A laptop that lights up, spins its fan for a moment and then switches off is not dead — it is trying to start and giving up. Something in the power-on sequence fails, and the board's protection or the embedded controller shuts everything down. The task is to catch which step fails before it disappears, which is harder than a fully dead board because the evidence lasts only a second.

What is happening inside

When you press the power button, the board switches on its rails in stages: memory rails, then system rails, then the CPU core rail. Each stage is usually confirmed by a power-good signal. If a rail does not come up, comes up at the wrong voltage, or a power-good never arrives, the EC or the chipset waits for a short time and then aborts the sequence, turning everything off again. Some boards retry a few times; some give up after one attempt.

That means the shutdown is a symptom of a specific failed step. The rest of the board may be perfectly healthy.

The usual causes

CauseTypical clue
Short on a rail that switches on at power-onCurrent spikes to a high value, then drops
Regulator failing to start (bad enable, damaged part)One rail missing while others appear briefly
Power-good not reportedAll rails appear, then everything shuts down
Battery or adapter problemsBehaviour changes with battery removed or another adapter
Memory or CPU faultsRails stay up a little longer before the shutdown
Corrupted BIOS or EC firmwareRepeating on-off cycles, sometimes with a pattern
Overheating protectionShutdown after several seconds or minutes, not instantly

Step 1: rule out the external causes

  1. Disconnect the battery and try with the adapter only.
  2. Try a known-good adapter of the correct rating.
  3. Remove memory modules and try one known-good module.
  4. Disconnect peripherals: keyboard ribbon, touchpad, display cable, daughterboards. A shorted keyboard or a damaged USB board can trigger a shutdown.

If the behaviour changes with any of these, you have found the area without touching the motherboard.

Step 2: watch the current draw

A bench supply or a USB-C power meter shows what the board tries to do in that one second:

  • A spike to the current limit, then off — almost always a short on a rail switched on at power-on. The system rail is fine in standby but a later rail collapses.
  • A modest rise, a pause, then off — rails come up but something is missing: a power-good, a clock, a reset.
  • Repeating cycles — the EC retries. Count the cycles; some boards retry a fixed number of times.

Set the supply's current limit sensibly so a short cannot do further damage — see Bench Power Supply for Laptop Repair: How to Inject Voltage.

Step 3: check rails for shorts with power off

Many power-on shutdowns are caused by a rail that is shorted but only matters after the button press. Such a short is invisible in standby, which is why a board that looks healthy while off can still fail every attempt to start. With the board unpowered, measure the resistance or diode-mode reading to ground of the main switched rails:

  • memory rails (VDDQ, VTT, VPP);
  • system rails (+3VS, +5VS, +1.05VS and similar);
  • the CPU core rail and other processor rails;
  • the GPU core rail on boards with a discrete GPU.

Compare with a known-good board or the reference values in the boardview. A rail reading close to zero is your first suspect. Finding the exact part is covered in How to Find a Short Circuit on a Laptop Motherboard.

Step 4: catch the failing rail

If no rail is shorted, you need to see which one fails during that short window:

  1. Use an oscilloscope with two or three channels on suspect rails and trigger on the power button signal or on the first rail that comes up. You will see which rail rises and which never does.
  2. Use a meter with min/max capture if you have no scope; it records the highest value reached by each rail during the attempt.
  3. Hold the board in its "on" state where the design allows. On some boards, keeping a specific signal asserted prevents the shutdown long enough to measure — only do this if the schematic makes the logic clear.

Step 5: check the power-good chain

When every rail appears but the board still shuts down, follow the power-good signals: each regulator's power-good output, the combined power-good to the chipset, and the platform reset. A power-good held low by a leaky part, or never generated because a regulator's feedback is wrong, aborts the sequence even though the voltages look fine. See Power Good Signals and Why They Matter.

Example: the shorted capacitor on a memory rail

A laptop lights its power LED for about a second after the button press and then goes dark. On a bench supply, the current jumps to the limit at the moment of the press and drops back.

In standby, the system and always-on rails are present and healthy. With power off, the memory VDDQ rail reads only a few ohms to ground, far below a known-good board. The rail is fed by a buck regulator that starts only after the sleep signals release — which is why standby looked normal.

Using voltage injection at a low voltage with a current limit, a thermal camera shows one small capacitor warming up next to the memory slot. Removing it brings the rail back to a normal reading, and the board now powers on and boots. A replacement capacitor of the same value and voltage rating finishes the job.

Common mistakes

  • Chasing the CPU first. Short rails and missing power-good signals are far more common than a dead processor.
  • Measuring only in standby. The failing rail usually exists only during the attempt.
  • Ignoring peripherals. A shorted keyboard or touchpad board can pull a rail down and trigger exactly this symptom.
  • Repeating power-on attempts with a hard short. Each attempt heats the shorted part and stresses the regulator. Find the short first.

FAQ

Is a laptop that turns on and off immediately worth repairing?

Usually yes. The most common cause is a single shorted capacitor or a failed regulator component on one rail — cheap parts, once found.

Can a bad battery cause it?

Yes. A battery with a failing cell or a faulty protection circuit can collapse the system rail under load. Test with the battery disconnected.

Why does it work with the adapter but not on battery?

The battery path and the adapter path differ. A weak battery, a faulty battery MOSFET or a charger configuration issue can stop the board when it relies on the battery alone.

Can overheating cause an immediate shutdown?

Not usually immediately. Thermal shutdowns happen after the processor has run for a while. An instant shutdown points at power, not heat.

Do I need an oscilloscope?

For shorts, no — a meter in resistance or diode mode finds them. For a rail that rises and falls within a second, an oscilloscope or a meter with min/max capture makes the job much easier.

Bottom line

A laptop that turns on and off immediately is aborting its power-on sequence. Rule out the battery, adapter, memory and peripherals, read the current draw, check the switched rails for shorts with power off, and if none are shorted, catch the failing rail or power-good during the attempt. The power sequence page of the schematic tells you which step comes next.