply company make my element to my exact specifications for about 100USD. There are various parameters to consider, and you need to play with the math until you've got something you're happy with. I made an Excel file that does the calculations that I'd be happy to share. Your most important parameter is the surface loading on the wire (i.e. power per unit area of the wire surface), keeping this as far below the maximum as possible will prolong the heating element's life. Kanthal forms a protective oxide layer, but it will eventually oxidize all the way through and break; too much power through too little wire is a good way to ensure that this happens very, very quickly.
3. Refractory "insulation": What I've often seen with other hobbyist furnaces (and what I did myself) is hot face bricks surrounded by refractory and then the outer shell. Commercial kilns seem to have the hot face bricks right up against the outer shell. In my furnace the refractory is, if I recall correctly, some mixture of fireclay, sand, and commercial refractory that I came up with after much experimentation (some people like pearlite mixed with commercial refractory, but that comes with its own set of problems). I would not use this approach again. What I failed to consider is that refractory does not necessarily imply insulating. My "insulation" stands up to the heat very well, but it conducts the heat a lot more quickly than the K26 (though, still less than air). What's worse, it's about 4" thick. This means, first, that it's incredibly heavy. Second, it keeps the outer shell of the furnace cool for quite a while, but once it heats up it has so much thermal mass that it takes forever to cool off again (sometimes over 24 hours). In fact, I've observed that the outer shell often gets hotter after I've turned the furnace off. The best I can figure is that this is the heat slowly but surely migrating outward through the "insulation." If I did it again, I would surround the hot face bricks with a few layers of kaowool. It's light, it's cheap, and it's a really good insulator.
4. Electronics: A cheap PID controller with a solid state relay mounted in an ATX power supply case, combined with a thermocouple, and, voila, you have professional-looking temperature control.
 
The following users thanked this post: ^_^

Offline Peter.L

  • Newbie
  • Posts: 1
  • Country: au
Re: Building electric furnace to melt aluminium
« Reply #23 on: April 10, 2017, 10:17:01 pm »
Hi All,
my first post here ...
I've built an electric kiln as well (seems to be popular hobby these days). Here is my 5 cents

To reduce thermal mass of the system you can use ceramic blankets on the outer shell and K26/K29 bricks or hard-face alumina bricks(non insulating) on inside. it helps kiln to warm up faster and keep mass down.

You can apply IR reflective ceramic on K25 bricks  something like Satanite or ITC-100 will do, it not cheap but helps with brittleness of silica bricks. Also it's IR reflective and helps to retain more heat in the chamber. Also it resistive to fluxes (borax) which attack silica.

I've built a simple TRIAC to control voltage with this thing and clamp amp meter you don't need to do precise calcs on your heating element (easier to replace heating element if you want to alter your design later)

An the last one Inkbird Digital Pid Temperatur?e Temp Controller Control Thermometer (~30$ delivered from ebay) is a gold and helps a lot.
 
 
 
The following users thanked this post: ^_^

Offline ^_^

  • Regular Contributor
  • *
  • Posts: 82
  • Country: au
  • EE
Re: Building electric furnace to melt aluminium
« Reply #24 on: April 11, 2017, 09:14:44 pm »
Thanks, thanks!  :)

As for the electronics, SCR, PID etc. temperature control overall, I'm gonna look into it after I get some results (melting aluminium is good enough indicator).
I want to do it step-by-step, because I used to be the kind of guy that buys it all and never finishes a project ::) so I'm taking the money-efficient approach here.
Surely gonna get back to that after I get some results!

I have ordered the firebricks, some kaowool and mortar today. All from http://www.vitcas.com/ as they have a distributor at my location.

The fails:
- The first bricks I had were not firebricks, just home insulation. That's why it didn't work. Their insulation coefficient was probably super-bad at high temperatures and they would break if reheated as you guys said. Thanks for pointing me to the right bricks!
- I did not know about proper wire surface load values then. I did read Kanthal datasheet though. They are being weird about this. The SL seems to depend on the application.

I was using this as a reference:

Though, it is provided for >=5mm diameter and short loop lengths. I guess it's for multi-kW furnaces as they give examples for 3x 40kW  :scared:

I've also read on some pottery forum that I'm supposed to obtain SL at the level of 2.5W/cm^2, which is similar to what you guys mentioned. Well, that'd be hard to do now, as I have only 1mm dia wire (2mm would be better).
I'm gonna design universal wire holders inside the furnace, so that the wire could be easily replaced (preparing for the future ;) ). Was thinking a lot about it, the use of ceramics and other heat-resistant elements or just "standard" grooves in the bricks. The easiest way I've though of was prepared pieces of steel, kind of "steel holders", put between the bricks that are glued together with mortar.
Could thermal expansion of that small steel element create problems?

New math, MATLAB script (https://github.com/stawiski/electric-furnace/tree/master/math) /* yes, it's overengineering to use MATLAB for that */ says:

Heating wire:
Code: [Select]
Power = 2000 W
grid voltage = 230 V
wire Ohm/m = 1.85 Ohm/m
wire dia = 1.00 mm
wire current = 8.70 A
wire length = 14.30 m
wire Surface Load = 4.45 W/cm^2

As I have 20m + 10m of Kanthal wire atm. that will give 2 phases, 4kW total.

And the furnace:
Code: [Select]
Power wire = 4000.00 W
Power loss = 1005.94 W
Furnace heating power = 2994.06 W
Furnace capacity = 15.63 L
Furnace heating power per capacity = 191.62 W/L
Estimated time to heat 1.00 L of alu to 800.00 *C = 10.87 mins

Seems it would even work with just one phase at 2kW, but would heat up long ;)
« Last Edit: April 11, 2017, 09:18:54 pm by ^_^ »
 

Offline 691175002

  • Regular Contributor
  • *
  • Posts: 64
Re: Building electric furnace to melt aluminium
« Reply #25 on: April 11, 2017, 09:30:38 pm »
Kilns are actually quite easy to build.

The most common brick is K23 which can be cemented or just stacked.  The heating elements are normally mounted in milled grooves, but you can also "staple" them to the brick with loops of wire.

Drill a hole for the thermocouple and element leads, and buy a $20 PID + SSR off ebay.
 

Offline CaptainNomihodai

  • Contributor
  • Posts: 37
  • Country: us
Re: Building electric furnace to melt aluminium
« Reply #26 on: April 12, 2017, 03:24:02 pm »
I'm gonna design universal wire holders inside the furnace, so that the wire could be easily replaced (preparing for the future ;) ). Was thinking a lot about it, the use of ceramics and other heat-resistant elements or just "standard" grooves in the bricks. The easiest way I've though of was prepared pieces of steel, kind of "steel holders", put between the bricks that are glued together with mortar.
Could thermal expansion of that small steel element create problems?

Absolutely do not use steel to hold your elements. At those temperatures steel will form a cool-looking black oxide layer. As it heats and cools the oxide will flake off, and these flakes can short out and destroy your element (I know this from experience using a steel crucible). They will also eventually oxidize away to nothing. You may also get the element and the steel "welding" themselves together (not an actual weld, just the oxide layers growing together). Finally, it just seems, intuitively, like having the element touching another metal that isn't Kanthal can't be good for it. Also, if you're using soft firebrick as your hotface, I would not recommend gluing them together with mortar. Those bricks are very fragile and you want to be able to easily replace one if it breaks.

Quote
I did not know about proper wire surface load values then. I did read Kanthal datasheet though. They are being weird about this. The SL seems to depend on the application.
Basically the more "stretched" your element is (i.e. larger diameter and pitch, for a coil), the more surface load it can take.
 
The following users thanked this post: ^_^

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 16035
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Building electric furnace to melt aluminium
« Reply #27 on: April 12, 2017, 04:42:58 pm »
Also, iron oxide is a flux, so it will melt into the firebrick, staining it at least, but probably causing melting, especially around the element.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline ^_^

  • Regular Contributor
  • *
  • Posts: 82
  • Country: au
  • EE
Re: Building electric furnace to melt aluminium
« Reply #28 on: April 15, 2017, 05:41:46 pm »
@691175002:
I have used a similar brick to K23: IFB23. They have good parameters as well.
PID and SSR coming in the near optimistic future.

@CaptainNomihodai:
That's interesting as I've seen guys build whole crucibles out of welded steel :D
But I did listen to your advise and went with the in-brick grooves.

As for the mortar - I've already bought it, and the bricks do not seem *that* fragile, so I've went with it.

### UPDATE ###

Inventor screenshot vs. Inventor 48h render:


Right now the mortar is setting, 4 (+1 divided into pieces) more bricks will be added on top.

The wire i have now:
Code: [Select]
### Main ###
Power = 2000 W
grid voltage = 230 V
 ### Wire ###
wire Ohm/m = 1.85 Ohm/m
wire dia = 1.00 mm
wire current = 8.70 A
wire resistance (24*C) = 26.45 Ohm
wire length = 14.30 m
wire Surface Load = 4.45 W/cm^2
 ### Coil ###
coil diameter = 15.00 mm
min. coil length = 0.91 m
max. recommended coil length = 1.82 m

With a coil length of 1m (dimensions of the base of the inside of the furnace are roughly 25 x 25 cm).
Having three levels of grooves I can add the same coil two times (and get overall 3-phase 6 kW furnace...  >:D).
However, realistically I can add it only once because of my house power limit.

If I had more money for this project and higher current limit per phase in my household I'd go for 2mm Kanthal, one-phase 4.3kW furnace:
Code: [Select]
### Main ###
Power = 4300 W
grid voltage = 230 V
 ### Wire ###
wire Ohm/m = 0.46 Ohm/m
wire dia = 2.00 mm
wire current = 18.70 A
wire resistance (24*C) = 12.30 Ohm
wire length = 26.63 m
wire Surface Load = 2.57 W/cm^2
 ### Coil ###
coil diameter = 15.00 mm
min. coil length = 3.39 m
max. recommended coil length = 6.78 m

And obtain a nice SL of 2.57 W/cm^2.
It would need 2 more levels of grooves in the top bricks to fit 5m coil.

... but 2mm Kanthal is expensive and I have 16A current limit, so no  ::)

Will get back to you guys soon.
 

Offline CaptainNomihodai

  • Contributor
  • Posts: 37
  • Country: us
Re: Building electric furnace to melt aluminium
« Reply #29 on: April 17, 2017, 02:08:29 pm »
@CaptainNomihodai:
That's interesting as I've seen guys build whole crucibles out of welded steel :D
But I did listen to your advise and went with the in-brick grooves.

Using a steel crucible is fine, in fact I usually use schedule 40 pipe welded to some plate, as long as you're not too worried about the precise chemistry of your melt (i.e. you're okay with a little bit of iron in it). My point was that you can't have steel touching the heating elements or the brick, because that's when this (see image) happens.
 
The following users thanked this post: ^_^

Offline ^_^

  • Regular Contributor
  • *
  • Posts: 82
  • Country: au
  • EE
Re: Building electric furnace to melt aluminium
« Reply #30 on: June 25, 2017, 01:14:20 pm »
Hello again after a long break :)

I've ordered stuff for PID from ebay (PID controller, K-type thermocouple rated to 1300 *C and 2x 40A SSRs).
They took almost whole May to get shipped from China.
Lately I've had some time to put it together and it's done!

The furnace still needs little work: putting the electronics/cables in boxes as well as the proper lid.
The lid will be made from 2 firebricks cemented together, firewool in the opening of the foundry and some welded steel elements for the holder and keeping it together.
With the proper lid I can make a comparison between calculated and real time to melt alu.

Thank you all again for contributing in this topic, your posts were of great insight  ;D

Here's a photo:
 

Offline SeanB

  • Super Contributor
  • ***
  • Posts: 15434
  • Country: za
Re: Building electric furnace to melt aluminium
« Reply #31 on: June 25, 2017, 04:38:24 pm »
Would not cement the 2 bricks together for the lid, but make a steel frame for them that holds them, and a steel handle to lift it with a lifter. Cement will fail in this use, it is too brittle with heat and stress.
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 16035
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Building electric furnace to melt aluminium
« Reply #32 on: June 25, 2017, 06:53:47 pm »
Presumably it's furnace cement, which is strong and refractory.

The bricks themselves don't look all that abrasion-resistant though, and for that reason, a frame would be a good idea.  They could also maybe be faced with cement too, if it doesn't cause expansion cracking problems.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 
The following users thanked this post: ^_^

Offline ^_^

  • Regular Contributor
  • *
  • Posts: 82
  • Country: au
  • EE
Re: Building electric furnace to melt aluminium
« Reply #33 on: June 29, 2017, 09:23:10 am »
Presumably it's furnace cement, which is strong and refractory.

The bricks themselves don't look all that abrasion-resistant though, and for that reason, a frame would be a good idea.  They could also maybe be faced with cement too, if it doesn't cause expansion cracking problems.

Tim

You're double-right  ;)
The cement is actually most heat-resistant element used in the furnace as it's rated to 1700*C (the firebricks are 1260 *C).

And yes, the firebricks are very, very NOT-abrasion-resistant. Putting them somewhere leaves tiny firebrick-dust. They need to be handled very carefully.

That is why the lid as a moving part needs special care. Hence the firewool as it will take care of the brick-to-brick friction problem.
I was thinking about steel frame as well to make it even more robust as the steel would hold everything and take on all the force.
The firewool in between would provide the thermal isolation.

Cheers!
 

Offline ^_^

  • Regular Contributor
  • *
  • Posts: 82
  • Country: au
  • EE
Re: Building electric furnace to melt aluminium
« Reply #34 on: September 09, 2017, 02:17:27 am »
Just before moving away from Europe, I managed to do first electric furnace cast.
Well, technically, electric furnace was used to heat the cast in a proper temperature profile.

I've used "Kerr Cast 2000" in 5 hour profile, which was in the end around 6h profile.
Cast temperature was around 400*C when I put the form out of the furnace.

The form cracked halfway through the temperature profile, but it did not affect the process. Maybe a form this big would need the longest, 12h profile.
I've also experienced problems to remove the wax, as I needed to physically remove the form, pour the melted wax out, and put the form back again in the furnace.

Though, it DID work, and I'm happy with this first time result.
Aluminium was melted in the old, coal powered foundry.

Pictures below.

Wax fist used to make a form (back in 12.2016 ::) ):
https://github.com/stawiski/electric-furnace/blob/master/photos/P_20161209_155233.jpg
https://github.com/stawiski/electric-furnace/blob/master/photos/P_20161209_155252.jpg
https://github.com/stawiski/electric-furnace/blob/master/photos/P_20161209_155338.jpg

Casting result:
https://github.com/stawiski/electric-furnace/blob/master/photos/IMG_2642.JPG
https://github.com/stawiski/electric-furnace/blob/master/photos/IMG_2647.JPG
https://github.com/stawiski/electric-furnace/blob/master/photos/IMG_2656.JPG
 

Offline negativ3

  • Regular Contributor
  • *
  • Posts: 121
  • Country: th
Re: Building electric furnace to melt aluminium
« Reply #35 on: September 16, 2017, 09:33:32 am »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf