“We’ll just power it from a phone charger — it’s only two watts, right?”
We hear this on almost every OEM enquiry. The two-watt figure is correct. It is also completely irrelevant to whether a USB-C charger can run the product.
Magnetic levitation is not an average-power load. It is a peak-power load with an extremely high crest factor. A supply that looks generous on its label can still fail at the one moment that decides everything: catching and locking the floater.
Four engineering reasons, with the math.
1. What a USB-C port can actually deliver
Start with the connector. The Type-C VBUS contact pair is rated for 5 A continuous. Not 6 A, not 8 A — 5 A is the physical ceiling, and no cable or negotiation can exceed it. Every step above 100 W comes from raising voltage, not current.
Without PD negotiation, a Type-C source is limited to:
| Condition | Output | Power |
|---|---|---|
| Default USB 2.0 | 5 V / 0.5 A | 2.5 W |
| Default USB 3.x | 5 V / 0.9 A | 4.5 W |
| Type-C 1.5 A | 5 V / 1.5 A | 7.5 W |
| Type-C 3 A | 5 V / 3 A | 15 W |
With USB PD, the ceiling rises:
| Specification | Max voltage | Max current | Max power |
|---|---|---|---|
| PD 3.0 (SPR) | 20 V | 5 A | 100 W |
| PD 3.1 (EPR) | 48 V | 5 A | 240 W |
The port is not the constraint. The constraint is which voltage the adapter is willing to negotiate.
2. 12 V is an optional, deprecated profile — not a guaranteed one
This is the part most spec sheets leave out.
USB PD 1.0 was designed around USB-A, and it defined 12 V. When the industry moved to Type-C and published PD 2.0, 12 V was replaced by 15 V. Only four fixed voltages are official PDOs today:
5 V · 9 V · 15 V · 20 V
12 V survives only as a legacy, optional profile. A compliant adapter may omit it entirely — and most modern ones do.
PD’s Power Rules make the tiers explicit:
| Rated adapter power | Voltages it must offer |
|---|---|
| > 15 W | 5 V, 9 V |
| > 27 W | 5 V, 9 V, 15 V |
| > 45 W | 5 V, 9 V, 15 V, 20 V |
| any rating | 12 V is never mandatory |
A public, checkable example: Framework’s own 60 W USB-C adapter, sold with their laptops, refuses to output 12 V. Ask it for 12 V and it hands back 9 V. Ask for 15 V or 20 V and it works fine. Framework’s engineers confirmed on their community forum that this is fully compliant — because 12 V is not one of the required profiles.
Apple’s 20 W and 30 W adapters behave the same way. So do most 65 W and 100 W GaN bricks.
In our own incoming inspection of retail PD adapters, fewer than half expose a 12 V fixed PDO, and above 65 W the number approaches zero. The higher the power rating, the more likely the designer skipped 12 V — 15 V and 20 V already cover the laptop use cases that high-power adapters are built for.
That gives us the first structural problem:
You cannot print “works with any USB-C charger” on a levitation product, because a compliant charger is allowed to omit the voltage you need.
3. Our load has a 15:1 crest factor
Now the electrical characteristic that turns an inconvenience into a design failure.
A levitation module idles cheaply and spikes hard:
- Steady levitation: 1.5–2.5 W (≈ 0.13–0.21 A at 12 V)
- Catch and lock: 12 V × 3 A = 36 W, held for tens to hundreds of milliseconds
The crest factor is 36 ÷ 2.4 ≈ 15:1. For comparison, a 65 W laptop pulling a 90 W transient has a crest factor of about 1.4. A phone is flatter still.
Why 3 A? Because the coil is an inductor, and the controller drives it at full rail voltage to build flux quickly: i(t) = (V/R)(1 − e−t/τ), with τ = L/R. For a coil with roughly 4 Ω of DC resistance, a 12 V rail commands 3 A once t is a few time constants in. That is not a fault condition — it is the intended operating point during catch.
And the force scales with the square of the current. For a simplified magnetic circuit with air gap lg: B = μ₀Ni/lg, and F = B²A/(2μ₀) = μ₀N²Ai²/(2lg²), which reduces to:
F/F₀ = (i/i₀)² = (V/V₀)²
Force is quadratic in voltage. A 10% droop costs 19% of your lifting force. A 25% droop costs 44%.
4. What a sagging supply actually does to the product
Put the two formulas together and the failure mode writes itself.
| Rail actually delivered | Current vs nominal | Lifting force |
|---|---|---|
| 12.0 V | 100% | 100% |
| 11.4 V (−5%) | 95% | 90% |
| 10.8 V (−10%) | 90% | 81% |
| 9.0 V (PD fallback) | 75% | 56% |
| 5.0 V (no PD) | 42% | 17% |
At 9 V the system retains barely half its design force. It may still appear to work on a light floater under bench conditions — and then fail in the field the moment someone bumps the table, swaps in a heavier display object, or the room warms up and coil resistance drifts upward.
At 5 V it does not levitate at all. The controller sees the floater, drives the coil, gets 17% of the force it needs, and gives up.
The symptoms our support team hears, in order of frequency:
- The floater shivers or orbits instead of locking.
- It locks, then drops when a hand approaches.
- It works on the bench and fails on the customer’s desk.
- It works with one charger and not another — same model, different batch.
None of these look like a power problem to the user. All of them are.
5. The failure mode that turns a nuisance into a warranty claim
There is a worse outcome than “it won’t levitate”.
PD is a negotiated contract, not a fixed rail. When load current steps to 36 W, the adapter’s output sags. If that sag trips over-current or under-voltage protection, the adapter hard-resets the contract. VBUS collapses, the PD controller renegotiates from scratch, and the rail returns at 5 V until a new contract is agreed.
On a levitation system that sequence is: 36 W step → rail sag → adapter protection trips → hard reset → VBUS collapses → coil current vanishes → floater falls.
Then the recovery attempt begins at 5 V, which cannot lift the floater, so the controller keeps retrying, so the adapter keeps tripping.
This is the same failure we described in our article on battery power: once the floater drops, the damage is mechanical and irreversible. A 0.5 kg floater falling 30 mm carries about 0.15 J; absorbed over 1 mm of housing deformation that averages roughly 150 N of impact force. On heavier systems the floater is considerably more massive. One hard drop can crack the housing or shift the magnet stack out of alignment — after which the unit may never balance again.
A marginal supply does not produce a marginal fault. It converts a recoverable electrical problem into an irreversible mechanical one.
6. Cables, connectors, and the global adapter lottery
Even when the adapter does offer 12 V, the rest of the chain has to hold up.
Cable drop. At 3 A, conductor resistance stops being academic: ΔV = 2Iρl/A. For a 1 m cable (2 m round trip) in copper:
| Conductor | R per metre | Drop at 3 A | Rail left | Force left |
|---|---|---|---|---|
| 20 AWG | 33 mΩ | 0.20 V | 11.80 V | 97% |
| 24 AWG | 84 mΩ | 0.50 V | 11.50 V | 92% |
| 28 AWG | 212 mΩ | 1.27 V | 10.73 V | 80% |
A thin, long, bundled “charging cable” quietly costs a fifth of your lifting force before the power even reaches the board.
Current rating. Above 3 A a cable must carry an E-Marker declaring its capability. Below 3 A there is no such interlock — a cable rated for 2 A will happily be asked for 3 A, heat up, and droop.
A-to-C is a dead end here. A USB-A port has no CC pin, so PD cannot run at all. An A-to-C cable falls back to Quick Charge or BC1.2, and QC 2.0’s ladder is 5 / 9 / 12 / 20 V — which does give you 12 V, but only from an A-port adapter and only within QC’s current limits. It is a workaround, not a design.
And then there is the adapter lottery. The same “65 W USB-C PD” label covers:
- adapters with 12 V and adapters without;
- single-port adapters and multi-port adapters that redistribute power when a second device is plugged in — plug in a phone and the levitation rail can drop a tier mid-levitation;
- adapters that thermal-derate after thirty minutes of continuous load;
- adapters that pass certification and adapters that merely print “PD” on the housing.
For a product that runs continuously and demands a repeatable 36 W transient on command, this is not a supply chain. It is a lottery with a warranty attached.
7. What we recommend instead
| Scenario | Recommendation |
|---|---|
| Any levitation product | Fixed 12 V DC adapter, barrel or locking connector, rated ≥ 12 V / 3 A, specified at 12 V / 5 A (60 W) for headroom |
| Customer requires a USB-C receptacle | Use Type-C as a physical connector only, fed by a fixed 12 V supply — do not design the product around PD negotiation |
| PD is unavoidable | Negotiate 15 V or 20 V (the mandatory tiers), then buck down to 12 V on-board. Never rely on the 12 V PDO |
| Battery-powered variants | See our separate article — workable for commercial, not recommended for consumer |
The 15 V route deserves a note. If a customer insists on a C-to-C cable, the defensible design is to negotiate 15 V — mandatory above 27 W — and regulate down to 12 V with an on-board buck converter. It costs a converter and roughly 8–10% efficiency, but it is deterministic, and it removes the dependency on an optional profile.
That is a different product from “plug in any USB-C charger”, and the spec sheet has to say so.
The one-line version
A USB-C charger is designed to fill a battery slowly and tolerate interruption. A levitation system needs a fixed rail that never blinks. Those are opposite requirements, and no amount of negotiation fixes it.
If an industrial design absolutely requires a USB-C receptacle, we will build it — with a fixed 12 V supply behind it and the limits printed on the label. What we will not do is ship a levitation product that depends on a charger having a voltage the specification never required it to have.
Frequently asked questions
Can a USB-C PD charger output 12 V?
Sometimes. 12 V is a legacy optional PDO. PD 2.0 replaced it with 15 V, and only 5 / 9 / 15 / 20 V are mandatory tiers. Many compliant adapters — including Framework’s 60 W and Apple’s 20 W and 30 W — omit 12 V entirely.
Our module only draws 2 W. Why does 36 W matter?
Because 36 W is the catch-and-lock transient, and the coil is driven at full rail voltage to build flux fast. Average power determines the electricity bill. Peak power determines whether the floater stays up.
What happens if the adapter falls back to 9 V?
Lifting force scales with the square of voltage, so 9 V leaves you about 56% of design force. That is enough to look like it works on a bench and not enough to survive a disturbance.
Why not negotiate 20 V and step it down?
You can, and it is the right answer if PD is mandatory: 20 V is required above 45 W, and a buck converter gives you a clean 12 V. You pay for a converter and some efficiency, but you stop depending on an optional profile.
Does a USB-A to USB-C cable work?
Not for PD — USB-A has no CC pin. It falls back to Quick Charge, which does provide 12 V, but only from A-port adapters and only within QC’s limits. It is a compatibility path, not a design.
Do you build USB-C levitation products if a customer insists?
Yes. We build the receptacle, put a fixed 12 V supply behind it, and print the limits in the manual. We just won’t claim it works with every charger on the market, because it doesn’t.
Engineering notes from our bench. Figures are typical measured values on our modules — for design reference, not product specifications or delivery commitments.
Related reading: Why We Don’t Recommend Battery Power · Why Lamp Power Is Small. Specifying a module? The Global Adapter Guide module selection guide covers supply requirements per model, or talk to our engineering team.