Slinky Solenoid · Guide · glidepathscience.org
GuideSlinky Solenoid
Contents
This Guide explains the ideas and equations behind every Walkthrough step, Concept Challenge question and Predict round, in the order a textbook would take them: first the field a coil makes, then the forces that field puts on the coil itself. Part 1 prepares the Intro track and Part 2 the Advanced track; each takes about fifteen minutes to read. Most sections end with a worked example that is not one of the sim's own problems.
Part 1 · The essentials
Which way the field of a coil points
Coils are how we make magnetic fields on purpose: a crane's electromagnet, a car's starter solenoid, a phone's speaker and an MRI magnet are all coils carrying a current. A Slinky is one too, wound from flat steel ribbon, and because it stretches it shows how a coil's shape sets its field.
The field of one straight wire circles it; the right-hand rule gives the sense, with the thumb along the current and the fingers curling the way the field goes. Bend the wire into a loop and all those little circles thread through the loop the same way, adding up to a field straight through it. Stack loops along a line and you have a coil, or solenoid, whose turns all add their fields along the axis. For a coil the rule is easiest in its grip form: curl your right hand's fingers the way the current goes around the turns, and your thumb points along the field inside. The field leaves one end, returns around the outside and comes back in at the other, like a bar magnet's; the end it leaves is the north pole.
Two mistakes are common: using the left hand, which reverses the answer, and thinking the field inside circles the axis as the current does. Circling is what a single straight wire's field does; a coil turns those circles into a field along its length.
Knowing which way the field points, we next ask how strong it is.
The field inside a long coil
Deep inside a long coil the field is uniform, the same at the axis as next to the winding and all along the coil. Its strength depends only on the current and on how tightly the turns are packed, measured by the turns per meter n: the number of turns N over the length ℓ, or one over the spacing of neighboring turns, center to center:
$B = \mu_0 nI$ $n = N/\ell$ = 1/(turn spacing)
Here I is the current in amperes, B comes out in tesla, and $\mu_0 = 4\pi \times 10^{-7}$ T·m/A is the magnetic constant. The radius does not appear: a fat coil and a thin one with the same turns per meter carry the same field inside. That makes stretching a Slinky a clean experiment. Twice the length keeps the turns and the current but halves n, so it halves the field. The mistake to avoid is putting the total N where n belongs; only the count per meter sets the field.
Uniform inside cannot hold everywhere. Somewhere the field has to leave the coil, and the next section looks at where.
Inside, outside and at the ends
A Hall probe slid along the axis of a real coil finds three regions. Deep inside, the field is $\mu_0 nI$. Beside the coil it is weak: the lines that run along the inside return around the outside, where they spread through all the space around the coil. And at the ends the field on the axis is about half its inside value. Picture a long coil as two half-coils end to end. At the joint each half gives half of $\mu_0 nI$; at the end of one alone, only that half is there:
at an end of a long coil: $B \approx \tfrac{1}{2}\mu_0 nI$
Each turn stands for one spacing of the coil, half on each side of it, so a coil of separate turns ends half a spacing past its last turn. That is where the Walkthrough puts its probe.
The usual mistake is to think "uniform inside" means full strength right up to the last turn. The field begins to drop a few radii before the end: one radius inside the end it is already about 15 % down, and beyond the end it keeps falling.
So far the field has been the result. From here on it is the cause: the coil's own field pushes on the coil's own current.
Parallel currents attract
A current in a magnetic field feels a force, $\vec F = I\vec L \times \vec B$, with $\vec L$ along a length L of wire. Two parallel wires each sit in the other's field, and the right-hand rule puts each force toward the other when the currents run the same way: parallel currents attract; antiparallel currents repel. For long straight wires d apart, the force on a length L of either is
$F/L = \mu_0 I_1 I_2/(2\pi d)$
Neighboring turns of a coil carry the same current the same way around, so they attract. Up close, two turns of radius R look like two long parallel wires of length $L = 2\pi R$, and the force between them becomes
two neighbors: $F \approx \mu_0 I^2 R/d$ ($d$ center to center)
The force goes as the square of the current, since each turn's current both makes a field and is pushed by the other's: double the current and the pull is four times larger. And d is center to center, the visible gap plus the ribbon's thickness; using the gap alone overstates the force.
One pair is only the start. In a coil of 86 turns, every turn pulls on every other.
Everyone squeezes the middle
Every turn on one side of the middle gap pulls every turn on the other side, so that gap carries the sum: with 43 turns on each side, 43 × 43 = 1,849 pairs. A gap near an end has fewer turns on one side, so the middle gap is squeezed hardest and closes first. A long coil offers a shortcut through all those pairs: the field itself carries the force. A field B exerts a magnetic pressure $B^2/2\mu_0$ (in pascals, with B in tesla), the same number as its energy per cubic meter, and across the middle it pulls the two halves together over the cross-section $\pi R^2$:
squeeze ≈ $(B^2/2\mu_0)\pi R^2 = \mu_0 n^2 I^2\pi R^2/2$
The second form, with $B = \mu_0 nI$, shows the squeeze growing as $I^2$ and as $n^2$. The easy slips are dropping the 2, which doubles the answer, and using the diameter for R, which quadruples it.
Whether that squeeze closes the coil depends on what pushes back.
The spring pushes back
A Slinky is famously soft: it walks down stairs. A coil spring's ribbon does not stretch when the coil does; it twists, and steel twists far more easily than it stretches. The N − 1 gaps between N turns act as springs in series: the same force passes through each, and their stretches add up, so each gap is N − 1 times stiffer than the whole coil. From the ribbon itself, a gap's stiffness comes from twisting a ribbon of shear modulus G and torsion constant J around a turn of radius R:
$k_{gap} = (N - 1)\,k_{coil}$ $k_{gap} = GJ/(2\pi R^3)$
The $R^3$ is why wide coils are so floppy. The common slip is to divide by N − 1 instead of multiplying; but a short piece of spring is always stiffer than a long piece of the same spring.
With a magnetic squeeze on one side and a spring on the other, a balance looks easy. It is not, for a reason the next section explains.
Why it snaps
Turn the current up slowly and at first the coil acts like any spring under a growing load: the gaps close a little until the spring's push matches the squeeze. But closing the gaps raises the squeeze itself, which grows as $1/d^2$ with the pitch d, while the spring's push grows only in proportion to how far they have closed, $(d_0 - d)$. This is positive feedback. Below a certain current the spring still wins each round and the coil settles shorter. Above it no balance is left, and the coil collapses on its own: the middle turns slam together and the rest pile in within a fraction of a second. That is the snap.
The mistake is to expect a smooth shortening in step with the current, as under a weight. A weight does not grow as the spring moves; the squeeze grows as the coil shortens.
The squeeze is not the only force. The same field also pushes outward on every turn.
The outward push
Cut the coil along its axis, as the Cross-section tab does: the current comes out of the page at the top of each turn and goes in at the bottom. Opposite sides of a turn carry opposite currents and repel, so every turn is pushed outward, whichever way the current flows. The field presses on the winding with the same $B^2/2\mu_0$, over the band $2\pi R \times d$ that one turn presents. With $B = \mu_0 I/d$, the outward push on one turn, adding up its sizes all the way around, is
outward on one turn ≈ $\pi\mu_0 I^2 R/d$
That is π times the pull between two neighbors. As vectors the pushes cancel, so the turn does not move sideways; together they stretch the ring. Yet the radius grows by microns at most, because steel resists opening a turn far more than the coil resists closing a gap (Part 2). The slip to avoid is thinking the push reverses with the current: every force here involves two currents, and reversing both leaves their product unchanged.
Forces are only half of what happens when a Slinky carries amperes. The other half is heat.
The hidden hazard: heat
Steel conducts about eleven times worse than copper, and a Slinky is a long, thin ribbon of it: a real resistor, turning electrical energy into heat at the rate $P = I^2R$. A uniform conductor of resistivity ρ, length L and cross-section A has
$R = \rho L/A$, $P = I^2R$, warming rate = $P/(mc)$
The last is how fast the temperature rises at first, before the coil loses heat to the air: the power over the heat capacity, mass m times specific heat c (470 J/(kg·K) for steel). As it warms it sheds heat by convection and radiation, and the temperature settles where the loss equals $I^2R$. The common slip is writing the power as $IR$, a voltage; doubling the current quadruples the heating.
Heat is not the only thing bare metal brings. A bare ribbon also lets the current take shortcuts.
Bare metal and short circuits
A real Slinky has no insulation. When turns touch, the current can cross the contact instead of going around. The contact $R_c$ is in parallel with the turn's own $R_{turn}$, and current divides inversely as the resistances, so the fraction still circulating around the turn is
$I_{turn}/I = R_c/(R_c + R_{turn})$
With a supply holding the total current fixed, a short weakens the squeeze there. If the spring then wins, the turns open, the short breaks, the full current returns and they slam shut again: the coil chatters. The mistake is to think a short makes more current flow; the supply fixes the total, and the short only changes the path.
Part 2 takes each of these ideas further: exact forces, real coil lengths, the field outside a helix, and the snap in equations.
Part 2 · Going further
Two rings, exactly
The wire estimate fails once the rings are about a radius apart. The exact force comes from the mutual inductance M of two coaxial rings of radius R a distance h apart, the flux through one per ampere in the other, a classic result in the complete elliptic integrals K(k) and E(k):
$M = \mu_0 R[(2/k - k)K(k) - (2/k)E(k)]$ $k^2 = 4R^2/(4R^2 + h^2)$
At constant current the force between the rings is $F = -I^2\,dM/dh$; M falls as the rings separate, so F is positive, an attraction. Differentiating gives the closed form
$F = \mu_0 I^2 hk/(4R(1 - k^2))\cdot[(2 - k^2)E - 2(1 - k^2)K]$
For $h \ll R$, k approaches 1 and F approaches $\mu_0 I^2 R/h$. Far apart, each ring is a dipole of moment $I\pi R^2$ and $F \to (3\pi/2)\mu_0 I^2 R^4/h^4$. Dividing by the wire estimate leaves a ratio that depends only on h/R; dividing the other way round, or by $\mu_0 I^2/h$ without the R, are the usual slips.
Exact forces need exact fields, and a real coil has ends.
A coil of finite length
Treat a coil as a uniform sheet of current, n turns per meter wrapped on a cylinder of radius R and length ℓ. Adding up the on-axis fields of its rings gives a compact result: at any point on the axis, $B = (\mu_0 nI/2)(\cos\theta_1 - \cos\theta_2)$, where $\theta_1$ and $\theta_2$ are the angles from the axis to the rims of the two ends, seen from that point. At the center and at an end it becomes
center: $B = \mu_0 nI\cdot(\ell/2)/\sqrt{(\ell/2)^2 + R^2}$ end: $B = (\mu_0 nI/2)\cdot\ell/\sqrt{\ell^2 + R^2}$
Both are long-coil values times a cosine. The center's, the cosine of the half-angle the coil subtends, is within 1 % of 1 once $\ell/2$ reaches about seven radii, a coil about seven diameters long. A shorter coil leaks: its lines bulge out of the ends, and $\mu_0 nI$ overstates even its center field. Mixing up ℓ/2 and ℓ, or R and the diameter, are the slips.
The sheet model treats each turn as a ring. A Slinky is a helix, and a helix has one more trick.
A helix is also a wire
As the current goes around a helix, it also advances along the axis by one pitch p per turn. Over the whole coil it carries the current I from one end to the other, exactly as a straight wire along the axis would. Outside the coil, that net current adds a field circling the axis; inside, it adds nothing, since no net current is enclosed there. Ampère's law gives it at once:
outside the helix: $B_{\varphi} = \mu_0 I/(2\pi r)$
Just outside the winding, at r ≈ R, compare it with the field inside, $\mu_0 I/p$: the ratio is $p/(2\pi R)$, a percent or two for a Slinky. The two fields count the current differently. Inside, every turn per meter adds its share; outside, the net current counts once. For a right-handed helix like the Slinky, $B_{\varphi}$ circles the same way the current goes around the turns. It therefore lies along the current, and $I\vec L \times \vec B_{\varphi}$ pushes nothing along the axis, so the squeeze is unchanged. Using the turns per meter in $B_{\varphi}$ is the usual slip; it gives the inside field instead.
With the fields in hand, we can look more closely at how a field pushes.
The field does the pushing
The magnetic pressure of Part 1 is more than a shortcut. In Maxwell's picture the field carries stress: along its lines it pulls like a stretched rubber band, and across them it pushes like a gas, both with size $B^2/2\mu_0$, which is also its energy per cubic meter. The force on everything on one side of a plane is that stress added up over the plane. Inside a coil the lines cross the middle plane and pull the halves together; at the winding they press outward.
magnetic pressure = $B^2/2\mu_0$
A short coil leaks. Shut, a Slinky is shorter than its own diameter: its center field falls well below $\mu_0 nI$ and varies across the middle plane, weakest on the axis. Added up over that plane, $B^2/2\mu_0$ gives only about half the long-coil squeeze; squaring the center field alone would not, because the field is stronger toward the winding.
Now we can put the snap of Part 1 into equations.
The snap current and hysteresis
Model the coil as compressed evenly to pitch p. The middle gap carries the long-coil squeeze $\mu_0 I^2\pi R^2/(2p^2)$, and the spring pushes back with $k(p_0 - p)$, where k is one gap's spring constant and $p_0$ the pitch with no current. A balance needs
$I^2 = 2kp^2(p_0 - p)/(\mu_0\pi R^2)$
As the current rises the pitch moves down this curve from $p_0$; the right-hand side rises, peaks and falls, as in the Part 1 figure. Its derivative vanishes at the fold, $p = 2p_0/3$. Past that current no balance is left and the coil snaps shut:
snap: $p = 2p_0/3$, $I_c^2 = 8kp_0^3/(27\mu_0\pi R^2)$
The pitch has closed by a third, not to a third, and the fold exists only if the turns have not touched first: if $2p_0/3$ is less than the ribbon thickness t, the coil closes smoothly. Shut, at p = t, the squeeze is far stronger, and the coil springs open only when it falls below the spring's push at contact, $\mu_0 I^2\pi R^2/(2t^2) = k(p_0 - t)$:
reopen: $I_{open}^2 = 2k(p_0 - t)t^2/(\mu_0\pi R^2)$
It reopens well below the snap current: hysteresis. The full model, summing every pair, corrects both. A coil a few diameters long is squeezed less than the long-coil estimate, most of all near its ends, so it holds on somewhat past $I_c$; and once shut its end gaps open first, so the rest springs open above the long-coil $I_{open}$. Forgetting the square root, or using the gap for $p_0$ or the diameter for R, are the usual slips.
The outward push was left as "microns at most". Here is why.
Bending versus hoop stretch
An outward load q per meter of ribbon can open a turn in two ways. If the coil's ends are free to turn, the helix can unwind slightly: fewer, bigger turns. Each turn then opens by bending the ribbon in its own plane, against the bending stiffness $E\cdot tb^3/12$, where E is Young's modulus and b the ribbon's width, measured in the plane of the turn. If the ends are held, the ring can only stretch around its circumference, against $E\cdot bt$:
$\Delta R_{bend} = qR^4/(E\cdot tb^3/12)$ $\Delta R_{hoop} = qR^2/(E\cdot bt)$ ratio $12R^2/b^2$
Bending is the soft way, by a factor of thousands, but even bending is stiff compared with closing a gap, because closing a gap only twists the ribbon. The ribbon's length is fixed, so a bigger radius means fewer turns: the free end rotates by $360^\circ \times N\,\Delta R/R$, which multiplies a tiny change in radius by the number of turns. Forgetting the 12, or using the diameter, are the slips.
There is another way to find every one of these forces, from energy alone.
Virtual work
A coil's self-inductance L measures the flux it makes through itself per ampere. For a long coil of N turns, radius R and length ℓ, the field $\mu_0 (N/\ell)I$ threads all N turns, so
$L = \mu_0 N^2\pi R^2/\ell$ $F = \tfrac{1}{2}I^2\,dL/dx$
The second equation is the force along any coordinate x at constant current: imagine moving that part of the coil a little, and ask how much the inductance changes. The sign surprises many students. The field's energy is $\tfrac{1}{2}LI^2$, and one might expect the force to be minus its slope. But at constant current, as the coil shortens and L grows by ΔL, the supply has to deliver an extra $I^2\Delta L$ to hold the current. Half of it goes into the field, $\tfrac{1}{2}I^2\Delta L$, and the other half does the mechanical work. So the force pushes toward larger inductance: the coil is pulled shorter. For the long coil, $|dL/d\ell| = L/\ell$, and the squeeze $\tfrac{1}{2}I^2 L/\ell$ is exactly the magnetic pressure times $\pi R^2$ again. It is the same physics seen through energy. Dropping the ½, or letting L grow as N instead of $N^2$, are the slips.
The supply has appeared twice now. What kind of supply it is changes how the coil behaves.
Battery or bench supply?
A bench supply set to constant current holds I fixed whatever the coil's resistance does. A car battery, and a bench supply set to constant voltage, holds V fixed instead, so the current is $I = V/R$. The difference matters as soon as bare turns short. A short puts a path in parallel with a turn, which lowers the coil's resistance. At constant current that changes nothing in total, the shorted turns carry less, their squeeze weakens, and the coil tends to chatter. At constant voltage the lower resistance raises the current, every unshorted turn carries more, the squeeze grows as $I^2$, more turns close and short, and the collapse runs away, with the heating rising too. The slip is to expect the two kinds of supply to behave the same when the coil's resistance changes.
Every model ends somewhere. The last section says where this one does.
Where the model ends
The sim treats every ribbon as non-magnetic, which copper, brass and aluminum really are. A real Slinky is carbon steel, and the coil's own field magnetizes it along the axis. Each turn then becomes a weak magnet stacked north to south on its neighbor, so the turns attract each other in addition to the pull between currents, and a steel coil snaps at a lower current than a copper coil of the same size and stiffness. A rough long-coil estimate, treating the steel as ideally soft and far from saturation, puts the Slinky's snap about 14 % lower. Shut, the thin contact gaps would hold the turns together far more strongly still.
Several smaller effects are also left out. A long squeezed coil can bow sideways or buckle; here every turn stays on the axis. The leads make their own field. The current spreads unevenly across the ribbon's width. And the coil's inductance delays any change in the current by its L/R time, about 70 µs for the open Slinky, far shorter than any motion the stage shows. Each of these is small, or clearly marked, and none changes the main story: the field inside is $\mu_0 nI$, parallel currents attract, and a soft spring loses the race.
Reference
Common mistakes
- Total turns for turns per meter. $B = \mu_0 nI$ uses n = N/ℓ, never N.
- The gap for the spacing. Spacing and pitch are center to center: the gap plus the ribbon's thickness.
- Full strength at the ends. On the axis at an end, a long coil has about half its inside field.
- Forces that go as I. Every force between currents goes as $I^2$, and reversing the current changes none of them.
- Losing the 2. The magnetic pressure is $B^2/2\mu_0$.
- The diameter for the radius, in $\pi R^2$, in $\mu_0 I^2 R/d$, and in the cosines of a finite coil.
- Softer gaps. Each gap is N − 1 times stiffer than the whole coil.
- $IR$ for the power. The heating is $I^2R$.
- A short that raises the current at constant current. It only changes where the current goes; at constant voltage it does raise it.
- The long-coil formula for a short coil. Multiply by the cosine of the half-angle.
Equation sheet
| Part 1 | |
|---|---|
| long coil | $B = \mu_0 nI$, $n$ = turns per meter; $\mu_0 = 4\pi \times 10^{-7}$ T·m/A |
| at an end | $B \approx \tfrac{1}{2}\mu_0 nI$ |
| two neighbors | $F \approx \mu_0 I^2 R/d$ ($d$ center to center) |
| squeeze | $(B^2/2\mu_0)\pi R^2 = \mu_0 n^2 I^2\pi R^2/2$ |
| outward | $\pi\mu_0 I^2 R/d$ on one turn |
| spring | $k_{gap} = (N - 1)\,k_{coil}$; $k_{gap} = GJ/(2\pi R^3)$ |
| heat | $R = \rho L/A$, $P = I^2R$, warming rate $P/(mc)$ |
| bare metal | $I_{turn}/I = R_c/(R_c + R_{turn})$ |
| Part 2 | |
|---|---|
| two rings | $F = \mu_0 I^2 hk/(4R(1 - k^2))\cdot[(2 - k^2)E - 2(1 - k^2)K]$, $k^2 = 4R^2/(4R^2 + h^2)$ |
| center | $B = \mu_0 nI\cdot(\ell/2)/\sqrt{(\ell/2)^2 + R^2}$ |
| end | $B = (\mu_0 nI/2)\cdot\ell/\sqrt{\ell^2 + R^2}$ |
| helix | $B_{\varphi} = \mu_0 I/(2\pi r)$ outside |
| pressure | $B^2/2\mu_0$ |
| snap | $p = 2p_0/3$, $I_c^2 = 8kp_0^3/(27\mu_0\pi R^2)$ |
| reopen | $I_{open}^2 = 2k(p_0 - t)t^2/(\mu_0\pi R^2)$ |
| radius | $\Delta R_{bend}/\Delta R_{hoop} = 12R^2/b^2$ |
| virtual work | $L = \mu_0 N^2\pi R^2/\ell$, $F = \tfrac{1}{2}I^2\,dL/dx$ |
Symbols: B field (T) · I current (A) · n turns per meter · N turns · ℓ coil length · R coil radius (or resistance, in the heat rows) · d, p center-to-center spacing (pitch), $p_0$ with no current · t, b ribbon thickness and width · h distance between two rings · K, E complete elliptic integrals · k one gap's spring constant (N/m), or the elliptic modulus in the two-ring row · G, E shear and Young's moduli · J torsion constant · ρ resistivity · m mass · c specific heat · $R_c$ contact resistance · L inductance (or ribbon length, in R = ρL/A).
Further reading
OpenStax, University Physics Volume 2, chapter 12 (sources of magnetic fields: parallel currents, loops, solenoids) and chapter 14 (inductance and the energy in a magnetic field), an open textbook licensed CC BY-NC-SA 4.0 (openstax.org); D. J. Griffiths, Introduction to Electrodynamics, on the Maxwell stress tensor; J. D. Jackson, Classical Electrodynamics, §5.5, the field of a circular loop in elliptic integrals; R. G. Budynas and J. K. Nisbett, Shigley's Mechanical Engineering Design, on coil springs and the elastic constants of steel. The Field Manual lists every source and every model choice.
Slinky Solenoid v1.6 · © 2026 Dr. David Meier · CC BY-NC-SA 4.0 · glidepathscience.org