Iron Ore Mining Magnetic separator project using electromagnets

[nsmorley]

I am working in an iron ore mine and we need to build a 3m long magnet x 5cm high x 5cm
variable from 500 - 10,000 gauss
x 30 rows (i.e. 30 identical EM's)

We need to know:-
the ideal core material, guage of electrical wire, number of turns/ layers to wind, current to pass

The ore is crushed 1mm to 5mm particles.

Well how did you decide that it should be 3 meters long and 5cm high and 5cm wide?
The current separator we have uses static 1000 gauss magnets, which we want to replace with EM's of the same overall size.
Also, what kind of current supply do you have (it takes a fair bit of current to energize an electromagnet)?
we have 1mw of power supply on site.

Also, why are the constraints so strict (500 to 10kG field strength), why that level?
the iron ore we are passing reacts best to these field strengths.

Also, how close to the surface of the magnet will the ore be placed, and how fast will it be moving?
6mm from the magnet to the ore, with a stainless steel rotating drum in between. This will rotate at 0.5-1m/s

 

[nsmorley]

mag separator pic

 

[moffy]

I assume you want DC for the magnets rather than alternating magnetic field. So you need high current low voltage supply. Mild steel is the best core material and since its DC it doesn't need laminations. The steel won't do a great deal of good as the air gap is large. You need to get a good reference on large rectangular inductors. There are various fudge factors used for various sizes. Once you have an inductance value for the given size you can evaluate the average flux density. Another way is to simply build a scaled inductor and measure its field strength vs current, then scale the results.

 

[MrAl]

View attachment 71556

Hello,

nsmorley:
I didnt know you already had magnets, but they are solid magnets not electromagnets right?
Also, they are 5cm by 5cm square cross section, by 3 meters long. So there are four sides 3 meters by 5cm each, and two sides 5cm by 5cm. So what are you calling the 'face' of the magnet, is one of the 3 meter by 5cm sides? (The face is the side that is used to attract the particles).
It's a little hard to tell what is what in the photo, maybe a few more angle shots would help and maybe label each part.

moffy:
What makes you think the steel (core) isnt going to do much good?

 

[nsmorley]

Hi,
yes they are solid magnets at the moment. approx 1000 gauss. we want to replace these with EMs of the same size which can fit into the same space in the frame/ drum. The current magnets are 20mm x 30mm x30mm. so there are 100 pcs lined up to make up a 3m span.
Can we wrap these with elecrtical wire and charge it to make the magnetic field stronger.
we need variable from 1000- 5000 gauss so please advise spec of cable and winding and elec current we would need?

many thanks for all your help!

Nick

 

[nsmorley]

**broken link removed**

for reference - video of the existing magnetic separators being built..

 

[MrAl]

Hello again,


Well when i ask you some questions you have to tell me the answer or else we wont be able to do anything

Questions you have to answer:
1. What side are you calling the 'face' of the magnet (the face is the side of the magnet that attracts the particles) ?
2. I think you said that you have a 1mw power available at the site, but did you really mean 1 Megawatt?
3. Is that power in #3 above intended for ALL the electromagnets or just one of them at a time?
4. About your power source, is it AC or DC and what voltage is it and what current can it put out (if you know)?

To answer your question about wrapping wire around the magnet to make it stronger...
The answer is it might, but you'd have to wrap a lot of wire to test it. It depends on the metal used, and usually magnet metal has different characteristics than steel made for transformers and electromagnets. For example, it may not be able to attain a high field strength due to saturation of the metal.

Also, can we see a picture of the magnet and the magnet where it is mounted in the machine itself? Vids didnt help here.

 

[moffy]

moffy:
What makes you think the steel (core) isnt going to do much good?
Because the air gap is large. You don't have the field going across a small air gap to more steel which is connected back to the coil. But in asking this question you show you don't have a basic knowledge of inductors. You really need an inductor/electromagnet specialist as there are lots of details that need to be correct. The construction is quite specialised.

 

[MrAl]

moffy:
What makes you think the steel (core) isnt going to do much good?

Because the air gap is large. You don't have the field going across a small air gap to more steel which is connected back to the coil. But in asking this question you show you don't have a basic knowledge of inductors. You really need an inductor/electromagnet specialist as there are lots of details that need to be correct. The construction is quite specialised.

Hello again,


I asked a simple question.

QUOTE:
"But in asking this question you show you don't have a basic knowledge of inductors."

I hope you dont mind if i chuckle for a while, like a long while here

I would hope i picked something up about inductors over my 30 or more years past in the field of custom power supplies including much design work and quite a bit of hands on work as well.

You happen to be right...right if this was an inductor, but since it's not an inductor what's the opposite of right?

It just so happens that the *total* inductance goes down as the air gap increases, you are absolutely right about that. That means that if we had to build an inductor we'd be in trouble here. But since we dont have to build an inductor we are doing better than that. It's only the distance from the face of the core to the object to be attracted that is of the most importance. That's because although the *total* inductance goes down, the inductance in *each element* may be higher, much higher. And if we choose the right metal we can get that up pretty darn high. That's why an electromagnet almost always has a metal core of some type or another, whether there is a small air gap or a huge air gap or as is often found an undefined air gap.

So i agree that the total inductance goes down, but that's not the most important aspect of the design. The more important is the inductance per circuit element where one element may be much, much higher than another.

I also agree however that if this is for a commercial business it would be wise to consult with an expert on electromagnets. We can still help here however if the budget is low, it will just take longer with some back and forth information gathering and even possibly some simple on site experiments.

 

[moffy]

It is an inductor. It generates a magnetic field in response to a current. Energy stored is L*I*I/2. It will have some series resistance due to the windings, but primarily inductance. V = L*di/dt = N*dPHI/dt. So inductance and flux are directly related. Flux is your measure of guass.

 

[dougy83]

What makes you think the steel (core) isnt going to do much good?

The permeability of soft iron may be as 50000 times that of a vacuum (ref https://en.wikipedia.org/wiki/Magnetic_core).
(ref https://en.wikipedia.org/wiki/Electromagnet) Where the 'L' values denote the length of the path through the core and air gap respectively. Therefore, So, unless the length of the core path is ~50000 times that of the air gap path length, it can be basically ignored from the equation, and just the path length of the air gap used. That said, adding soft iron to make up half the path length should roughly double the field strength. A core taking up 3/4 of the path length should roughly quadruple the field strength (w.r.t. a coreless solenoid).

Please excuse the small size of the inline equations... clicking on them makes them larger

EDIT: note that the second equation shows "D = NI ur / ..." whereas the ur should be u0

 

[MrAl]

It is an inductor. It generates a magnetic field in response to a current. Energy stored is L*I*I/2. It will have some series resistance due to the windings, but primarily inductance. V = L*di/dt = N*dPHI/dt. So inductance and flux are directly related. Flux is your measure of guass.

Hello again moffy and dougy,

moffy:
QUOTE: "Flux is your measure of Gauss"
Sorry but no, it's flux *density*. And the flux *density* is much higher in the steel than in the air. In the steel it is confined, in the air it spreads out. That's why the steel can do so much good and the air does so much harm.
But as long as you want to relate this to an inductor, you cant tell me that if you take an air core coil and measure the inductance with an without a 5000u core that the inductance isnt going to go up significantly, even if it is a straight coil that does not curve back to meet the other end.

moffy and dougy:
There's not much room for argument as i have direct experience with this in actual applications that measure the field strength. So i know for a fact that when you add steel you get a higher concentration of flux which in turn results in higher field strength near the ends of the steel. Without the steel the measurement may even be too low to be useful, but with the steel the measurement picks up significantly enough to be measured in a given application.

Also, you two must still be looking at the properties of a finished inductor. In a normal inductor with an air gap, the air gap reduces the total inductance measured across the leads electrically, but that measurement does not reflect what is happening inside the steel. The steel acts as a sort of 'lens' that concentrates the flux into a much smaller cross sectional area than can be possible in the air. So the flux is the same, but the flux density is much greater in the steel. So the air does in fact affect the *overall* inductance to a great extent, but it does not modify the actual properties of the steel itself in that you'll still see a higher measurement of Gausses at the end of the coil with the steel than without.

But you dont have to take my word for it, so a simple experiment. Get a linear hall effect device and wind a straight cylinder coil. Apply an appropriate current. Measure the field strength at one end of the coil using the hall device. Next, insert a 10 penny nail inside the coil, then repeat the measurement. Note how much it went up. For best results find some 20kG steel and try that and see how MUCH more it went up.
You can also measure the inductance if you like, but you'll have to watch the frequency and match it to the steel and it might not work too well with the nail depending on how much current you use.

Also one more small note:
We are still in the information gathering phase here anyway. What this is starting to look like is the magnet face will be one side of the 3 meter length. That means the front face will be 5cm x 3 meters, so the distance from the front of the magnet to the back of the magnet may only be 5cm.

Another quick question for the OP:
Have you tried rare earth magnets yet?

 

[dougy83]

But as long as you want to relate this to an inductor, you cant tell me that if you take an air core coil and measure the inductance with an without a 5000u core that the inductance isnt going to go up significantly, even if it is a straight coil that does not curve back to meet the other end.

Based on the equation, the permeability of the core is not as important (so long as it's large) as the proportion of the path that contains the core.

Also, you two must still be looking at the properties of a finished inductor.

Not particularly... I was just ignoring the fact that the path length of the air gap changes as the air is replaced by iron

To keep the shape of the flux lines similar to what is currently being used, perhaps the core should be a similar shape to the current magnets, e.g. not take up more than say 1/2 of the flux path... that should mean that 120 turns at 10A should give around 0.5T (5000 Gauss), no?

 

[MrAl]

Hello again dougy,


Yes it may not be that significant, but it has to go up at least by some degree.

But i am still more or less waiting for a reply to what side the face is considered to be. I am picturing a construction that looks like a giant guitar pickup
Funny as that sounds, if the face really is one of the 5cm by 3 meter side then the wire would have to wrap around the LENGTH of the 'core' (air or metal) not around the 5cm x 5cm cross section like a transformer. I have doubts as to how well this will work. I think i would go with rare earth and just vary the distance instead of the actual coil excitation. Maybe have to think about this some more, but one turn around that thing is going to be 3+3=6 meters plus a little more. And that's only one turn (6+ meters). Next comes finding the ideal wire size and that requires knowing a lot about what his power supply is going to look like. Or we could do a mock up with some random wire gauge and see what fits and what the effect at the 'ends' of the 'core' are.

120 turns at 10 amps gives 5000 Gauss...how did you calculate that?

So we have a ways to go yet on the information collection phase. Maybe this is just one of those things that is left to someone who can actually visit the site and make the appropriate judgments after more careful measurements and observations. At the rate we are going, it's going to take 2 months just to get all the info.

On a related note, here is an article on calculating the magnetic field of a point outside a wire:
https://www.electro-tech-online.com/threads/analog-to-digital-tlc548.591/

LATER:
A quick calculation shows that with a core material with initial permeability of around 200 (even less would work) would increase the effective permeability from 0.5 to close to 1.0, which is double. Not a lot, but still significant as this effectively halves the current in principle.

 

[dougy83]

120 turns at 10 amps gives 5000 Gauss...how did you calculate that?

I just used the formula from above of [LATEX]B=\frac{NI\mu_0}{L_{gap}}[/LATEX] rearranged to [LATEX]N=\frac{BL_{gap}}{I\mu_0}[/LATEX]. B is in Teslas, and is therefore 0.5T for 5000 Gauss. I is 10A. Lgap was set to 6mm... but should actually be closer to 6cm, so perhaps you'll need 15km of wire instead.. Ok, that's enough mistakes for me for one night.

 

[MrAl]

Hello again dougy,


Mistakes or not i am happy to read your input here as it has been very interesting and you've found some interesting information too.

I was wondering what time it was down under as here it's just around 5am in the morning (Eastern USA).

Oh before i forget again (he he) another interesting thought experiment...

Take a small bar magnet not too strong, see if it attracts a steel sewing pin say 2 or 3 inches away. If it does, move it away a little farther until it just doesnt pull it to the magnet. Then lets say it's at 3.1 inches now, so now take a stack of say 1/2 x 1/2 inch transformer laminations 3 inches long and stick the magnet on one end so the pole is facing the laminations, now place the magnet in the original position with the laminations pointing toward the pin. That means the laminations take up part of the space between the pin and the magnet now. Now see if the other end of the laminations attracts the pin weaker but sort of like the magnet did.
Then ask this question:
The pin did not move with the magnet alone, but the pin did move after the laminations were added (if it's done right of course). Why did it move after the laminations where added?
One possible answer: the laminations provided a lower reluctance to the magnetic path between the magnet and the pin because of their higher permeability than air.
Apply: This is similar to how the laminations can link the turns of the coil together, so that one end of the coil has more effect at the other end when the steel core is added inside the coil. Without the steel core, the first turn has the most effect while the last turn has the least effect. With the core in place, this is still the case but the difference is less because each turn links to the first turn better and the first turn is closest to the place where we need the force.

 

[dougy83]

We're 15 hours ahead of you, at GMT + 10, so it's 8pm here when it's 5am there.

Yes, interesting experiment indeed. The needle will be attracted more easily anyway on the second time even without the iron because the needle becomes slightly magnetised! But yes, I get the experiment, and it works as you suggest.

One possible answer: the laminations provided a lower reluctance to the magnetic path between the magnet and the pin because of their higher permeability than air

Sounds good to me. You can draw the field lines from N to S with both the short magnet and the long magnet (i.e. the one with the iron) - the short one should have a high density of flux lines on the short sides, without much leaking out into the rest of the world while the long magnet will not have the high density and will therefore have the lines extended out further away from the magnet ends... That's my take on it anyway. I had to write an electrostatic solver for some numerical programming course in uni, which could be used to show the fields here, but I misplaced the code.

 

[nsmorley]

Hi guys, looks like i should check in more often!
Yes the EM face dimensions are 3m long x 5cm wide so will need to wind it accordingly.
We have looked at rare earth magnets - neodymium in particular which is available in small 5cm x 5cm blocks at 5000 gauss
but they are $30 each and we would need 2000 to complete the system.
I was hoping an EM would be a bit cheaper and also the ability to vary the magnetic field intensity will be very useful to us in separating different grades of magnetic iron ore.

Would it be possible for you to let me know your calculated values for the original questions regarding the construction of the EM?

thanks again!
Nick

 

[shortbus=]

@nsmorley, just question about your operation. Is this separating being done wet(slurry) or dry?

 

[nsmorley]

dry separation. we are located in the atacama desert..