Sciencemadness Discussion Board

Carbon arc Furnace

lvjrf - 13-10-2004 at 23:56

Hi All
I known The Theory Of Carbon arc for lighting welding

but are thereany body show me how made electric furnace (carbon arc)
to melting metal ?


Oxydro - 14-10-2004 at 17:23

Try this site for a little bit of info, it was the first thing I found.... bond with google for more information


lvjrf - 17-10-2004 at 03:21

thank Oxydro

Oxydro - 17-10-2004 at 06:25

No problem, it's an interest of mine :).

There was also an article in (IIRC) PopSci by Theodore Gray (guy who collects elements, owner of "The periodic table table";) on making a simple arc furnace for melting various metals, but it no longer seems to be available online.

There is also a section in "How to make and use a small chemical laboratory" about making an electric arc furnace. If you can't get that book or file (it's on the ftp), let me know and I'll post pics of the relevent pages on my website.

I always wanted to build one of these, but I never got around to getting the proper eye protection -- I don't want to risk my eyes, they're bad enough already. If you try it, be sure to keep us informed on how it goes.

Edit: too

[Edited on 18-10-2004 by Oxydro]

ignaro - 20-10-2004 at 08:46


I was thinking (and procrastinating) about to build an electric furnace also.

The most obvious way was (from my viewpoint) to use an arc welder: Low voltage, easy current regulation, isolated secondary, etc.

I was reading some of the Oxydro links, an I saw that they use a direct connection to distribution lines (shock hazards) and resistors (energy waste) or reactances as current limiters.

Is there some cause (other than low cost of equipment) for such methods?

Twospoons - 20-10-2004 at 13:58

Simplicity. Robustness.

I like it!

Oxydro - 20-10-2004 at 19:26

Oh, an arc welder would be wonderful... if I happened to have one. Although it's probably the best current source, you do have to go out and buy the welder, if you don't already have one... a tub of salt water as a resistor is a wee bit easier on the wallet.

What might be the best alternative is to rewire one or more microwave oven transformers to provide low voltage high current.

vulture - 21-10-2004 at 02:02


What might be the best alternative is to rewire one or more microwave oven transformers to provide low voltage high current.

Dumb question: How are you going to sustain an arc of reasonable current with low voltage? Resistance of air is something like 10kOhm/cm IIRC.

Organikum - 21-10-2004 at 03:06

The easiest way is a salt-water rheostat. A description how to build one and an arc-furnace was already posted here by me, do a search for "DIY".

There is a electrical misconception here too, you dont need a transformer but a ballast for an arc furnace. thats not the same.

[Edited on 21-10-2004 by Organikum]

chemoleo - 21-10-2004 at 04:14

Vulture, an arc of low voltage/high current is absolutely fine. I got my arc to work with alternating DC current, 42 V and 120 Amps. The reason being? As soon as the electrodes touch, you create superheated carbon, which evaporates (and burns). The ionised gasses conduct the current.
Nonetheless of course, the higher the voltage, the further apart the electrodes can be. So with very low voltages (i.e. 10 V) you get an arc that is extremely short - so short that the electrodes virtually touch. Therefore it's harder to keep the electrodes at the correct distance, i.e. where they don't touch each other, but also where the arc can sustain itself.
I remember with my 42 V arc, every dozens of seconds or so I had to move up the electrodes. I am sure this wouldn't be quite as frequent had I used a higher voltage.

[Edited on 21-10-2004 by chemoleo]

Organikum - 21-10-2004 at 04:30

70V to 120V is the preferred voltage for arc furnaces, DC is preferred over AC.

A salt-water rheostat with an rectifier after the rheostat works just fine.

For small furnaces is understood.

ignaro - 21-10-2004 at 08:24


What might be the best alternative is to rewire one or more microwave oven transformers to provide low voltage high current.

A search for "arc welder" microwave brings a lot of pages.

I like this:

ignaro - 21-10-2004 at 09:06


70V to 120V is the preferred voltage for arc furnaces, DC is preferred over AC.
A salt-water rheostat with an rectifier after the rheostat works just fine.

By my (only) experience, a rectifier can be made using the lowest-cost-per-amper-diodes, conected in paralell.

I read that such thing must not be done, because there are not two equal diodes, then a diode starts conduction before the others diodes, it conducts the entire current and it dies in a Jeanne d'Arc way.
Maybe my diodes never read such argument, because they worked. I used the rectifier in a very old DC motor (220 V, 0.25 HP).
They worked a couple of years (4-5 hours per week) then the motor jammed and expelled some smoke; then, the diodes were sorrow because the motor death, or they tried to emulate their coworker, and they died also.

Oxydro - 21-10-2004 at 15:49

I think, with the parallel diodes, in practice the resistance of the diodes and their leads creates enough of a resistance to overcome the slight differences in the forward voltage drop of the diodes.... if that makes any sense -- just my best guess.

So parallel arrays *can* certainly screw up, it's just that real life situations usually have balancing factors.

Edit: Ignaro, that is in fact where I found the idea of a rewired transformer... More accurately first on

Regarding Chemoleo's comments on arc length:
I think that the voltage doesn't have much effect on the ultimate length of the arc. Only the current affects the maximum length (of course you have to have a high enough voltage to overcome resistance, but the resistance of an arc is tiny).

What the voltage affects, however, is the maximum gap the arc will start at. With air's breakdown voltage of approximately 1100v /mm, at voltages under a few hundred you essentially have to touch the electrodes together. That is how the arc is stated in an arc welder, by eithertapping or scraping the electrode briefly on the surface of the work. Tapping for DC and a scraping motion (I don't know how to describe it better) for AC, as far as I recall.

[Edited on 21-10-2004 by Oxydro]

Twospoons - 21-10-2004 at 17:15

Parallel diode arrays work better if the diodes are in close thermal contact. As diodes heat up their forward voltage drops, so if one is hotter than the others it takes more current and so gets hotter, and you get thermal runaway leading to destruction.

uber luminal - 21-10-2004 at 21:13

We worked on an arc furnace about a year ago in lab. Used a 220 stick welder and copperclad carbon arc gouging rods. we made it work, but it was far more trouble than it was worth. problems:
-The arc is hot. 10,000 F. hot. everything melted, or burned.
-Even with the current turned down considerably (if you turn it down too much like below 100 amps, you lose your arc.) the carbon rods would burn up quickly, and would have to be ajdusted frequently. On the up side, it didnt take long for anything in the middle to melt. problem, the stuff in the mold would melt whereever the arc was, so you have to heat it longer... and this deformed the mold.

you have to keep the mold under whatever it is you are melting, becuase otherwise the load cools too quickly to cast it. Also, We tried using a graphite crucible in doing so, we destroyed the graphite.
-Every fume coming out of there is toxic. You are ionizing our air which contains Nitrogen and Oxygen. Needs to be well ventilated. (where as normal welding operation is either done with an inert gas, or hydrogen gas(for stick welding) and thus doesnt need nearly the same precausions that you need when arcing regular air)

-Did I mention it was hot? Kevlar gloves and a #11 lense on your helmet. trouble is you have to stand there, and watch your puddle, and so you have to be looking at it.
-Oh yea, your refractory or whatever you build it out of will experience thermal shock, decompose, or both.

Good things:
-Able to make a casting of mildsteel (thats pretty cool, since you typcaly cast ... cast iron, not mild steel)
-Tried to cast stainless... destroyed it.
-Was able to melt #7 Tungsten electrodes (pure)
-Was able to melt 2% thoriated tungsten electrodes (what a waste)
-Aluminum... was meant to be funny, but it turned to dust/white smoke instead of melting like we thought it would.
-castings were crude, and had silica from the melted sand stuck to the outside in the gaps, but you could generaly see their shapes.

Now im not saying all arc-furnaces are not practical, just ones made from welders. If you looking to cast high carbon iron, copper, pewter or aluminum, just stick with a gas burning furnace. takes a little longer, but I think its far less of a hazard and a little more fun.

ignaro - 22-10-2004 at 08:04


Ignaro, that is in fact where I found the idea of a rewired transformer... More accurately first on

Yes, I saw it. I preferred the other design because its quick-and-dirtyness, very similar to my creations. I have some kind of neatness disability :)


Parallel diode arrays work better if the diodes are in close thermal contact. As diodes heat up their forward voltage drops, so if one is hotter than the others it takes more current and so gets hotter, and you get thermal runaway leading to destruction.

Maybe the best method is to bound parallel diodes with a cooper sheet or similar, I think that I'll do next time, if there is a next time.
For info only, I don't want to disagree: according to this criterium, I made the rectifier in the worst way.
I used 20 or 24 diodes, I dont remember it. Four copper wires nailed perpendicular to a wood piece. Wires length and distance nearly 5 cm long. I soldered diodes trying to maintain them apart, some of them were near the wood.

Uber luminal:
It was a rod to rod arc, or a rod to metal one?
The rod to metal arc transfers a little more energy from arc to metal, and the current in the metal develops a little heath in it... maybe the sum of the two little quantities can be helpful, at least while the starting, if the metal is a set of pieces: the contact resistance can increase the heat development in metal.
I read, I dont remember where, that primitive laboratory arc furnaces were made from CaO.

An (maybe crazy) alternative to the resistor for current control

A half value resistor and a single diode (or parallel diodes equivalent to a single diode), then we obtain the same average current, and a capacitor parallel to arc, for to maintain the arc in the no conduction intervals. Less energy waste.

rikkitikkitavi - 22-10-2004 at 22:31

sorry ignaro, but the capacitor has to be a BIG one. And you will not achieve current regulation either. Only way to do this is by putting an IMPENDANCE in series with the circuit,causing a small voltage drop, depening on current.
This impendance can be made by a resistor (high power loss, hot, bulky) or a inductor
( bulky, heavy, possibly expensive if purchased, difficult to design if homemade from old transformer) or a capacitor (not realistic).

This said, many transformers used for welding are a special design with a magnetic shunt in the core, which is adjustable. This way the transformer itself is current limited somewhat (A MOT sometimes has a magnetic shunt built in, aswell as neon light transformers)

Using half wave rectifying:

Ripple voltage over a capacitor is
U = I *20/C for your halv wave rectifier , for a full wave it is I*10/C

where I = current (mA) , C capacitance (uF)

if 20 A is drawn, and ripple voltage < 5V (about 10 % of arc voltage) gives a capacitance of 80000 !!!uF . That is double the size of a coke can! Also, the ripple currents charging and discharging such capacitor will heat it extensivly , so it has to be first grade (expensive, about 50-100 US$) for a voltage rating of about 63V. (use rating = double working voltage)

This said, it can be realized by using multiple parrallell connected caps of smaller values, that can be found through various surplus for cheap. This way the wiring is important so that the current is shared equally through all caps, since this type of surplus caps usually are of less sturdy quality.


[Edited on 23-10-2004 by rikkitikkitavi]

ignaro - 25-10-2004 at 06:38

80000 uF? That is fine!
I have a lot of ceramic capacitors, I can get more.
Voltage rating is low, but I can do serial connections (equal capacity in each serial group), and afterward connect groups in parallel.
Soldering cost could be a little high, I guess; and space consumed a little high also :D

Maybe a short arc (i.e. low voltage) could be more stable for a half wave rectifier if using thick electrodes and high current: The high temperature of electrodes point could help the arc restart, by heating the air in the gap.
But high current excludes the use of direct connection to mains...

And I'll never try to make an arc with the mains voltage. That position comes from local environment conditions: We have 220-380 volts here, I prefer other forms of suicide :). Better is to use transformers.

More Seriously: I am reading "The Electic Furnange" (Sic). From ;
Really, it is The Electric Furnace by Alfred Stansfield, there are some typos in the books list.
better is to download the file TheElecticFurnange_orig.tif (148.8M), not TheElecticFurnange.tif (67.3M) The second one is easy to download but hard to read.
It is a good book, date is 1914. It don't refer material and devices by brand names, it use substance names, or describe the devices.

iron core laminations

Magpie - 5-12-2005 at 19:49

I'm planning on building a current limiting inductor and start doing some carbon arc testing using 120VAC/60Hz as input power. I'm shooting for a current draw of around 10 amps with the electrodes touching.

The subject here could equally be called "lamentations." I'm trying to find a source of sheet metal or strips of the silicon iron used for iron cores for transformers, inductors, etc, and I'm not having any luck. :(

Because I can't find a source I thought I'd build my own laminations. But you still have to find out what kind of sheet metal to buy and how thick. I found a very old book today that said the thickness is usually 14 mils (0.014"). Material is "stovepipe iron" or "Russia iron." My guess is that this is just silicon iron.

If I can't find a source for the right material I will likely just buy some thin sheet steel and cut out my own strips. I will then coat them with insulating varnish.

If anyone can tell me where to get these laminations I will be grateful.

Twospoons - 5-12-2005 at 20:18

You need to talk to a transformer manufacturer. The material is grain oriented silicon iron, and the manufacturer will have it already cut into E and I sections. Better still is to ask them to wind the inductor for you - which they will do, with the right size core,wire, insulation for your job. If you don't want to do that you can scavenge the laminations from an old, dead, or surplus tranny. Cutting from raw sheet stock would be a LOT of work. I assume you don't have a CNC Turret punch?

Magpie - 5-12-2005 at 20:42

Thanks twospoons.

I understand what you are saying about making the laminations being a lot of work. I only have a jigsaw. I won't be using E and I configuration. I'll just wrap an "I" piece. This is going to be fairly crude.

I do have a phone number for a transformer manufacturer. I'll give him a call. (I also read that the grain was oriented and the metal annealed.) If I make my own I'm assuming that the thinner the metal the better (within strength limitations).

Where would I find an old transformer? A TV? A microwave oven? City landfills don't allow much salvaging anymore. I don't have a TV or microwave that needs destruction right now. ;)

Am I being too fussy? I'm not trying to optimize this design. Overkill by 200% won't break my budget. Could I just get by with say 0.031" (22 ga) steel as long as the core doesn't heat up too much? I'm planning on winding the core with insulated solid wire #12 (AWG). Building it myself is most of the fun.

[Edited on 6-12-2005 by Magpie]

Mr. Wizard - 5-12-2005 at 22:25

Your best bet, as far as easily available materials would be the transformer from a dead microwave oven. They are built to handle 900 to 1000 watts (look on the ratings sticker). This guy shows how it's done:
The microwaves are a dime a dozen at junkyards, swap meets, next to the trash cans on trash day. The transformers are rarely bad, if they are it's usually the high voltage side, which you will remove anyway. What out for those High Voltage capacitors, they can bite, even after the thing is unplugged.

Edit to add: Google microwave transformer welder to get additional ideas and info. The first site mentioned had a great page, but he's selling his plans now and isn't as detailed as he used to be.

[Edited on 6-12-2005 by Mr. Wizard]

Magpie - 6-12-2005 at 09:48

Thanks for the good info Mr. Wizard. It looks like I just need to be a little more aggressive in my scavenging.

Magpie - 6-12-2005 at 19:51

With a bit of dumb luck I picked up a 1600w microwave oven today at a junk yard for $5. I don't think they usually have them. It has an E-I type core transformer which should work out nicely for my inductor. :D

Fleaker - 7-12-2005 at 16:45

A search for "arc welder" microwave brings a lot of pages.

I like this: "

Actually, the person who built that welder is a rather frequent contributor to this forum, perhaps you should U2U 12AX7 who designed that unit. Also, I suggest you email Theodore Gray, he's a very nice person and a fellow science enthusiast; I'm sure he'd be glad to supply some details.

Making an arc furnace is a loosing endeavor without the proper refractory materials. Just using 1200C rated high insulating kiln brick, and after one use you will notice the brick has fused together on the inside from the radiant heat. Tungsten or graphite electrodes can be used, regardless, they will volatize and conduct the arc (gradually disappearing as they do). It's the radiant heat that melts the metal in most cases, not direct contact and the metal to be melted is usually in some sort of cooled basin (copper typically IIRC) in an inert, reduced pressure atmosphere. I suggest you look into magnesia or zircon as refractory to reflect the heat and keep from melting so much as a more simple mullite refractory would.

neutrino - 7-12-2005 at 17:32

High vacuum melting on a piece of cooled copper? Are you sure you're not confusing this with e-beam melting?

I've never seen a DIY carbon arc furnace with an inert atmosphere and reduced pressure, although I guess I haven't looked that much. Still, it seems beyond the ameteur's capabilities.

Fleaker - 7-12-2005 at 17:40

Not entirely beyond the amateur's capabilities. My friend, nearly anything is possible with money. Check out Max Whitby's reduced pressure Ar, arc furnace.

I might have it confused with e-beam melting (although I doubt it, e-beam melting is the epitome of refining, it is after all an upgrading step and they would not introduce anything that could be dissolved, even soluble copper) I can always e-mail a few people and ask what type of basin they use.

Mr. Wizard - 7-12-2005 at 20:18

Magpie, Funny how it happens when you are looking for something you see them all around you :) A 1600 watt transformer sounds like a good one. Don't forget to save the diode(s), fan, capacitors, and the magnetron. The magnetron has a big magnet in it. The thing about MOTs (MicroWave Oven Transformers) is they are wound for power, not for efficiency, as I understand it. Most transformers used on the mains won't draw much current when their secondaries are not loaded, but MOTs don't generate enough counter EMF (voltage) in their primaries to resist the line voltage, and thus draw a significant current even when not loaded. You might want to keep that fan to cool it down.

Magpie - 7-12-2005 at 21:32

I have turned my kitchen into an electrical development laboratory, assuring my wife it will only be for a few days. :D It's much warmer than the garage.

I removed the transformer from the microwave oven. Then removed its secondary winding. It was an over-the-oven type so scavenged 2 pilot lights (30w) also, the electrical cord, a little wiring, a large capacitor, and a dual squirrel cage fan. Thanks for the tip Mr Wizard, I'll go out to my garbage can and get the magnetron and diodes too.

I assembled the parts on a piece of 5/8" plywood and began testing. First test was with a 75w incandescent lamp as load just to verify the circuit. Light was dim which was a good sign. Then loaded it with a 1600w space heater. Amp draw was 5.5 amps vs a possible 13.3 without the inductor. So far things are looking good. I will keep testing with loads of smaller resistance until I feel I can test a short circuit load. If the short circuit test yields less than 10 amps I feel the power supply should be ready. I will post a picture of the set-up in my next post.

I understand that this arc furnace will be very hot. It will make my muffle furnace, which is limited to about 1200C, look like a wimp. I am looking for appropriate refractory materials. So far the best I have found is 2800C firebrick and 3200C castable. I appreciate any leads on more appropriate refractories.

I plan to use ~1/4" graphite electrodes and a graphite crucible of about 200 mL capacity. I realize that there will be consumption of these items, especially the electrodes.

[Edited on 8-12-2005 by Magpie]

[Edited on 8-12-2005 by Magpie]

power supply test set-up

Magpie - 8-12-2005 at 07:30

Reduced pic size. Chemoleo

[Edited on 28-11-2008 by chemoleo]

inductor.jpg - 98kBinductor with load.JPG - 468kB

chemoleo - 8-12-2005 at 09:14

A while ago, I ran 42 V @ 120 Amps through two 0.8 mm thick carbon electrodes, where the electrodes touched in a chamber of a 4 cm diameter, made of cast firecement (you know the cement that is used to stick together a fireplace). I couldnt use this setup properly because after a few seconds only the firecement melted into a white glowing mass. I was able to make some Calcium carbide that way, within a few few seconds run. At the time I realised making a proper setup would be difficult, you need a remotely controlled setup to move the carbon electrodes forward, it's quite dangerous to be that close to the arc. Then, a decent distance from the arc to the chamber is better, to avoid the problems described above.
The biggest problem, as I see it, are the electrodes, and their required steady movement. I'd use alumina-based fire brick/plates which you can buy from hardware suppliers, and which are used to line ovens with. On my coal furnace, they glowed a happy white but did not crack. All you need to do is to cut them into the right sized pieces, but and hold a square setup together with an outer layer of fire cement. Drill a hole into the brick for the electrode, and devise some setup to move the electrodes smoothly. Then you are all ready to go!

Did anyone ever dig out how a professional carbon arc furnace is built?

S.C. Wack - 8-12-2005 at 11:22


Attachment: jacs_20_769_1898.pdf (259kB)
This file has been downloaded 1128 times

12AX7 - 8-12-2005 at 12:35

So you're using the MOT primary as a ballast inductor?

Cut off the welds holding the "I" segment on, then add a layer of thick plastic and add a C-clamp to hold it in place. The air gap will reduce inductance. Varying gap (practical limit is about a half inch, beyond which you might as well just have the "I" seperate) varies inductance.

Measure the voltage across the space heater next time you run it. R = V/I gives a figure for its resistance. Kirchoff's rule says 120 - VS = VL, and currents being equal you can calculate phase angle (and power factor) and reactance. Given a constant supply frequency, you can thus establish the inductance of the inductor. Then graph L vs. gap for reference. :D


12AX7 - 8-12-2005 at 12:39

Originally posted by chemoleo
Did anyone ever dig out how a professional carbon arc furnace is built?

Hum, I've walked past a small one before. Three 1" (or so) dia. graphite rods dropped into a refractory chamber. This was a direct-melting furnace, three-phase (delta) powered obviously, probably for melting cast iron, steel and stainless steel (that's what the foundry pours).

Large direct-melting furnaces are the same idea, just bigger rods (a few *feet* in diameter..) and, I'd hate to think what the ballast inductors look like. :P

Indirect-arc furnaces usually use a rotating chamber, IIRC, so the radiation heats the walls and the charge makes contact with as much of the heated surface as possible.

Limestone, dolomite and magnesite (magnesium carbonate, or make some yourself by precipitation) are suitable half-assed refractory materials. They really do need to be fired to around 2000°C first but if you tolerate the mess you'll make of them, you can use it raw.


neutrino - 8-12-2005 at 17:29

>...All you need to do is to cut them into the right sized pieces, but and hold a square setup together with an outer layer of fire cement. Drill a hole into the brick for the electrode...

How hard is it do machine aluminum oxide? I know that this material is very hard (~9), so I imagine it wouldn't be so easy.

chemoleo - 8-12-2005 at 18:15

Well the bricks I purchased are in fact quite crumbly, and break easily. Yet, using an angle cutter (is that the name?) it DID abuse the rotating disk a fair bit.
But it's possible.

more testing

Magpie - 8-12-2005 at 19:38

I had more fun today with my test set-up. To get the final 1.1 ohms resistive load I had to place every heating device I could get my hands on into a parallel array. It is a good thing my wife was away at work. What is beautiful to a madscientist can be overwhelming to most others. :D

My method was to measure the resistance load, R, cold with an ohmmeter. Then turn on the power and measure the resulting amp draw, I. The impedance, Z, was then calculated as Z= E/I, with E being 120v. The inductive reactance, XL, was then calculated as XL = [(Z^2)-(R^2)]^0.5. Inductance, L, was calculated from L= XL/(2*pi*f) with f= 60Hz.

The data are as follows. In each case the current value was measured quickly before the resistance items had time to heat up (no lamps were involved).

R, ohms-------I, amps-------- L, henrys (calculated)
1.1------------10.0------------- 0.032

Plotting R vs L yields a fairly straight line. Extrapolating this line I estimate my short circuit current will be 13.8 amps. This may or may not blow the microwave fuse I have in the circuit. My guess is that it won't as to meet the nameplate 1600w it would have to draw 13.3 amps.

I find it interesting that the inductance, L, drops off with increasing current. I didn't expect this. Maybe it is due to "core saturation?" Perhaps somone who is better versed in this phenomenon can tell me.

[Edited on 9-12-2005 by Magpie]

[Edited on 9-12-2005 by Magpie]

12AX7 - 9-12-2005 at 08:39

It certainly could be. Saturation usually approaches suddenly, though. I would expect something like "50mH ... 49mH ... 47 ... 44 ... 25 ... 10mH", where L starts dropping significantly only above a certain value.

For low currents, L will also be lower, due to initial permeability. Starting at 5A you should be well out of that region though.

If you have the equipment, it's a lot easier to observe the properties of an inductor by pulsing a constant voltage across it. It starts to draw current according to the equation V/L = dI/dt. V and L being constant initially, it increases at a constant rate. In nice materials (last night, I was testing ferrite cored inductors), it stays constant until it saturates, at which point current curves up and starts rising faster. In iron cores, it starts low, goes up pretty smoothly, and then saturates.

Since you then have energy stored in the inductor, you have to release it. Most times you use a diode to direct the energy into a capacitor, where the power is burned off by a resistor inbetween cycles (since you'd be operating this continuously to get a repeating, viewable waveform, energy pulses over time equals power).

Last night, I was testing a ferrite core to see if I can get enough inductance and current capacity. You can increase saturation by adding air gap. I went from about 250AT to 750AT with about 3/32" gap (I measure 2.5mm), reducing inductance from 0.25 to 0.1 microhenries per turn squared.


Magpie - 9-12-2005 at 16:37

Thanks 12AX7 for the insight on core saturation and permeability. That is interesting stuff that I hope to take the time to understand better someday. I don't have much for electrical test equipment other that a Heathkit volt-ohm meter and a RadioSchack clamp-on ammeter. Someday I hope to pick up a cheap oscilloscope on eBay.

Also thanks to S. C. Wack for posting the old reference. Without these old references (I have 3 now) I wouldn't know where to start in building a home arc furnace. Yours will be especially valuable when I get to "furnace design."

I obtained two more test points today, i.e.:

R, ohms---------I, amps-----------L, henrys (calculated)

I now feel that the power supply testing is complete and satisfactory.

I purchased the hardware needed for an arc test stand today. I will use 5/16" carbon gouging rod as electrodes. I also purchased a welder's helmet ($20) so that I can look at the arc. :cool:

S.C. Wack - 10-12-2005 at 21:01

2 like articles from JCE - 9, 148 (1932) and 14, 492 (1937):

Magpie - 12-12-2005 at 14:57

Did my first arc testing today with the 5/16" (8mm) graphite gouging rods. I did the first test outside (T=0C). I couldn't get a sustainable arc, even with a minimal gap. So I brought everything inside, let it warm up, then tried it again. This time I was able to get a sustainable arc using a minimal gap. The current was 13.7 amp. Then I sharpened the rods with a pencil sharpener to about a 3mm diameter point. I seemed to get a better looking arc here but it was not sustainable. These results were disappointing as I expected a decent sized, sustainable arc, using a gap at least as wide as 3mm. Please critque this experiment as you see fit.

Perhaps this does not indicate failure for my ultimate purpose which is to heat various carbon containing dry powders to temperatures in the 1600C - 3000C range in a properly designed furnace. I suppose it will depend on the conductivity of the powder and how well the furnace is insulated.

Shown below is a picture of the arc test stand:

[Edited on 28-11-2008 by chemoleo]

arc.jpg - 130kBarc test stand.JPG - 483kB

chemoleo - 12-12-2005 at 15:15

Very nice, I suppose you move the electrodes forward by threading it through the wood?

Also, I don't know what voltage you used, but 13.4 Amp sounds like WAY too little. The arc I described above had 120 Amps at 42 V. I am not suggesting you need that much, but 13.4 Amps does not sound sufficient to me.
Also, is that guy in the link above not using 6 or eight microwave transformers (modified) in parallel?

Attached are the designs on an image electric furnace - they use mirrors to focus the light. That way they melt ZrO2, MgO and Cr2O3 :o

Slightly more sophisticated. However it might give you some design ideas.

[Edited on 12-12-2005 by chemoleo]

Attachment: carbon arc furnace.pdf (864kB)
This file has been downloaded 1342 times

Magpie - 12-12-2005 at 15:54


I don't know if it is amps I'm looking for or just power. Power I would think.

You delivered (120a)(42v)=5040w! That's a lot of power. If I'm drawing 13.7amps this means my gap resistance is next to 0 according to my earlier power supply testing data. With a gap resistance of next to zero I'm not producing much power. It is a very weak arc. For me to draw 5040w using 120VAC line voltage I would have to install a 50 amp breaker and use #6 wire. I would have to beef up the whole system for that kind of current.

What kind of power supply and electrodes were you using for your production of calcium carbide?

I haven't been able to access those last S.C. Wack references as I don't have access to RapidShare, even on the free basis. I must be on a proxy server? I can get reprints on an inter-library loan, however.

Yes, my safety equipment is visible in the picture (welder's helmet & leather gloves). I have to change the gap while the system is "live" by pulling or pushing one electrode and with helmet on. I do this with gloves on and a wooden handled bar-b-que tong. Crude but effective.

I first thought I would mount a nut on the wood and turn the "all-thread" through that for exact gap setting. Then I realized that not only was that not necessary for this preliminary testing but is tough to do with a wire attached. :D

chemoleo - 12-12-2005 at 16:15

Indeed, I do think you need several kW, at high Amps. No, high voltage does not equal high amps here, you want the number of electrons, not the 'speed' of them, as it were!
The arc is made by ionised air/carbon/CO/CO2, not by voltage bridging!

My powerlines were very thick, deliberately, to minimise resistance. As in about 10 cables of the type you showed in your pictures, twisted together - so as to make sure the most resistance is encountered where the electrodes meet. The setup to me looks like you are rather heating the wires than the electrode.
I used the same electrodes as you, about 20 cm long/8 mm thick and Cu coated for better conductance. THe supply - well it was an industrial transformer that I scavenged at a company where I once worked. It weighs roughly 20 kg - very heavy. Still it became quite warm during my treatment - I suppose I should have put in some current limiters. Another problem of the supply was that it gave an impulse onto the mains when switched off, which resulted in the house fuse being switched too - so every operation needed the house fuse to be turned on again. I think I discussed that elsewhere, can't remember if we solved the problem or not.

Apart from that, once you got a better powersupply (there is unfortunately no way around it I think), I really suggest you do this in some sort of crucible. Once it gets going you'll see what I mean - it is awesomely hot and dangerous, you want it at least partially covered!And I can almost guarantee you, you will be too scared to move the carbon electrodes by threading them through, you'll be much too close to the arc.

I uploaded SCWacks references (which are good, similar in fact to the design I did ) to megaupload. I know the trouble with rapidshare, I also have it with certian files. And proxies hardly ever work for me :-/

PS I just noticed I wrote 0.8 mm in one of the posts above, it was meant to be 0.8 cm, or 8 mm of course! 0.8 mm could never hold that much power!

[Edited on 13-12-2005 by chemoleo]

Magpie - 13-12-2005 at 13:01

This is a very interesting problem. The more articles I read the less clear it is to me exactly how to proceed. But I'm going to keep reading.

I was bothered by the fact that I had no guidance on what the power should be for a desired temperature. So I did calculations for a simplified model using the following assumptions (my apologies for the barbaric units):

furnace geometry: spherical with radius, R2 = 6 inches
crucible spherical cavity radius, R1: 2 inches
cavity wall temperature, T1: 1600C
outside surface temperature, T2: 80C
insulating material conductivity, k: 2 BTU/(hr-ft-F)

design equation: Q= 4*pi*k(T2-T1)[1/(1/R2-1/R1)]

With these assumptions, Q = 5037w

This is remarkably close to what chemoleo had used.

For a 3000C furnace the power required would be 9345w.

Note: I realize that a basketball sized furnace would have a tough time radiating 5200w with a surface temperature of 40C.

Edit 1: I have just discovered the mother lode of electric furnace information: ignaro's 10/25/04 reference, i.e., "The Electric Furnace," 1914, by Alfred Stansfield of McGill
University. This reference has everything including power requirements.

Edit 2: Corrected design equation and used a more realistic surface temperature of 80C and conductivity.
[Edited on 14-12-2005 by Magpie]

[Edited on 14-12-2005 by Magpie]

Magpie - 16-12-2005 at 11:22

With a cursory read of Stanfield I believe I now see how to construct an electric furnace for both production of calcium carbide and other metals via carbon reduction.

In summary, this is what I've learned:

1. It is power that is required to reach the desired temperature.

2. A minimal voltage of 40VAC is required to sustain an arc in air.

3. Because of the very low resistance of the arc, or carbonaceous mineral matrix, it is necessary to use very high currents to get the necessary power delivery. [If you play around with the equations P = (I^2)R and P = (E^2)/R you'll understand.]

4. An AC stick welder, specifically the Lincoln AC-225C, looks like the ideal power supply from the standpoint of control, current capacity, and cost. It is capable of supplying about 10kw at currents up to 225 amps. (My household mains circuit breaker is rated at 22kw.)

chemoleo - 16-12-2005 at 11:50

Yes I fully agree, this is why I put the development of my own furnace on ice, due to control issues.
I'd really just get an arc welder, the cheap ones aren't even that expensive (soemthing like 80 euros they used to be). You'll have everytthing in a box, and then can finally make some beloved CaC2!

Believe me, once you take your sample and you drop a few solid chunks (tiny!) into water, and bubbles that burn with a pop, you'll be wondering why you bothered with MOTs and all - when it can be done so easily!

The difficult issue will be the crucible design, and steady movement of the electrodes....

Dodoman - 18-1-2006 at 01:44

I want to build my own vacuum arc furnace so I came up with a design but I need some opinions.

I think the diagram is very explanatory. I have a few doubts though. Will my cooling system work effectively ? Where and how do I place the second electrode so that I'm able to move them closer and even stir them if I need to? I think if the 2nd electrode was a right angle I could move it farther and closer to the first by rotating the rode. How do I seal the cover do I just use heavy grease and count on the vacuum to seal it? Any thoughts would be very helpful and much appreciated.

Arc furnace my design.jpg - 36kB

Magpie - 18-1-2006 at 08:34

Interesting design Dodoman.

1. Read the Stansfield book I reference above if you haven't already. It is a wealth of information and ideas.
2. Why do you need a vacuum? If this wasn't required you could just run in the inert gas as a purge without worrying about perfect seals.
3. Telescopic seals are a possibility for moving electrodes.
4. Can the crucible itself be the 2nd electrode? Some designs use this (see Stansfield).

Your design is close to what I've been contemplating. I have put this project aside until I can justify buying an AC stick welder for a power supply.

[Edited on 19-1-2006 by Magpie]

toxin - 3-5-2006 at 10:21

I don't trust using power directly from a wall outlet to produce a spark, is there anything wrong with poducing a spark for this furnace from something with a lower current yet high voltage like a stun gun powered from a salvaged computer power supply ?

12AX7 - 3-5-2006 at 11:38

It's not a spark, it's an arc. You want high current because well, high voltage sucks.

A stun gun especially. That has single pulses of limited energy. You need continuous power. Nothing a computer supply is capable of producing, either. More than four times that.


chromium - 3-5-2006 at 12:36

I once played with computer PSU connecting two graphite rods to 12V and gnd. This rail was rated to 8A at 12V and after some unsuccesfull atempts i was able to produce microscopic arc. I rubbed gently graphite rods with each other and sometimes there emerged small unstable arc. It was seen as very bright point of light and rods got hot almost immediately. Normally I managed to hold this arc 10 sec or less but in some cases it burned more than 1 min continously.

Such small and unstable arc has no practical value in my opinion but it was quite a fun to play with.

a_bab - 4-5-2006 at 09:04

I have a way of producing a low powered arc easy enough, with two 110/220v transformers. I'll put the plan soon.

I discovered it by mistake, and you can create and sustain an 5 mm to 1 cm arc for as long as you want. You don't need carbon electrods either; it'll form from metal electrodes aswell. Not good for welding though, due to the low power.

12AX7 - 4-5-2006 at 10:31

Oh yeah, heh, I was messing with my induction heater project one day, testing with a load resistor- on accident I loaded it across the power supply rails and got a handsome arc. This was +200VDC and a 50 ohm wirewound resistor, so around 4A flowing = 800W, until the resistor heats up and the value drifts anyway (depending on the tempco of the metal used for the resistor).

Beautiful arc, reasonably thick at 4A, good length about 1/8 or was it 1/4", hissing since it's a filtered DC supply, only popping when struck or broke. :D