Pages:
1
2
3 |
watson.fawkes
International Hazard
Posts: 2793
Registered: 16-8-2008
Member Is Offline
Mood: No Mood
|
|
As a pointer for anybody who's serious about building a good arc furnace, I can recommend reading up on the power supplies and, particularly, the
control circuitry for EDM, electro-discharge machining. An EDM supply is pulsed, and an arc is continuous (usually), but the important similarity is
that there's a servo control loop that keeps the EDM electrode at a constant distance to the work. By analogy, the same kind of control can keep the
electrodes at constant distance from each other, automatically adjusting for wear and allowing a fairly good measure of temperature regulation.
|
|
densest
Hazard to Others
Posts: 359
Registered: 1-10-2005
Location: in the lehr
Member Is Offline
Mood: slowly warming to strain point
|
|
@watson.fawkes: good advice... though perhaps the topic of power supplies has come up often enough to warrant an FAQ for experimentalists not totally
immersed in power electronics? I'd be glad to contribute a few paragraphs.
As an addendum to what watson.fawkes said, old electric arc lighting used mostly mechanical ways to gradually feed the carbon rods inward as they
burned off. At a first approximation, once the arc is established and the system is stabilized at a particular temperature, feeding inwards an
experimentally determined number of millimeters an hour would do a lot to keep an arc running.
EDM arcs are interesting since they're under fluid. I'm only passingly familiar with them; I do know that they operate as a "short arc" meaning that
direct ionization from the intense field at the moving electrode keeps the plasma going, and the fluid is in motion to wash away the material removed
from the work piece. I'm less sure that a carbon arc, which could be 100 times longer, behaves exactly the same.
An electronic control system for temperature stabilizing an arc furnace could be interesting.... 'fuzzy logic' or proportional-integral-derivative
temp controllers with integral thermocouple conversion aren't -too- expensive ($US 100 or less for a cheap one new, less for a used one). Proportional
electronic power control at KW levels can get more expensive: a new proportional voltage-controlled "dimmer" rated at 20A costs $US 75 or so.
The interesting part would be extra logic to detect arc failure, running out of carbon rod, and other out-of-the-linear-region conditions. Or watch it
all the time and have a big OFF switch. What I don't know is whether or not arc length increases or decreases power output
over a wide region of length.
A purist's approach might be to use two control loops: the "primary" one regulates current into the arc sensing temperature while another one
attempted to keep the length stable sensing voltage, since voltage across the arc changes with length and current. It is complicated because many arcs
have effectively negative resistance: the voltage across them decreases with increasing current, at least over a large region of
operation. I believe that voltage across an arc at atmospheric pressure is roughly proportional to length since it must continually ionize new gas
molecules to replace ones leaving.
|
|
watson.fawkes
International Hazard
Posts: 2793
Registered: 16-8-2008
Member Is Offline
Mood: No Mood
|
|
Arc discharge is full of non-linearities and various kinds of parameter dependencies. In my opinion, it's a problem for purpose-built software. Now I
do a lot of software, so this doesn't seem as daunting as it might to others, but when you have a system that's full of, well, behavior like an arc is
it's time to step back and think about what's important, which is not killing yourself, and this is mostly about thinking about the edge conditions:
start-up, shut-down, faults, and emergencies. I've thought some about building an arc furnace, though it's not on top of my priority list. Then again,
I really have to care. I have 277/480 V 3-phase-wye power in my shop on 50 A circuits. When something goes seriously wrong with that, it's 70 kW
before the breaker even thinks about tripping. If you don't want to make something this big, you may not need to concern yourself with the things I
do.Top of my list is built-in ground-fault sensing, both a shutdown circuit in hardware for faults and, for smaller faults, a
software monitor on the sensors to keep an eye on leakage current, say, from failing insulation or corona discharge. This subsystem includes current
sense on the electrode feeds, because a fault there is potentially more dangerous that a fault on the mains.
A crowbar circuit for failure, so that when it shuts off, it stays off. Were I to build a DC arc furnace, I'd be tempted to feed the mains
through a mechanical contactor to the arc supply section, wired so that start up requires a mechanical button press and that the crowbar shuts it off.
Current-limiting engineering, including picking a saturable-core transformer, fusing, output current sense (feeding the crowbar), perhaps an
output ballast coil, etc.
A high-frequency arc start generator. This one's not as obvious. Striking an arc directly is just chock full of transients, high inrush
currents, brand new non-linearities, and generally a whole lot of extra things to go wrong. It's a whole lot more predictable.
After all this packaging, I'd use, internally, generic kinds of switching power supply blocks: a DC power source, a switch matrix with isolated
triggers, sensors, passive output network, and a central microprocessor controller. To this add servos and sensors for the carbon rod feeding
subsystem. If you're building something like this, it's a project that will be a constant subject of tinkering, as applications change. So it's a good
thing to build it with a relatively good amount of reconfigurability.
|
|
12AX7
Post Harlot
Posts: 4803
Registered: 8-3-2005
Location: oscillating
Member Is Offline
Mood: informative
|
|
Why does that need a microcontroller? It'll sit around twiddling its thumbs. All those features are already covered by the very hardware you listed.
Maybe you can put in some digital panel meters (DPMs), but that hardly deserves a uC when off-the-shelf units are available.
Tim
|
|
densest
Hazard to Others
Posts: 359
Registered: 1-10-2005
Location: in the lehr
Member Is Offline
Mood: slowly warming to strain point
|
|
For a system as complex as watson.fawkes suggests (and it sounds like a good idea to me), an industrial process controller should work well. I'm
thinking of one of the uC controlled relay logic emulators which can do computation as well. I/O is their business, so they come with lots of inputs
and outputs which is what's necessary. Sequencing startup and shutdown, for instance, would be much easier with one. It would be very nice to have a
simple "on" button, a "temp" knob, and a big "off" button, and after that a way to set a temperature profile over time. A high frequency ignition
scheme could be lifted from a commercial welder.
Don't forget the exhaust fan for a 70KW arc! CO, NOx, O3, metal vapors, evaporated firebrick, .....
Stepping back a little further, what would an arc furnace be able to do that (for instance) an induction furnace couldn't do? That's a lot cleaner,
heats more selectively, and by choosing susceptors one can heat in very precisely controlled regions. Obviously, something which needs 10,000C from
the arc itself is probably not a candidate. The arc is a lot more dramatic, too.
|
|
watson.fawkes
International Hazard
Posts: 2793
Registered: 16-8-2008
Member Is Offline
Mood: No Mood
|
|
Quote: Originally posted by 12AX7 | Why does that need a microcontroller? It'll sit around twiddling its thumbs. All those features are already covered by the very hardware you listed.
| The safety gear I spent time talking about would, indeed, be outside the microcontroller's reach, and on
purpose, since you don't want your fail-safes to fail unsafely because of software defects. I didn't talk nearly so much about what I'd put in a
microprocessor for. What I was imagining the microcontroller for was for holding parameters of the arc model, largely to keep away from phase change
points, such as arc dropout and sudden conductivity drops. As such it would be controlling the electrode servo system and setting the operational
points of the power supply. And, for certain experiments, it provides a ready place for remote-control operation, if you don't want to be standing
next to the gear during the run.
|
|
watson.fawkes
International Hazard
Posts: 2793
Registered: 16-8-2008
Member Is Offline
Mood: No Mood
|
|
Quote: Originally posted by densest | For a system as complex as watson.fawkes suggests (and it sounds like a good idea to me), an industrial process controller should work well.[...]It
would be very nice to have a simple "on" button, a "temp" knob, and a big "off" button, and after that a way to set a temperature profile over time.
| Personally, I'll probably use an Arduino, since they're quite handy and the software development environment
very clean. If you are more comfortable with a process controller, I'm sure it would work too.
It's true I'm thinking about a complex system. The utility of such a device is exactly that the external controls are simpler, as you point out. In my
thinking, there are two kinds of reasonable ways to do this. The first is the kind that you pay attention to full time while it's in operation. These
require less internal complexity at the cost of requiring manual control. But you shouldn't walk away from one while it's operating, because you are
the safety device. The second is the one as I described, which is capable of taking care of itself to a great extent, and you can walk away from it.
Both, of course, should have a big red emergency stop button you can slam down on. Quote: | Don't forget the exhaust fan for a 70KW arc! CO, NOx, O3, metal vapors, evaporated firebrick, ..... | Oh, yes.
I was just addressing the power supply. Quote: | Stepping back a little further, what would an arc furnace be able to do that (for instance) an induction furnace couldn't do? [...] Obviously,
something which needs 10,000C from the arc itself is probably not a candidate. The arc is a lot more dramatic, too. | You have it exactly. Some processes need the higher temperature of an arc.
And as long as we're stepping back, I should point out that there's a lot of overlap between the supply as I've outlined it for an arc and that for a
induction furnace. In particular, the safety-specific features are almost all the same. Then, if you've got a solid DC stage, you can put different
switch topologies on it for different applications. If you're designing for hacking, you can make each individual switch unit (the switch device such
as a MOSFET or IGBT, gate trigger isolation, gate bias supply, and optional sensor for switch failure) a separate module, allowing reconfiguration by
bolting to new bus bars. Finally, you load in new control software.
|
|
12AX7
Post Harlot
Posts: 4803
Registered: 8-3-2005
Location: oscillating
Member Is Offline
Mood: informative
|
|
You still don't need a controller, arcs are not nearly as complex as you imagine them to be. Constant current supply and you're set. Electrode
spacing can even be done mechanically, but if you must have an electric servo, match it to arc voltage. Nothing more mysterious than that.
What a microcontroller would be useful for is reducing parts count (if you have one or a few controllers and PWM systems, you can nix a lots of chips
and resistors this way) and adding far more complex features, like something above and beyond PID (possibly higher order time constants and nonlinear
terms -- assuming, that is, that you've mapped the model quite thoroughly enough to be confident that you can actually write the correct inverse
transfer function), temperature profiles, digital I/O, etc.
Microcontrollers are useful, but please don't fool yourself on what they are used for. They are often good for reducing parts count. They aren't
good for reducing price, and they have no effect on development time whatsoever.
Note also that microcontrollers are sensitive equipment and not exactly robust around power electronics of ameteur construction.
Tim
|
|
IrC
International Hazard
Posts: 2710
Registered: 7-3-2005
Location: Eureka
Member Is Offline
Mood: Discovering
|
|
It is well known that the impedance of a typical burning arc is around 0.02 to 0.05 ohms. If we say the impedance of the heater coil is around 19 ohms
then the bulk of the energy is being wasted in the heater coil. Assuming 15 amps in a series circuit the power in an arc of say 0.035 ohms is just a
few watts, meaning you are wasting large amounts of energy.
A neon transformer has a built in shunt which limits short circuit current to 30 or 50 milliamps. Using one constantly will boil the potting tar and
burn it out. Far better is the idea of using a transformer from an oil burner furnace, where the transformer is designed to operate in this type of
setup.
|
|
bquirky
Hazard to Others
Posts: 316
Registered: 22-10-2008
Location: Perth Western Australia
Member Is Offline
Mood: No Mood
|
|
I remember reading about alot of high power carbon arc electric lighting before the lightglobe was around
One of the ways that was reguarly used to maintain a good arc was by mechanicly coupling an electro magnet to the carbon electrodes and then running
the windings in serise.
The idea was that when the current passing through the arc dropped the magnetic feild weekened and allowed the spring loaded electrode to advance a
little more.
There where all kinds of variations on this them incuding electromagnetic rachets that cranked the arc rods acording to current variations.
There was evan one scheem called a Yablochkov candle that had the two electrodes running parralell to each other like this l l the space in
between was filled with (i think) plaster. the idea being that the plaster whould crumble away at about the same rate as the carbon rods did.
I have evan made a miniature arc lamp by busting open a relay and gluing a graphite pencil lead to the moving part and wiring the relay in series
with the two electrodes.
Carbon arc lamps i believe are still used in some parts of the world in cinema projectors using that kind of system.
But if you want an arc furnace... how about doing something with a ballasted microwave oven transformer. or a cheap arc welder in an argon atmosphere
perhaps using cooled tugston electrodes ?
|
|
12AX7
Post Harlot
Posts: 4803
Registered: 8-3-2005
Location: oscillating
Member Is Offline
Mood: informative
|
|
Quote: Originally posted by IrC | It is well known that the impedance of a typical burning arc is around 0.02 to 0.05 ohms. If we say the impedance of the heater coil is around 19 ohms
then the bulk of the energy is being wasted in the heater coil. Assuming 15 amps in a series circuit the power in an arc of say 0.035 ohms is just a
few watts, meaning you are wasting large amounts of energy.
|
No, that can't be right, and at best you must be confusing incremental with average resistance. A short arc at 15A might drop 20V (a typical welding
voltage), which is 20/15 = 1.3 ohms average. It might be 30V at 200A, which suggests 0.15 ohms average and (30 - 20) / (200 - 15) = 0.05 ohms
incremental.
Regardless, an arc is nonohmic, so it's useless to rate one in terms of ohms. I can burn a handsome thermal arc from a flyback transformer, that
might be dissipating 20W at 20kV and 1mA (apparently 200kohms average resistance). Or I can burn steel at 20V, 100A (2kW and 0.2 ohms). Or I can
leave it to the power company to occasionally produce some quite spectacular accidents in the 500kV, 10kA range -- hundreds of megawatts, and still
only a few ohms.
Tim
|
|
watson.fawkes
International Hazard
Posts: 2793
Registered: 16-8-2008
Member Is Offline
Mood: No Mood
|
|
I should mention two more safety items that are not complicated, and which I should have listed before:Emergency stop button.
You know the big red kind, preferably one that requires some second motion to reset.A power-on lamp, possibly on a pole. Just so that you know
it's on, and more importantly, so that other people do.
For those interested in understanding how fault-sense circuits work, first read up on residual-current devices. Then to understand why the picture in that article applies to alternating current, see Ampere's law and electromagnetic induction. Note that if you're planning on a DC arc, you'll need a different technique, either using sense resistors and op amps,
or perhaps the Hall effect.
|
|
12AX7
Post Harlot
Posts: 4803
Registered: 8-3-2005
Location: oscillating
Member Is Offline
Mood: informative
|
|
Wow, a brief, informative post by Fawkes. I believe that's a first!
FYI, they're called GFCI receptacles over here.
|
|
densest
Hazard to Others
Posts: 359
Registered: 1-10-2005
Location: in the lehr
Member Is Offline
Mood: slowly warming to strain point
|
|
@12AX7, @watson.fawkes: one could build an AC leakage breaker, but they're commercially available for a wide range of power levels and leakage levels
and show up on the surplus market for reasonable prices.
It would be easy enough to make a DC leakage sensor with an old laminated
core, a hacksaw, and a few turns of heavy bifilar wire. Hmmm... quick calculation suggests that you'd want the gap as small as possible to admit the
thinnest possible (.5mm?) sensor. 100 turns @ 10mA difference @ .5mm = 25 gauss, enough to be sensed with an analog hall effect sensor but it would
need amplification. It would be well out of the noise and earth's magnetic field.
I don't believe that commercially available arc welding apparatus (AC or DC) come equipped with such a sensor. The arc circuit is not referenced to
earth and the voltage is low enough that leakage is not a safety issue.
I still think that for anything up to a 5 or 10KW arc or so a welding supply with perhaps modified controls and run at a level where it is capable of
continuous operation would be by far the most cost effective solution and easiest to get working. Constant current, constant voltage, limits, high
frequency arc ignition, all are ready made. The electrode feed would be the major area for implementation, I think.
For the truly insane, take a 20-100HP (15-75KW) VFD (variable frequency motor drive), reverse engineer the I/O on the CPU, and replace the firmware
chip. AC, DC, up to 7 or 15 KHz ignition drive... use a couple of very high
current chokes on the output if necessary to give the CPU enough time to recognize shorts, perhaps. Total outlay (on EBay) of about $500. Again, the
only hardware construction necessary is the electrode feed. The VFD will have uncommitted digital and analog I/O sufficient for (say) a set of
photocell arc length and status monitors, an electrode drive motor, etc. If reverse engineering the whole thing is too much trouble, it's quite likely
that with a little study the existing CPU can be cut away and a new unit of one's choice installed instead using the IGBT drives, current sensors, and
general purpose I/O already existing.
Anything to avoid creating the -whole thing- from scratch if it's not necessary and/or not fun.
|
|
watson.fawkes
International Hazard
Posts: 2793
Registered: 16-8-2008
Member Is Offline
Mood: No Mood
|
|
Quote: Originally posted by densest | one could build an AC leakage breaker, but they're commercially available for a wide range of power levels and leakage levels and show up on the
surplus market for reasonable prices. | This is true, if you want a separate module for leakage all by itself.
And that's not a bad idea at all, really. On the other hand, these are very easy to make, particularly if you pick up a current transformer on the
surplus market. I have several at this point, which is occasionally amusing, since the sizes aren't always apparent from the photos, and I've got one
that has capacity way larger than I would ever need (but which I'll use anyway).
I have two reasons for making them myself. The first is that I want two of them in a power supply. One is for the device as a whole, right as the
mains enter the cabinet. The other is for the high-power supply separate from the other circuitry in the cabinet. The difference, for me, is that the
outer unit passes all four conductors of the three-phase wye circuit through the inductor, and that the inner unit passes only the inner three wires.
The inner unit enforces a design criterion on the power supply that it draw balanced phase currents. Thus the outer unit detects earth faults, where
net current is leaving the device as a whole. The reaction to this fault should be device shutdown. The inner unit detects a more general class of
leakages, including, particularly, neutral currents (that is, imbalanced phase draw) and phase-to-ground faults (as opposed to phase-to-neutral
current, which is ordinary). The reaction to these faults need not be immediate shutdown, since they're not posing any immediate life hazard. In such
a fault condition, having a parameter logging facility in the controller is useful for forensics.
The second reason is that I want to have an interruption system that has more inputs than just that for the leakage current. This is really for the
inner interrupter. One is a deadman switch so that if the control circuit loses voltage, it shuts off the main power supply for the arc. Another is an
exceptional stop switch, which shuts off arc power without shutting down control circuits, which is useful if you do parameter logging. (There's also
an emergency shutdown switch that cuts all power.) There's also one for leakage current on the inner interrupter, using a second, higher threshold
than the one that only yields a fault indicator.
So, in summary, if I find an adequate fault interrupter cheap enough, I'll certainly use it for the outer unit.
|
|
watson.fawkes
International Hazard
Posts: 2793
Registered: 16-8-2008
Member Is Offline
Mood: No Mood
|
|
Quote: Originally posted by densest | It would be easy enough to make a DC leakage sensor with an old laminated
core, a hacksaw, and a few turns of heavy bifilar wire. | This would be one way to do it. Another would be to
use the Hall effect directly. Take a strip of soft iron sheet metal; this is an adequately low resistivity and high enough permeability to capture
most of the field from a conductor. Bind the two DC conductors to either side of this piece of metal, passing the conductors in opposite directions
from the feed so that their induced fields should cancel. (You could also use small H-channel for this.) Now pass a bias current across the narrow
edge of the strip; perhaps you braze on some copper strips to distribute the current. Then measure the induced EMF between the two ends of the strip.
This can be made pretty sensitive since you can mode-lock to a modulated bias current. Quote: | I still think that for anything up to a 5 or 10KW arc or so a welding supply with perhaps modified controls and run at a level where it is capable of
continuous operation would be by far the most cost effective solution and easiest to get working. | I'm in
agreement. The old big-iron versions of these are ideal for this kind of hacking, since they're cheap and the modules inside are easy to get to and
rewire.
|
|
densest
Hazard to Others
Posts: 359
Registered: 1-10-2005
Location: in the lehr
Member Is Offline
Mood: slowly warming to strain point
|
|
Quote: Originally posted by watson.fawkes | Quote: Originally posted by densest | It would be easy enough to make a DC leakage sensor with an old laminated
core, a hacksaw, and a few turns of heavy bifilar wire. | This would be one way to do it. Another would be to
use the Hall effect directly. ... Then measure the induced EMF between the two ends of the strip. This can be made pretty sensitive since you can
mode-lock to a modulated bias current. |
I'd have to disagree and go for something like the Honeywell linear current sensors ($10-$75) with Hall effect (3ms) or giant magnetorestrictive
(0.5us !) semiconductor elements. The magnetic properties of iron change with stress and temperature history sufficiently to make getting a good
sensor pretty difficult. Getting truly balanced magnetic properties while folding & bending sheet iron is not a project I would like to undertake.
I've already lost too much hair doing amateur magnetics!
Besides, the Hall effect in metals is very much smaller (1000x or more) than in semiconductors and the GMR sensors are even better. The Honeywell
units come with the magnetics integrated and a fully temperature compensated and linearized output. It's hard to beat that.
Now, if you want to learn all about grain and residual stress and impurity inclusion distribution, go right ahead. Electronic
Goldmine is selling sheets of super high permeability magnetic material right now...
Alternatively, one could break open a GFCI of the appropriate size and cut the wires to the trip coil. I had three of various sizes but I threw them
away last year Various size sensors are up on EBay right now for what I consider
too much ($100-400). A deal might wander by... The brand name "LineGard" seems to be the most available.
One could also disassemble a single phase GFCI and run more wires through the sense transformer & put a SSR or optocoupler on the trip output -
total cost about $30 (one or two tries learning how to open the thing without destroying the parts you want) and a couple of hours messing around.
|
|
watson.fawkes
International Hazard
Posts: 2793
Registered: 16-8-2008
Member Is Offline
Mood: No Mood
|
|
Quote: Originally posted by densest | I'd have to disagree and go for something like the Honeywell linear current sensors ($10-$75) with Hall effect (3ms) or giant magnetorestrictive
(0.5us !) semiconductor elements. [...] Now, if you want to learn all about grain and residual stress and impurity inclusion
distribution, go right ahead. Electronic Goldmine is selling sheets of super high permeability magnetic material right now. | Good information. I hadn't heard about the giant magnetostrictive materials until your post. (I assume these are the same as gaint magnetorestrictive materials, no?)
In any case, there's no need for a very fast reaction times in the sensor and some reason to avoid it. There's always some capacitance in the load
device and this leads to current imbalances when the current changes. If you're using a pulsed DC arc, for example, each arc-start and arc-stop
operation would definitely show off this capacitance. On the other hand, if you can detect "steady-state" operation, say, by getting a signal from
your controller, then you could switch on a fast sensor when it was appropriate. Regardless of this, capacitance limits the sensitivity of any kind of
leakage device. If building your own, it's a good idea to put a trimmable lower limit on it, so as to avoid nuisance trips.
Another consideration for a coil-wound DC sensor is the inductance. This wouldn't matter in a continuous setting, but it might in a repetitive pulse
one. I haven't done any estimates about how significant this might be.
I bought a sheet of that mu-metal from EG, but not for this. Hmm.
|
|
densest
Hazard to Others
Posts: 359
Registered: 1-10-2005
Location: in the lehr
Member Is Offline
Mood: slowly warming to strain point
|
|
My bad. Giant magnetoresistive not magnetostrictive. Both GMR, just entirely different mechanisms... though the side trail through
MRAM and other applied quantum physics objects was interesting to look up (4Mbit 35ns static MRAMs for about $30 each - what could one build?)
And what might one make in a high power arc furnace? A high power resistive furnace is used to make CaC2, a route to acetylene and with another, lower
temperature, step, to dicyanodiamide, two interesting raw materials.
|
|
UnintentionalChaos
International Hazard
Posts: 1454
Registered: 9-12-2006
Location: Mars
Member Is Offline
Mood: Nucleophilic
|
|
Quote: Originally posted by densest | My bad. Giant magnetoresistive not magnetostrictive. Both GMR, just entirely different mechanisms... though the side trail through
MRAM and other applied quantum physics objects was interesting to look up (4Mbit 35ns static MRAMs for about $30 each - what could one build?)
And what might one make in a high power arc furnace? A high power resistive furnace is used to make CaC2, a route to acetylene and with another, lower
temperature, step, to dicyanodiamide, two interesting raw materials. |
We have what should be a very nice, clean preparation of calcium cyanamide already, via the pyrolysis of calcium cyanurate:
https://www.sciencemadness.org/talk/viewthread.php?tid=2762
http://www.sciencemadness.org/talk/viewthread.php?tid=8594#p...
Department of Redundancy Department - Now with paperwork!
'In organic synthesis, we call decomposition products "crap", however this is not a IUPAC approved nomenclature.' -Nicodem
|
|
watson.fawkes
International Hazard
Posts: 2793
Registered: 16-8-2008
Member Is Offline
Mood: No Mood
|
|
Birkeland-Eyde with a pulsed arc
Why funny you should ask. I've been thinking about
pulsed power because it's occurred to me that you might be able to improve the efficiency of an arc for burning nitrogen (for the Birkeland-Eyde
process) by running a pulsed arc rather than a continuous one. The principle is to increase the surface-area--time product of the arc. Reactions
happen at the surface of an arc, not its interior. They don't happen in the arc itself, since that's plasma and is too thermally energetic to form
stable bonds. It's at the surface of contact between the arc and its atmosphere that reactivity is high, as ions (free radicals) leak out of the
plasma and react with the neutral atoms nearby. So by periodically quenching the arc you liberate the ions inside for reactions.
Now this whole process, in order to work well, seems to require that you put an electric field transverse to the arc, providing a certain amount of
separation of the ions in the plasma before they recombine with each other. So now you've got a high voltage bias supply to deal with and all its
attendant risks, plus the possibility of arcing to your bias plates. So this is a little challenging.
Practically speaking, you'd really like to have a residual ion channel between the electrodes, which means a high-frequency arc stabilizer (not just
arc start) that can keep an ion path open, even if little current flows through it when the main arc is off. This reduces the initial overpotential
required to reestablish the main arc at the start of a pulse.
[Edited on 4-10-2009 by watson.fawkes]
|
|
watson.fawkes
International Hazard
Posts: 2793
Registered: 16-8-2008
Member Is Offline
Mood: No Mood
|
|
Ah. Not both GMR, but GMR and GMM, which was the abbreviation I was always seeing for "giant magnetostrictive materials".
|
|
Panache
International Hazard
Posts: 1290
Registered: 18-10-2007
Member Is Offline
Mood: Instead of being my deliverance, she had a resemblance to a Kat named Frankenstein
|
|
Funnily after reading this thread i stumbled upon this, do you think any publisher would allow such advice in a trade text/manual these days. I
especially like how he uses a gray lead pencil, you have a built insulator with the wood.
It comes from R.H. Wights manual on lab glass blowing
Attachment: Pt Welding.tiff (174kB) This file has been downloaded 1043 times
|
|
12AX7
Post Harlot
Posts: 4803
Registered: 8-3-2005
Location: oscillating
Member Is Offline
Mood: informative
|
|
Yup, that'll work fine. If you're good enough at it, you don't even need the smoked glass, just close your eyes, tap and it's done.
I prefer burning the pencil lead first, because it may explode from wax impregnated into it. Use a soft pencil, because the higher graphite content
makes them more conductive. The clay and graphite mixture vaporizes somewhat, producing a conductive plasma and reducing atmosphere better than plain
graphite.
Tim
|
|
metalresearcher
National Hazard
Posts: 758
Registered: 7-9-2010
Member Is Offline
Mood: Reactive
|
|
Nice try !
But only direct on the mains voltage is DANGEROUS !!
I have done the same but with a $100 hardware-store-around-the-corner welder which yields 40V with much more Amps.
http://www.metallab.net/arcmelt/
I can boil Aluminum, melt steel and melt MgO easily.
|
|
Pages:
1
2
3 |
|