Inrush when hot-plugging
Why do connectors spark and supplies trip when you hot-plug a board with large input capacitance? What do you do?
| Level | ●● onsite technical |
| Track | Power |
| Really testing | Inrush current and its control. |
Going deeper
Extra explanation that is not in the book: worked numbers, where people go wrong, and how to say it in the room.
Worked example — how big the spike is. 470 µF at the input, plugged into 12 V with 50 mΩ of cable and contact resistance: up to 240 A, limited in practice only by the cable inductance.
Worked example — the controlled ramp. Ramp the same 470 µF from 0 to 12 V in 10 ms: I = C·dV/dt = 470 µF × 1.2 V/ms ≈ 0.56 A. The series FET drops 12 V → 0 V over the ramp at that current, an average of about 3.4 W for 10 ms. Check that point on the FET's SOA curve.
The ringing problem. 1 µH of cable with 10 µF of ceramic rings at about 50 kHz, with a characteristic impedance √(L/C) ≈ 0.3 Ω. A bulk electrolytic or polymer cap with about that much ESR, and several times the ceramic capacitance, damps it.
Where people go wrong
- An NTC on a product that gets unplugged and plugged back quickly — the NTC is still hot.
- Forgetting that the downstream converter starts drawing current during the ramp.
Errata
No corrections known for this question. All errata
Ask about this question
Stuck on a step, or think something is wrong? Send it with PI-05 in the subject. Good questions get answered on this page, and corrections go into the next printing.