PYTHON · NUMPY

The magnet, switched

Model an electropermanent magnet: hold force versus gap, and why switching beats holding.

01Challenge

An electropermanent magnet holds 14.6 N at contact, falling to 2.34 N across a one-millimetre gap, then lets go. The starter hold_force() ignores the gap. Implement the interpolation, then compare the energy of the two ways to hold: an EPM pays 0.3 J once to switch and then holds for free, while an electromagnet burns power the entire time. The magnet you can switch is what lets a robot reconfigure its own body without spending power to stay together.

Predict firstOver one minute of holding, which costs less energy?

02Model

A permanent magnet holds for free but you can't turn it off; an electromagnet turns off but burns current to hold. An electropermanent magnet is both: a pulse flips it, and then it holds with zero standing power. That is the substrate of docking, self-assembly, and modular reconfiguration -- a body whose shape becomes part of the policy.

03Guided practice

Step A (worked): hold_force returns F0 for every gap, so the 1 mm and 2 mm rows are wrong -- run it. Step B (the fix): interpolate F0 + (F1 - F0)*gap for gaps up to 1 mm, and return 0 beyond. Step C (the point): the energy cell shows the EPM flat at 0.6 J for two switches while the electromagnet climbs to 120 J over a minute. Step D: PASS when the force curve and the energy comparison are both right.

04Feedback

PASS WHEN PASS when hold_force gives 14.6 N at 0 mm, 2.34 N at 1 mm, 0 beyond -- and switching costs less than holding.

If your run fails, check:

Read the bench’s printed result — it passes when hold_force gives 14.6 N at 0 mm, 2.34 N at 1 mm, 0 beyond -- and switching costs less than holding. Run the cells top to bottom; if one errors, the FAIL line names the step to fix.
hold_force still ignores the gap -- interpolate between 14.6 and 2.34 across the first millimetre.

05Retrieve & space

What can an electropermanent magnet do that a plain permanent magnet cannot?
Why does an EPM's energy stay flat while an electromagnet's climbs?
Name one physical-AI use for a magnet you can switch.

06Mastery & project

Pass the lab: your force model matches the real EPM, and you have shown why switching beats holding. That is the energy story behind self-reconfiguring robots.

The reconfiguration and self-assembly frontier -- M-Blocks, SMORES-EP, the FluxGrip gripper -- rests on this. The interactive lets you drag a module into the capture basin and watch it self-align.

GO FURTHER · THE LIVE PLAYGROUND

Latch it live: the electropermanent-magnet sim ↗ — pulse it on, drag a module into the basin, watch the energy stay flat.

← The graph is a policy From teleop to a released policy →

PYTHON · NUMPY

The magnet, switched

Model an electropermanent magnet: hold force versus gap, and why switching beats holding.

You read

the arrays and values already in scope

You change

the code you write in each cell

Fixed

the dataset and the checks that grade you