Various coil shapes, FEMM data / Electromagnetic Projectile Accelerators / Forums

Forums

4hv.org :: Forums :: Electromagnetic Projectile Accelerators

« Previous topic | Next topic »

Various coil shapes, FEMM data

previous 1 2

Move Thread

LAN_403

Pinkamena

Tue Mar 06 2012, 11:52PM

Registered Member #4237 Joined: Tue Nov 29 2011, 02:49PM
Location:
Posts: 117

Saz43 wrote ...

I've seen it posted that a greater change in coil inductance as the projectile goes from outside of the coil to the coil's center means better magnetic coupeling- and thus is a good thing. I have to do some more reading though to understand the physics behind that.

Well, that seems to be about right... These two curves are inductance and flux coupling for different projectile positions. They're completely proportionate.
Fluxxouplingvsinductanc

Oh, and you know what's neat about the flux coupling graph? If you differentiate the function that describes it, you can use it to find a function for the force on the projectile, like this:
Solenoid Force Equation

Pinkamena

Sun Mar 18 2012, 10:35PM

Registered Member #4237 Joined: Tue Nov 29 2011, 02:49PM
Location:
Posts: 117

Allright.
A little thread-bumping, but Forty asked about this before, and the simulations also brought up something very interesting.

I ran some simulations where I changed the material of the projectile and got force curves for the displacement of the projectile. Here's the dimensions of the coil and projectile:
Projectile: 1.6cm x 10cm, round nose
coil: Length=8cm, inner diameter=1.8cm, outer diameter=3.8cm
Outer sleeve: 1cm

I tested pretty much every magnetic material preset that FEMM had available, and got these curves (click for large version):

It is interesting to note that nearly all the metals behave like iron. Except for mu-metal and 416-steel, all the metals gets pretty much the same force curve as iron.
There are however two metals that act very different from the rest of the bunch, and that's "Supermalloy" and "Carpenter silicon iron core". They do not get the same plateau as the other metals do, but rather just keeps increasing, before dropping off as the projectile arrives at the scenter of the coil.

I assumed that this increase was due to the metals not being magnetically saturated, so I yanked the current density from 17A/mm to 1000A/mm, to see what would happen. This brings out a force curve that looks much like the force curve for an iron projectile where the coil haven't got an iron sleeve. That's because the sleeve has been completely saturated, and doesn't do much for flux linkage.

Changing the sleeve material to supermalloy, while also keeping the projectile supermalloy gives this force curve:

This looks a lot like the Supermalloy forcecurve from the first graph, which might mean that the supermalloy sleeve isn't saturated.

I'm honestly not sure what to think of this. Is supermalloy a coilgun builders wet dream?
There's not a lot on Wikipedia about this metal, but what little information htere is fits well with the simulation (extremely high magnetic permeability ( approximately 800000 N/A2)). It also has a low coercivity, which is good. Then it won't be permanently magnetized easily.
Read about it here.

What are you guys thoughts on this?

Adrenaline

Mon Mar 19 2012, 01:19PM

Registered Member #235 Joined: Wed Feb 22 2006, 04:59PM
Location:
Posts: 80

I am surprised the Hiperco-50 and Vanadium Permedur didn't do so well considering their high magnetic saturation.

Pinkamena

Mon Mar 19 2012, 03:47PM

Registered Member #4237 Joined: Tue Nov 29 2011, 02:49PM
Location:
Posts: 117

Adrenaline wrote ...

I am surprised the Hiperco-50 and Vanadium Permedur didn't do so well considering their high magnetic saturation.

"Vanadium permendur, on the other hand is suitable for very high flux-densities but has fairly low values of permeability"
Taken from this site:

I think that mean you'd get better results if the current density was higher. I'll check it.
EDIT:

Here we go, ran the simulation with 1000 A/mm^2. The iron sleeve becomes saturated, that's the reason for the smooth force curves. Anyways, Vanadium Permendur is clearly better than iron in this case. But I don't think it's worth using it for hobbyists!
31510394

Adrenaline

Mon Mar 19 2012, 04:01PM

Registered Member #235 Joined: Wed Feb 22 2006, 04:59PM
Location:
Posts: 80

Pinkamena wrote ...

"Vanadium permendur, on the other hand is suitable for very high flux-densities but has fairly low values of permeability"
Taken from this site:

I think that mean you'd get better results if the current density was higher. I'll check it.

Yea I could see that with Permedur u of ~6850 and Supermalloy of ~590,000.

Forty

Tue Mar 20 2012, 04:21PM

Registered Member #3888 Joined: Sun May 15 2011, 09:50PM
Location: Erie, PA
Posts: 649

goldmine has been trying to sell this stuff for a while. Might be worth a look (mu=400,000.) You could try it with one of the ideas presented in your projectile thread and roll it up into a tight spiral to make a projectile. Of course if it's conductive you might want to give it a light coat of something on one side so that the spiral design can still reduce eddy currents.

I'm glad that the super magnetic alloys gave better force results because on most other types of curves I see, they don't look that impressive so I have to wonder what's the point of the alloy. The downside is that they are still quite hard to obtain for us amateurs. It would also be hard for us to make directly because of the high melting temp of molybdenum. Not even an iron thermite reaction gets hot enough.

I'm also glad that the silicon steel was a little better too. The projectiles I've been working up to using are from solenoids and I believe to be silicon steel, although a closer analysis might be a good idea. I wonder how I could check.

It might be possible to piece together a coil shield from the pieces of metal found in hard drives (it's what the magnet is stuck to.) I believe these are made of mu-metal (mu=80,000-100,000)

Of course these pieces aren't nicely shaped, so you'd probably need quite a few in order to slice and dice your way into a decent sleeve (at least there wouldn't be eddy currents!)

That 1000 A/mm^2 value sounds like a lot, but am I right to conclude that it's simply 1000A flowing in something like 16 or 18awg wire, in which case it is actually a typical coilgun value?

Pinkamena

Tue Mar 20 2012, 09:45PM

Registered Member #4237 Joined: Tue Nov 29 2011, 02:49PM
Location:
Posts: 117

Forty wrote ...

I'm glad that the super magnetic alloys gave better force results because on most other types of curves I see, they don't look that impressive so I have to wonder what's the point of the alloy. The downside is that they are still quite hard to obtain for us amateurs. It would also be hard for us to make directly because of the high melting temp of molybdenum. Not even an iron thermite reaction gets hot enough.

Indeed. A quick search on google turns up nothing. I'm pretty sure Supermalloy is completely out of the question...

Forty wrote ...

I'm also glad that the silicon steel was a little better too. The projectiles I've been working up to using are from solenoids and I believe to be silicon steel, although a closer analysis might be a good idea. I wonder how I could check.

Steel from transformers is very good for coilgun projectiles, due to the thin hysteresis curve. You practically won't have to worry about wasting energy by making your projectile into a permanent magnet (That will likely happen with an iron projectile, due to the thick hysteresis curve).

Forty wrote ...

That 1000 A/mm^2 value sounds like a lot, but am I right to conclude that it's simply 1000A flowing in something like 16 or 18awg wire, in which case it is actually a typical coilgun value?

Ah, no. I just wanted a really high magnetic field, so I simply changed the coil current from 17A to 1000A. The turns and AWG was still the same, 22AWG @ ~2000 turns!

Oh, and regarding testing some other kinds of metals, I'm not sure I can do that. I mean, I can define my own kind of materials in FEMM, but I can't figure out the settings:
29385937

klugesmith

Thu Mar 22 2012, 02:44AM

Registered Member #2099 Joined: Wed Apr 29 2009, 12:22AM
Location: Los Altos, California
Posts: 1716

Pinkamena wrote ...
I tested pretty much every magnetic material preset that FEMM had available, and got these curves (click for large version):

...
What are you guys thoughts on this?

Might there be an error or two in the FEMM material libraries?

If a simulation gives a surprising and counter-intuitive result, I suggest an independent investigation of the material properties.
Also make sure the material name is not ambiguous, and not too general. Should you find an error, notify Dr. Meeker and don't be too proud of yourself. It's like finding your first compiler bug.

If you are right (or wrong) but not arrogant, most program authors will respond with a gracious acknowledgement.

[edit]

Pinkamena wrote ...
That 1000 A/mm^2 value sounds like a lot, but am I right to conclude that it's simply 1000A flowing in something like 16 or 18awg wire, in which case it is actually a typical coilgun value?

Yes, well figured.
From references like

, we can extrapolate the impulse current for adiabatic heating of Cu from initial 75 degrees C to 150 degrees C.
As Bjorn points out time and again, you can get the same result from Cu resistivity and specific heat.
The tabulated values agree, and the error from ignoring the tempco of resistance is only a few percent.
I^2.t is about 10900 A^2.s per square mm squared. That's 1000 A/mm^2 for 10.9 ms, or 100 A/mm^2 for 1.09 seconds.
Matches the 5-second ratings in the tables: about 47 A/mm^2.

The state of the art in cooling allows world-class Bitter solenoids to run continuously at over 200 A/mm^2, to generate uniform DC fields up to 35 teslas in useful volumes. Can't get there with just superconductors.

Forty

Mon Apr 02 2012, 06:12PM

Registered Member #3888 Joined: Sun May 15 2011, 09:50PM
Location: Erie, PA
Posts: 649

Hey I've got another funky idea you could simulate if you want.
A core with a triangular cross section that is large at one end and small at the other. An example could have the length being twice that of the projectile length and the maximum thickness being a projectile length (ie, triangle sides are L and 2L, so that it has the same area as a square L x L coil.)
The idea is that if the projectile were starting at the small end, it would feel a stronger field in front of it and be attracted throughout the length of the coil.
Of course the coils behind projectile would be slowing it down at the same time, so the net result would be small, but I was thinking that applying a smaller force over a longer distance might be a good way of initially accelerating the projectile from rest (this would be the first stage coil)
The pulse could also be much longer, and a sensor on the outside of the large end could be used to turn it off (since the "ideal" place for the projectile would actually have it protruding from the end rather than in the center of the coil.)

So basically the part to simulate, if possible, is the L x 2L triangular coil versus a L x L square coil on a L length projectile starting at a distance of 0 from the coil and ending at 2.5L for the triangle coil (sticking halfway out) and at L for the square coil (completely inside.)
Then from the force curves at identical A/mm^2 conditions, compare the total work done on the projectile starting from rest.

If you don't have the time or interest to do it then that's fine, I could just build one to see how (poorly) it works.

previous 1 2

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob