Advice on short duration, high current electromagnet circuit

I have participated in an informal competition held by the electronics society at my uni to design a working high strength electromagnet using commodity off the shelf (COTS) components. The electromagnet is to be powered by any commonly available means with a maximum budget of =A3100. There are basically no rules to the competition except the budget and COTS requirement.

As far as I know, everyone else is working on the principle of a steady state electromagnet circuit. Typically, with the appropriate AC

-> DC voltage rectification circuit and the maximum current determined by the ampacity of the windings on the electromagnets. Alternatively, other people are working on electromagnets powered off batteries which obviously have limitations of internal resistance.

I think that since the measurements of the electromagnet's strength will be done over perhaps 30 seconds in the competition, the ampacity of the windings is not particularly relevant, i.e. it is possible to overload the current and rely on the thermal mass of the assembly to soak up the excess heat generated. Therefore, I would like to go for a circuit that can deliver a sustained, high current pulse to an external load.

I would like to try to build a circuit that charges up a capacitor bank in parallel and discharges it across the electromagnet in series. A Marx cascade seems to be of the sort I need but Marx cascades work with high voltage capacitors (to trigger breakdown of the spark junctions) and I think for a ~30 second sustained current, I will need capacitors with a high capacitance. Capacitors with a high capacitance tend to have a low voltage (I could find a 30F aluminium electrolytic capacitor on Farnell with a voltage of 2.3V)

As term has just started, please assume I just have knowledge of electronics from A-level physics.

Does anyone have any comments on my approach?

Will the maximum magnetic field strength I can achieve be entirely limited by the saturation of the core? I know the saturation of different ferromagnetic materials varies, but wouldn't the magnetic field be mostly generated by current flow rather than alignment of magnetic domains in the core?

Where can I source good materials like soft iron to build my core from (other than steel bolts)?

Do you have any suggestions on circuit topologies I could look into?

Thanks.

Jessie xx

Reply to
Jessie
Loading thread data ...

The shape of the core is crucial. A straight rod is the worst possible choice because most of the magnetic flux path is through air. The best would be a toroidal (circular) core with a tiny air gap across which the field is measured. Better still, if the rules permit, have no air gap at all, confine the field to the core and let the judges worry about measuring it.

-- Joe

Reply to
J.A. Legris

You don't need a Marx circuit, because you don't need a huge voltage. But you'll never be able to afford enough capacitors to produce more than a feeble current for 30 seconds... do the math.

If you want a high field for, say, 100 microseconds, a cap bank is good. I did that when I was a kid... a lot of old teeve set electrolytics in parallel, charged to a few hundred volts, discharged into a 10-turn coil of heavy wire. It would magnetically saturate anything placed inside.

An iron core will saturate at low fields. Serious field strength must be ironless, a lot of current into a copper coil. The tinier the region, the easier it is to get a high field in it.

For 30 seconds, a car battery is pretty good.

Is the 30 second thing a rule of the contest? A copper coil *will* get very hot at high currents in 30 seconds, a lot hotter than it would in

30 microseconds.

The important thing here is: do the math. You can calculate the expected current, resulting solenoid field, coil power dissipation, and coil temperature. Do it.

John

Reply to
John Larkin

As far as I know, everyone else is working on the principle of a steady state electromagnet circuit. Typically, with the appropriate AC

-> DC voltage rectification circuit and the maximum current determined by the ampacity of the windings on the electromagnets. Alternatively, other people are working on electromagnets powered off batteries which obviously have limitations of internal resistance.

I think that since the measurements of the electromagnet's strength will be done over perhaps 30 seconds in the competition, the ampacity of the windings is not particularly relevant, i.e. it is possible to overload the current and rely on the thermal mass of the assembly to soak up the excess heat generated. Therefore, I would like to go for a circuit that can deliver a sustained, high current pulse to an external load.

I would like to try to build a circuit that charges up a capacitor bank in parallel and discharges it across the electromagnet in series. A Marx cascade seems to be of the sort I need but Marx cascades work with high voltage capacitors (to trigger breakdown of the spark junctions) and I think for a ~30 second sustained current, I will need capacitors with a high capacitance. Capacitors with a high capacitance tend to have a low voltage (I could find a 30F aluminium electrolytic capacitor on Farnell with a voltage of 2.3V)

As term has just started, please assume I just have knowledge of electronics from A-level physics.

Does anyone have any comments on my approach?

Will the maximum magnetic field strength I can achieve be entirely limited by the saturation of the core? I know the saturation of different ferromagnetic materials varies, but wouldn't the magnetic field be mostly generated by current flow rather than alignment of magnetic domains in the core?

Where can I source good materials like soft iron to build my core from (other than steel bolts)?

Do you have any suggestions on circuit topologies I could look into?

Thanks.

Jessie xx

Since iron or any other core material saturates at under 2T, the only way to get high flux densities, say 10T is with an air core coil. This means as high a current as possible into as many turns of copper as possible. How much current can you draw from the mains? Can you get 100 Amps?, and at what voltage? 240 Volts? To get the highest flux density, the area to be measured or center of the coil must be as small as possible. Can it be less than one sq cm? In any case use the highest power source available, probably the mains. 24KW is not out of the question. Batteries could work for a short duration. Rectify the AC if used..

The coil should be nearly square in cross section and a short as possible. Flux density in a solenoid is: B=uo*N*I/l, uo = 4*pi*10^-7 which means NI has to be huge and l in meters has to be short, say several cm.

So now it's just a matter of cramming as many turns of wire into the cross section say 5cm by 5cm possibly larger as you can. Given the maximum current pick a wire gauge that won't fuse at that current over the duration, say one minute. Given that wire gauge and the available area, calculate the wire length and resistance. Now, your required voltage must give you the required current at that resistance. Re-adjust as necessary to maximize amp-turns, NI.

Keep in mind that you are dissipating lots of power in a small volume so the duty cycle is necessarily very low. Allow plenty of time between runs for the coil to cool.

Let us know how it comes out.

Reply to
Bob Eld

The =A3100 limit is not absolute - it is based on a merit function of strength/cost. The winner has the highest strength for the lowest cost. I just said =A3100 because we were told the club can certainly subsidise project costs under this, but beyond this and designs will need to be considered on a case-by-case basis. The merit function still applies across the =A3100 threshold though.

The property of the Marx circuit I was hoping to use was that it charged the capacitors in parallel but discharged them in series. I can't think at the moment of how else to achieve a convenient and rapid circuit reconfiguration like this.

10 turns? Wow, that's very little. I thought the magnetic force was proportional to the number of amp-turns - I suppose you were going entirely for the current? Did you wrap the turns over each other to form concentric coils or was it just a single coil (better heat dissipation)?

A quick back of the envelope calculation showed that although resistance is proportional to wire length, the number of turns is also proportional to wire length. Therefore from a perspective of magnetic force alone, the wire length plays no contribution.

Measurement of the field strength would be by checking the electromagnet's pull strength against a set of scales.

Then the car battery's cost must be factored in. Running off mains with a rectifier is definitely cheapest.

No, it isn't a rule of the contest. It is what I'm guessing would be a reasonable time for the judges to switch on the thing and measure pull strength. It is most unfortunate in my case my pull strength would be decaying from the start...

Yes, you're right of course that I should do the math before I foray into costly purchases. I am mildly allergic to maths but here goes...

Current decay from a discharging capacitor: I =3D Io(e^-t/RC) If we consider 50% of the initial current to be the "spent" point and this to occur within 30 seconds,

0.5 =3D e^-30/RC With C =3D 30F, ln(0.5) =3D -1/R. So R =3D 1.44 Ohms

Therefore the solenoid must have a resistance of at most 1.44 Ohms. Oh dear, that means I've only got 1.6A going through it :(

I will be back to this thread in a moment, it seems I do have some maths to go through to optimise the thing!

Jessie xx

Reply to
Jessie

That's an entirely different problem from creating a high field. What is the scale going to be attached to?

John

Reply to
John Larkin

Are you allowed to use junkyard/surplus parts? Do they have junkyards/ surplus parts in UK? I know in the US, you can get some pretty awesome stuff, if you know where to shop (and have time to look at all the neat stuff you can get.)

My primary point being, if the parts aren't required to be NEW off-the- shelf, you can probably save from 50-90% of the cost of an equivalent assembly made from new, or conversely, double to quintuple the results you can get for the same money.

Good Luck! Rich

Reply to
Rich Grise

It is, but resistance is inverse (roughly speaking).

An ordinary snap-in capacitor will dump its load in about 10us.

This was 400V, switched with an ordinary xenon flash tube (t_r ~ 5us).

formatting link

The kind of caps you'll find in an old TV will have ESR in the couple- ohms range, and slower discharges, essentially RC, with lots of power burned in the cap itself. (Heck, even the snap-in gets a little warm after a few shots of the flash tube.)

So what you want is a coil with just enough inductance that R, L and C are critically damped. If you figure 470uF, 0.1 ohm (typical snap-in ESR), then L =3D 4.7uH. Then L/R =3D 47us, R*C =3D 47us and sqrt(L*C) =3D

47us (the "resonant time constant", because 1/2*pi*sqrt(L*C) is the frequency). I think that's critically damped. So you'll get a neat exponential which is shorter than any other combination: more R than L makes it slower; more L than R rings, and decays slower as a result.

Such a coil might be 10-20 turns in a 1 inch size. Probably you'll want to use copper strip encased in epoxy. Peak current will be on the order of 1kA, or 10-20kAt for the coil, which is on the order of

100kAt/m, which is close to what, 0.1T in free space (mu_0 ~=3D 1e-6)? Well, gettin' closer.

Now just increase the energy in C by a few orders of magnitude, play around with values, and stretch that exponential out to the 30 seconds you need (time constants will be closer to 60 seconds to keep field up during that time). Offhand, you'll need lots of farads and henries to do that, while keeping R under an ohm. Definitely not a cost effective method.

The direct power method is the way to go. You will need water cooling at least. Something quite cold would be a good idea, since copper has an awful tempco. You could even invest in thermos of liquid nitrogen, which isn't as expensive as it sounds. Though I'd be worried that the coil would heat up too much and geyser all the LN2 out instead of heating it up slowly. There isn't as much heat capacity in LN2 as there is in water, not by an order of magnitude or two.

Speaking of LN2, I wonder if high Tc superconducting cable is available for less than your immortal soul (or a research grant, same thing). That would be the way to go: drop in a few cups of LN2 and zap away. For a couple tesla, you don't need much wire, nor does it have to be very cold. Guess it comes down to the cable... if that stuff is worth more than gold, even if you don't need much of it, it'll still be the most expensive part, much more expensive than the cooling.

Tim

Reply to
Tim Williams

Then you should match your coil/core to the characteristics of what's being pulled. Assuming an iron slab is placed in contact with the face of the magnet's core during measurement, you should maximize the area of contact and flux density while taking into consideration the flux distribution within the slab and its own saturation point. There are software tools available that will let you calculate the flux distribution. Eg:

formatting link

-- Joe

Reply to
J.A. Legris

In the "exceeds your budget considerably, but is off the shelf" line, a lab I worked in used a commercial spotwelding supply for driving large electromagnets to contain a plasma discharge. Something like 13,000 amps for 5 seconds or so - the thing took up most of a room, and the "wires" were 6 inches wide and 1/2 an inch thick.

A normal stick (arc) welder as used surplus equipment might just fit your budget and get you decent current and rectification, albeit only in the hundreds of amps range - which will still provide you with a cooling challenge. Copper pipe as the wire with de-ionized water run through it might work, but the pump will probably blow your budget again. And I can't find any welders for sale on a random selection of UK cities on craiglist, so they may not be as common there as here.

--
Cats, coffee, chocolate...vices to live by
Reply to
Ecnerwal

to

what

measured

one

current

one

required

the

Maximizing force a magnet produces is a very different thing than maximizing flux density as I assumed you meant when you said "magnets strength" above.

You must use an iron core to convey the flux to the pole tips where the test piece will grab. Your pole tips should "match" the test piece and be more or less of the same total area or you will be wasting flux. Both the north and south pole tips should touch the test piece with the largest area available. This makes a complete magnetic circuit that includes the test piece. The air gap between the pole tips and the test piece should be minimized. This means grind or polish the tips absolutely flat or to the contour as the test piece if it's not flat.

The same rule about maximizing amp-turns applies as above in the air core case, but there is no point it going for a flux density larger that about

1.8T, the saturation point of the iron. Because the air gap is so small, about .002cm with polished pole tips, this flux density is easily reached with far fewer amp-turns than in the air core case.

Design the core to maximize it's area at all points around the magnetic circuit to the same or larger area as the each pole tip. If multiple pieces of metal are used be sure to minimize the air gap which is always present when two piece of metal are in contact by grinding and/or polishing the contact areas.

Your magnet should be designed with the test piece in mind and need not be larger than the area of the test piece dictates.

A well designed lift magnet can operate on very low power, maybe 100 watts or less and pull thousands of Kg when the air gap(s) are very short. There is no need for capacitive discharge or other gimmicks. Study the way professional lift magnets are designed and constructed.

Reply to
Bob Eld

Sorry, I thought where there is smoke there's fire so they were one and the same.

In Year 12 physics what we learnt about magnetism was not very deep, so please pardon my ignorance.

I know magnetic field strength decays in an inverse square law but what is the maximum pull strength a modern, well designed resistive and superconducting electromagnet can achieve on a given ferromagnetic test mass at various distances (e.g. 1m, 2m, ...)

How can this pull strength be increased?

But even if a ferromagnetic test mass has been saturated, surely increasing the electromagnet's field strength will increase the pull on it?

I'm just trying to see if this problem can be tackled from many angles.

Would having a very fine ferromagnetic dust on my electromagnet's poles (analogous to thermal paste for cooling) help improve the contact and reduce flux leakage?

Thanks for your input.

Jessie xx

Reply to
Jessie

Find out what it costs to rent an MRI for a few minutes. ;-)

Cheers! Rich

Reply to
Rich Grise

maximizing

above.

They are related but as soon as iron saturates, the increase in flux density is of little value for developing force. Secondly air core coils have, by definition, a very long gap. All things being equal, higher forces result from shorter gaps. Therefore designing for maximum force must reduce the gap to the absolute minimum.

test

more or

and

available.

air

means

piece

I have no idea about the forces with large gaps or air cores of meters outside of the magnet. The simple equation for force in a solenoid is: F(newtons) = uo*N^2*I^2*Ae/(2*lg). I in amps, Ae=effective area in sq meters. lg= length of gap in meters This is dependant on geometry so it's hard to generalize. It also ignores and magnetic path in iron or other high permeability material. It assumes that (all) magnetizing force is concentrated in the air gap. It's kind of like placing a 10 ohm resistor in series with a 100K resistor. Most all of the voltage drop will be across the

100k resistor, so it is with magnetics, the bulk of the magnetizing force will be across the high reluctance air gap, the iron is just a conductor to get there. Large air core coils can have high forces that's why your not supposed to take a monkey wrench into an MRI examination. You might be hit on the head with a flying wrench.

Make lg very small, maximize NI, maximize the effective area, Ae.

core

about

reached

That is true but the permeability of a saturated magnetic material drops from thousands down to unity when saturated so the forces will increase at a rate thousands of times less than when the material is not saturated. In otherwords diminishing returns sets in making the gain useless.

pieces

present

Don't know, it depends on the dust. Remember fine dusts have millions of tiny air gaps between the myriad of particles. That's how certain powdered iron inductor cores control permeability. The best is a machine or ground and polished fit between surfaces.

be

watts

There

Reply to
Bob Eld

Dear all,

I thought I'd post an update of what I've been up to. I decided to use a 500W ATX power supply for a computer as my current source. It rectifies the AC for me, provides a stable known output, has cooling and has surge control. At the end of the project I will still have a high current, high wattage power supply which is always handy - proper high current lab bench power supplies seem to be really expensive!

I hope to be using the 12V, 35A lead to power the solenoid, but will start with smaller wattages and see how quickly it gets hot and how well it conducts heat away.

Jessie xx

Reply to
Jessie

Yeah if they are putting a piece of iron on the end of the spring, then the maximum field it can 'measure' is in the 1-2 Tesla range depending on the piece of iron. You are also going to 'get points' for having a large field gradient. After all homogeneous magnetic fields don't make any force, only a torque.

George H.

Reply to
George Herold

izing

ove.

test

re or

and

able.

e air

means

piece

re

ut

,

ed

eces

nt

be

tts

ere

"> I know magnetic field strength decays in an inverse square law but

Opps, Electric fields from point charges decay as 1/r squared. There are no magnetic monoploes, so for magnetic dipoles the field decays as

1/r cubed. The induced force from a magnetic dipole on another piece of magnetic material is like the induced electric dipoles in Van der Waals attraction and goes as 1/r to the sixth power. You really notice this when bringing a screw driver near a super conducting magnet. Not much force and then... hold on!

George H.

Reply to
George Herold

Excellent, Now that you've selected the power supply you 'know' the optimal resistance of the coil. (Call it about 0.33 ohms.) Now you need to design the coil with that number in mind.

George H.

Reply to
George Herold

If they use a spring scale, its extension vs force curve complicates life.

John

Reply to
John Larkin

Sure, but just add calibrated marks. I'd be more worried about saturating the iron, as noted previously by Bob E. Won't the force go away once the iron saturates? The force is also going to be very position dependent. They might do better to measure a torque.

George H.

Reply to
George Herold

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.