Every term here is a decision a magnetic design can get right or wrong. From coercivity to flux closure, tagged honestly: what we design and simulate, what we design with, and what we’re fluent in.
0terms across nine domains
Core Designed and simulated hereApplied Designed with routinelyField Fluent, not built on
Why magnetic design takes expertise
Magnetics is unforgiving. Six ways designs quietly fail.
Magnets look simple: a block that sticks. The physics underneath is nonlinear, steeply distance-dependent, temperature-sensitive and easy to destroy. None of it shows up on a datasheet. All of it shows up in the field.
01B ∝ 1/r³
Distance is brutal
Double the distance from a magnet and roughly seven-eighths of its far field is gone. A part that holds hard at contact can be untouched a few millimeters away.
Our approachEvery fraction of a millimeter of gap, coating or paint is designed for, not discovered.
02F ∝ ∇(B²)
Strength isn’t force
A strong, uniform field pulls on almost nothing. The force on a loose part comes from how steeply the field changes, so the shape of the field matters more than its peak.
Our approachThe gradient at the working surface is engineered, not assumed from a datasheet value.
03≈2.1 T
Steel fills up
Past saturation, adding magnet buys nothing. The steel is full, and the extra flux leaks out as stray field instead of doing work.
Our approachFinding which section saturates first is routine diagnosis, done before anything is built.
04−0.11 %/°C
Heat steals strength
Neodymium loses remanence as it warms, and past the knee of its curve the loss is permanent. A bench result is a room-temperature result.
Our approachDesigns are checked on the hottest day they will ever see, not the average one.
05Q2
Shape decides survival
Where a magnet operates on its curve is set by its own geometry and the circuit around it. Get the load line wrong and a magnet can demagnetize itself with no outside help.
Our approachIn a switchable design the released state is the harshest. Margin is checked there first.
061 lever
Models have to be checked
Simulation is an instrument, and instruments drift. Change two parameters at once and nothing is attributable; read a point where an average was meant and the answer is wrong by design.
Our approachSweeps move one factor at a time, and every result is checked before it is believed.
Every term below is a decision a magnetic design can get right or wrong. That’s the work.