Nearly every customer who benches a PU system for the first time sends me the same message: empty, the float sat at 48 mm; they screwed their product on, and the gap visibly shrank. Is the unit weak? No. The gap is doing exactly what the physics tells it to do. Here is the mechanism, with real bench numbers.
First, why this costs almost no power
Our current low-power PU architecture splits the job in two. The ring magnet carries the weight — the whole gravity balance comes from the ring. Inside the ring sit four coil groups, and they only correct the position when the float drifts. That split is the core of the design: once the float is stable, the slow-servo MID loop searches for the lowest-power hover point and trims the coil current down to almost nothing. A balanced float needs no ongoing push, so the system runs at a few watts and stays only slightly warm.
Why does that matter for the gap question? Because it means the hover height is set by magnet geometry alone — the coils are not spending force to hold the float up, they only nudge it back to center. So the gap is a pure balance problem.
The ring magnet's one rule
The push force from the ring is strong close to the ring and weak far from it. A float settles where the upward push exactly equals the weight: Fring(h) = m·g. That height is the float gap (same thing people mean by “levitation distance”).
For a given system, the ring is fixed, so the field is fixed. A light object finds its balance far out, where the weak field is already enough — and you see a tall gap. A heavy object needs more force, and more force only exists closer to the ring — so it settles lower and the gap shrinks. Nothing is lost; the load simply spent it.
Measured on the bench: PU2000G
Real numbers from a PU2000G unit:
| Payload | Float gap | Gap cost per added kg |
|---|---|---|
| empty | 48 mm | — |
| 400 g | 40 mm | 20 mm/kg |
| 1 kg | 30 mm | ≈16.7 mm/kg |
| 2 kg (full) | 20 mm | 10 mm/kg |
Look at the last column. The first kilogram costs about 18 mm of gap; the second kilogram costs only 10. The decay slows down as you load up — because near the ring the field gradient is steep, so each extra kilogram only needs a short move down to find the extra force. Empty-space millimeters are expensive; loaded millimeters are cheap. Worth remembering when you size a display.
The sweet spot: load it to about half
Levitation exists to be seen — a taller visible gap is the whole point of the effect. From the bench data and from field projects, our recommendation is to size for roughly 50% load: a 1 kg product on a PU2000G still holds a 30 mm gap, stability margins are wide, and you are not paying for a system twice the size to win back millimeters. That is the best trade of viewing impact against cost.
Need more air anyway? The two honest moves are stepping up one system size, or taking weight out of the float — we covered both in the castle & building model display guide.
This article consolidates engineering practice. The formulas and measured figures are for design reference — not product specifications or delivery commitments.
Related reading: PU300G 300g push-up system, Why Our Push-Up Systems Need No Positioning Aid, Next-Generation Push-Up Levitation. Sizing a project? Talk to the engineering team.