Advice on building a 1.4T electromagnet

[ephedyn]

I'm trying to build a C or H-frame electromagnet that produces a flux of 1.4T in the gap between its two poles. (See: http://cyclotron.lbl.gov/images/Cyclotron%20yoke.JPG ) But I can't find any references on building C/H-frame electromagnets.

I calculated with permittivity of free space 4piE-7, gap 5.08E-2 in metres that I'd need either 3000 turns with 21.5A, or 2000 turns with 32.2A.

Now, the problem is, I have absolutely no idea of how the yoke works. I only know that I need a C/H-frame. Correct me if I am wrong, do I simply wrap 1500 turns (running 21.5A)/1000 turns ( running 32.2A) on each cylindrical section nearest to the gap (refer to picture)? (1)

Secondly, I'd think it is extremely difficult to shape a block of iron like that, so I'm planning to build a laminated core. In which case, what exactly do I use to hold the iron sheets and insulator sheets of the laminated core together? (2) I'm thinking that an adhesive would not hold them together because of the overheating, but a mechanical joint is not feasible either as the core will go out of shape. Which brings me to, what type of material would you use for the insulator sheets anyway? (3)

Lastly, I'm not adverse to reading up and figuring these out myself; so I'd be extremely appreciative if someone could just point out any references. (4) Thanks in advance!

 

[Bob S]

I agree with your calculations on B field. Two important questions:
1) What is the shape/size of your pole tip? This determines the overall size of your magnet. Rectangular is easier to make.
2) Is your coil water cooled or air (conduction) cooled? If air cooled, how do you plan to get the heat out? What are you planning for your power supply? What gauge wire (and insulation) are you planning to use?

 

[ephedyn]

OK, good to know that my calculations are OK.

I agree with your calculations on B field. Two important questions:
1) What is the shape/size of your pole tip? This determines the overall size of your magnet. Rectangular is easier to make.

Hmm, I have in mind exactly what is on the picture above. A rectangular frame with two cylindrical poles inside the frame; hence circular pole tips, with diameter ~40.0cm. I made a quick drawing on powerpoint to illustrate (but it's horribly out of scale and I'm bad punctuating the lines to show the projection). But anyway, the top picture is the 2D face-on view, while the bottom is a sort of a badly drawn 3D projection view.

http://home.vs.moe.edu.sg/linl/untitled2.jpg

My plan was to get iron sheets, say 1/8" thick, each in the shape of the 2D face-on view above, and laminate them to 40cm thickness, then machine the cylindrical poles to shape. This way I won't have to produce 4 cuboids and 2 cylinders out of a ton of scrap metal and bolt the whole thing together, if you get what I mean from the poor description. The only problem I forsee is that I won't yield uniform flux lines between the poles.

Can I ask if the rest of the yoke's dimensions matter? I can't find anything on this.

2) Is your coil water cooled or air (conduction) cooled? If air cooled, how do you plan to get the heat out? What are you planning for your power supply? What gauge wire (and insulation) are you planning to use?

I intend to use 12 gauge (AWG), double-cotton covered copper wires. Hmm, yes it's air-cooled; I don't intend to keep the electromagnet running for long intervals (~1min after saturating) and hence don't think I'll implement any other further cooling method.

 

[berkeman]

http://cyclotron.lbl.gov/images/Cyclotron%20yoke.JPG

Holy mother of

How big are you planning on making your version? What are you going to use for X-ray shielding...?

 

[ephedyn]

Not that large! :shy:

As mentioned, the pole faces are ~40.0cm in diameter, so the yoke would be about 1.8m x 1.2m x 0.4m. The plane of acceleration has an external diameter of 35.914cm; but the electrodes are 'pseudo-semicircular' because of a diametrical gap, and have chord length of 35.502cm after correction.

I'm using a 1.0m x 1.0m, 4" thick 75% gold plate for shielding. But that's just a safety precaution... my device will not produce any radiation whatsoever.

No, I was kidding. Most of the x-rays will be in the 'hard x-ray' range. Some radiation will be distinctly energetic (17.1+/-0.1 MeV gamma photons) from the decay of beryllium-8. Compton effect is predominant in high energy photons, and the best recommendation is to be behind high density concrete. To be accurate, I'll be behind 12.901cm of concrete. Photoelectric effect is predominant in low energy x-rays and gamma radiation (<3 MeV from acceleration of protons), which are 'attenuated' by 1/4" of steel, though definitely insufficient. The protons and alpha-particles have a higher weighing factor, and are fully stopped by the steel, which is most important. Most of all, I'll be at half-length of a lab away, which more than quarters the otherwise integrated dose.

I learned from literature review that the integrated dose is low, and this is consistent with the low time of exposure and distance from the radiation source, which are both equally important (I cannot quantify 'effective', but some parallelism is shared between these two terms) as shielding in the control of exposure. So I have not planned additional measures.

Lastly, I can always improvise; in due honesty, I haven't put my mind to the shielding since the more important concern is to get the apparatus operational first.

 

[berkeman]

I'm using a 1.0m x 1.0m, 4" thick 75% gold plate for shielding. But that's just a safety precaution... my device will not produce any radiation whatsoever.

No, I was kidding.

Oh, whew! Had me stopping and re-reading those sentences, going WTH? :rofl:

Sounds like you have a good handle on what you are doing. Fun stuff.

 

[ephedyn]

Haha, glad you enjoyed the joke. I thought of it because my Firefox crashed exactly when I finished my post; and I was thinking that I should perhaps write some random description to act as a filler since I would hate to rewrite the post.

Thanks, glad to have your confidence.

Edit: It's strange to see the two 'glad's lined up like that. There must be some explanation why it looks so aesthetically unappealing.

 

[Bob S]

This is a serious magnet. I have woorked around particle accelerators for many years, and have built many magnets this size and larger. First, using laminated iron for this is a bad idea, because the permeability of the air gaps between laminations have a magnetic permeability about 5000 times lower (i.e., reluctance 5000 times higher) than the iron, and this alone will destroy the efficiency of the magnet (tesla per amp-turn). So think of solid iron for the return flux bars. My ARRL handbook shows 12 Ga conductor has about 9 amps rating. 8 Ga conductor has about 24 amps (@ 700 circular mils per amp). Your next job is to design two 1500-turn coils to fit around the two circular pole tips. These should be roughly square cross section, i.e, about 40 layers of 40 turns each. Once this is done, the flux return bars can be designed. You should also calculate the total IR drop and calculate the voltage needed for both coils in series. In the flux return bars, you want to keep B below 1 Tesla.
By the way, water-cooled magnet wire is rated at about 600 amps per square cm of copper; I have used square magnet wire that is 3/16" on a side.

 

[ephedyn]

This is a serious magnet. I have woorked around particle accelerators for many years, and have built many magnets this size and larger. First, using laminated iron for this is a bad idea, because the permeability of the air gaps between laminations have a magnetic permeability about 5000 times lower (i.e., reluctance 5000 times higher) than the iron, and this alone will destroy the efficiency of the magnet (tesla per amp-turn). So think of solid iron for the return flux bars.

I see, that's enlightening! In that case, I apologize if for being inexperienced, but how did you usually procure such a large quantity of iron and shape it? I have to try to minimize the project cost, quite unfortunately.

My ARRL handbook shows 12 Ga conductor has about 9 amps rating. 8 Ga conductor has about 24 amps (@ 700 circular mils per amp). Your next job is to design two 1500-turn coils to fit around the two circular pole tips. These should be roughly square cross section, i.e, about 40 layers of 40 turns each. Once this is done, the flux return bars can be designed. You should also calculate the total IR drop and calculate the voltage needed for both coils in series. In the flux return bars, you want to keep B below 1 Tesla.
By the way, water-cooled magnet wire is rated at about 600 amps per square cm of copper; I have used square magnet wire that is 3/16" on a side.

OK, it looks like I'll work on these over the weekend. If you don't mind, can I consult you again via PM or on this thread (whichever you prefer) once I've met another problem?
Edit: I'll use the square magnet wire suggested to make things easy for the mounting.

Thanks!

 

[Bob S]

I see, that's enlightening! In that case, I apologize if for being inexperienced, but how did you usually procure such a large quantity of iron and shape it? I have to try to minimize the project cost, quite unfortunately.!

Fortunately, all I had to do is specify the magnets I wanted, and mechanical engineers and draftsmen designed them. Someone else paid for everything.

OK, it looks like I'll work on these over the weekend. If you don't mind, can I consult you again via PM or on this thread (whichever you prefer) once I've met another problem?
Edit: I'll use the square magnet wire suggested to make things easy for the mounting.

Two things on water cooled copper conductor. 1)If a conductor is too long, the pressure needed to force water through is too high. You need to understand Reynolds number and turbulence in narrow tubes. 2) If the water gets to hot in the tube, it will start boiling, expand to gas, and plug the tube.

 

[Bob S]

You might look into this Walker HF-12H magnet on eBay:
http://cgi.ebay.com/Walker-HF-12H-Research-Lab-Electromagnet-STEAL_W0QQitemZ270374662971QQcmdZViewItemQQptZLH_DefaultDomain_0?hash=item3ef395f33b&_trksid=p3286.m20.l1116
It has only a 30 cm diameter pole tip, but otherwise it meets your requirements.