iPad won't turn on? What board-level PMIC repair actually involves

The complaint is almost always the same: plug it in, nothing happens. No Apple logo, no charging animation, sometimes the case gets faintly warm near the charging port and sometimes it doesn't warm up at all. People assume it's the battery, because a flat battery is the only power problem most people have ever experienced. On a laptop or phone that's a fair guess. On these boards it's usually wrong.

What actually happens when you press the power button

An iPad doesn't just connect a battery to a processor and switch it on. Power comes in through the charging IC, then gets handed to the PMIC — the power management IC — which is responsible for sequencing a whole set of separate voltage rails to the CPU, the NAND storage, the display driver and various sub-systems, each one in a specific order and within a specific timing window. If any single rail doesn't come up cleanly, the SoC never gets a valid power-good signal and refuses to start the boot process at all. From the outside that looks exactly like "the board is dead." From the inside, 90% of the board is usually fine — one rail just isn't behaving.

Common causes we actually see on the bench

• Liquid ingress corrosion. Even a small amount of moisture that's dried out can leave conductive residue under the PMIC or the USB/charging IC, creating a partial short that only shows up under load.
• A shorted rail from a swollen or damaged battery. A battery that's degraded internally can pull a rail down hard enough that the PMIC shuts everything off as a protection response.
• Physical shock damage. A drop can crack a coil or lift a ball under a chip without leaving any visible mark on the case.
• Damage from a previous repair. We see boards that have already had a battery or screen replaced elsewhere, where a connector was reseated with a short across it, or excess flux was left bridging two pads.

How we actually diagnose it — not guess

The board goes on a bench power supply with the battery disconnected, current limited low, and we watch what the board draws the instant power is applied. A dead short shows as the current limit hitting almost immediately with the voltage collapsing to near zero. No draw at all usually points to an open somewhere before the rail even starts. A moderate, steady draw that's higher than a healthy board's baseline usually means something downstream is partially shorted and slowly heating up.

From there it's a thermal camera or, just as often, a finger and some isopropyl alcohol — a shorted component will get warm faster than everything around it, and alcohol evaporates visibly faster off a warm spot. Combined with the board's schematic and known-good rail voltages, that narrows the fault down to a specific rail and usually a specific component, not "somewhere on this board."

Why this isn't a reflow-and-hope job

These are dense, multilayer boards with components a few millimetres apart, all rated for very tight thermal profiles. Reflowing or replacing a PMIC without the right preheat, the right air temperature and the right no-clean flux risks lifting pads on the very next component over, which turns a repairable board into scrap. Static discipline matters too — these ICs are sensitive enough that an ungrounded iron or a static discharge from a chair can damage a brand new replacement part before it's even soldered down.

The actual repair, once the fault is isolated, is straightforward in principle: remove the failed IC under a microscope, clean and reballed pads, place and reflow the new PMIC to the correct profile, then re-test rail voltages under the bench supply before the board goes back into the case. That last step matters — it's the difference between a repair that's actually verified and one that just looks finished until the customer gets it home.

This is the same reason we diagnose before we quote on anything that comes through the door. A battery swap that doesn't fix a PMIC fault wastes the customer's money and delays the actual repair. Board-level work costs more than a parts swap because it's solving the fault instead of guessing at it.