Every maker eventually designs a magnet pocket. It goes into the lid of an enclosure, the back of a wall-mounted tool holder, the seam of a cosplay pauldron. And nearly everyone designs the same pocket: a shallow cylindrical recess sized for a small disc magnet, printed slightly undersize so the magnet wedges in. It works beautifully on day one. On day three hundred, the magnet is rattling around in a parts bin. A July 19 piece by Maya Posch at Hackaday, drawing on a Slant 3D video, catalogues seven ways to get magnets into FDM parts without glue — and the most useful thing in it is not any single technique but the explanation of why the default one degrades.
The failure mode: thermal creep
Thermoplastics under sustained compressive load do not hold their dimensions. A press-fit magnet is exactly that: the pocket walls are elastically deformed outward, and the restoring force is what grips the magnet. Over weeks and months — accelerated by every warm afternoon in a garage, a car, or a sunlit window — the polymer relaxes into the deformed shape. Elastic strain becomes plastic strain. The grip goes away. Hackaday points to thermal creep as the reason a simple friction fit will loosen over time, which makes it a poor choice as the sole retention strategy.
That is worth sitting with, because creep is invisible during the only test most of us run. You push the magnet in, it feels tight, you call it done. The failure is scheduled for a date well past the point where you would connect it to the design decision.
The approaches, and what they actually buy you
The roundup's methods are best read as a spectrum running from "friction fit, improved" to "friction fit, replaced."
Better friction fit. If you are going to rely on interference, use cylindrical magnets rather than flat ones — Hackaday notes this increases grip with the thermoplastic material. A taller magnet presents more wall area for the pocket to bear on, and gives creep more material to eat through before the hold fails. Seat it with an arbor press rather than, as the article puts it, human primate hand power: square, consistent insertion force lets you use a tighter pocket than you could manage by hand without cracking something.
Side slots. Instead of embedding the magnet inside the part mid-print, cut a slot into the side and slide the magnet in afterward. This avoids the mid-print logistics entirely — no pause, no timing, no reaching into a hot machine — but it carries a cost Hackaday names directly: you end up with a plastic barrier between the magnet and its target. Magnetic force falls off steeply with distance, so a wall added for convenience is a wall that weakens the hold.
Press-fit past extra material. Here the geometry does the retention work. You design in extra plastic — a lip, or an undersized throat — and use the arbor press to force the magnet past it. Once seated, the magnet is mechanically captured and prevented from falling back out, regardless of what creep does to the walls. This is the answer to the creep problem for most enclosure and lid applications.
Spherical magnets. Ball magnets automatically orient themselves to an opposing magnet, which sidesteps the polarity problem and tolerates some misalignment between mating parts. Hackaday notes they can be locked in using a similar press-fit method, so retention is not the trade-off.
Pair a magnet with plain steel. The best trick in the list is the one that halves your magnet count. Hackaday's advice here is blunt: never use two magnets in a magnetic lock — use a cheaper ball bearing or a similar plain metal part on the other side instead. It costs less, and it is immune to the reversed-polarity mistake, because there is no polarity on the steel side to get backwards.
Running underneath all of this is a quieter benefit: the video's framing is seven ways to install magnets without resorting to glue. Skipping adhesive removes the cure time, the fogging, and the fingertip bonding.
The slot math and the pause workflow
Design-side technique only pays off if the mid-print insertion actually goes cleanly, and printer manufacturer Sovol's guide to using magnets in 3D printed models is unusually specific about the numbers. Slot depth should match magnet thickness — if the magnet is 2mm thick, Sovol says, the slot should also be 2mm deep. The slot should be made slightly bigger than the magnet so it fits easily, with a snug fit as the target; a loose slot lets the magnet shift after insertion, and a shifted magnet is a nozzle strike waiting to happen when the print resumes. Sovol also calls for at least three layers of plastic above the magnet, for strength and appearance.
The pause itself is handled either through your slicer's "pause at height" option or by inserting custom G-code such as M601. The layer arithmetic, per Sovol, is straightforward: pause height equals slot start height plus magnet thickness. If your slot starts at 0.4mm and the magnet is 2mm thick, you pause at 2.4mm.
Mark polarity before you are standing over a paused printer — Sovol suggests using a marker or tape to label the north and south sides. Reversed polarity is not a tolerance problem you can sand out; it is a reprint.
If you would rather glue after all, Sovol recommends super glue or epoxy: apply a small amount and press the magnet firmly into place, having lightly sanded the magnet and slot first so the glue sticks better. The other listed pitfalls are worth a pre-flight check. For corrosion, choose corrosion-resistant magnets such as nickel-coated neodymium, and for anything that will live outdoors, use waterproof coatings or design an enclosure that seals the magnet in. For warping around the embedded metal, lower the cooling fan speed near the magnet; Sovol also notes PLA warps less than ABS or PETG, which makes it the easier material to embed into.
Bottom line for makers
Treat friction fit as a prototyping convenience, not a production strategy. If the part will be handled repeatedly, live somewhere warm, or ship to someone who cannot re-seat a magnet, design mechanical capture into the geometry — a lip the magnet gets pressed past, or a slot that traps it — and let interference be a bonus rather than the load path. Then take the free win at the end of the list: replace the second magnet with a ball bearing. You will spend less, and you will never print a part backwards again.
