Almost every important rail on a laptop board — the always-on rails, memory supply, chipset rails, processor and graphics cores — is made by a buck converter. It takes the higher system voltage and turns it into a lower one efficiently, by switching rather than burning off the difference as heat. Once you understand how a buck converter works, its parts on the board make sense, its failure modes become predictable, and its measurements tell you much more. This guide explains the principle and what it means for repair.
The principle
A buck converter has four essential parts:
- A high-side switch that connects the input to the switch node.
- A low-side switch that connects the switch node to ground.
- An inductor from the switch node to the output.
- Output capacitors that smooth the result.
The controller turns the high-side switch on for a short time, then off while the low-side switch turns on, many hundreds of thousands of times per second. The inductor smooths the pulses into a steady current, and the capacitors smooth the voltage. The fraction of time the high side is on — the duty cycle — roughly equals the output voltage divided by the input voltage. The principle is described under buck converter, and the arithmetic can be explored with Buck Converter Calculator (Duty Cycle and Inductor Ripple).
Synchronous buck converters
Older and simpler designs used a diode instead of the low-side switch. Laptop converters almost always use a MOSFET there instead — a synchronous design — because a MOSFET wastes much less energy than a diode at high currents. The two MOSFETs may be separate parts, a dual MOSFET in one package, or an integrated power stage that includes the gate driver. Testing them is covered in MOSFETs on Laptop Motherboards: How to Test Them.
The bootstrap supply
The high-side N-channel MOSFET needs a gate voltage above the input voltage to turn fully on. A small capacitor between the boot pin and the switch node — the bootstrap capacitor — provides it, recharged every time the low-side switch is on. A failed bootstrap capacitor or its diode stops the high side switching properly; the rail is then missing or weak. The part is described under bootstrap capacitor.
How the controller decides
Controllers regulate the output in different ways:
| Control style | How it works | Notes |
|---|---|---|
| Voltage mode | Compares output with a reference and adjusts duty cycle | Classic, needs compensation network |
| Current mode | Also senses inductor current each cycle | Better response, built-in current limit |
| Constant on-time | Fixed on-time, varies frequency | Common in laptops for fast response |
The schematic's controller page and its datasheet tell you which one is used; it matters mostly when reading the compensation and frequency-setting parts. How to read those pages is covered in Reading Voltage Regulator Pages in Laptop Schematics.
Light-load behaviour
At light load — a laptop in sleep or idle — switching at full frequency wastes energy. Most laptop controllers switch to a power-saving mode, skipping pulses or lowering the frequency. On an oscilloscope the switch node then shows bursts of pulses with gaps; on a meter nothing seems different. Some boards set the mode with a pin; a wrong mode setting can increase idle consumption.
Soft start
When enabled, a converter raises its output gradually rather than instantly, to limit inrush current into the output capacitors and avoid tripping protections. A soft-start capacitor or an internal timer sets the ramp. A shorted soft-start capacitor can keep the output at zero even with a valid enable. The concept is described under soft start.
Protection features
| Protection | What triggers it | What you see |
|---|---|---|
| Overcurrent | Load current above a limit | Output drops, often with hiccup restarts |
| Short circuit | Output near zero while switching | Hiccup mode: repeated restart attempts |
| Overvoltage | Output above a threshold | Controller shuts down, sometimes latches |
| Undervoltage | Output stays below a threshold | Shutdown or hiccup |
| Overtemperature | Controller too hot | Shutdown until it cools |
Hiccup mode is especially recognisable: a multimeter on the output shows a small, flickering voltage instead of a steady value. It is a strong hint that the output is shorted or overloaded.
What you can measure
| Measurement | Healthy result |
|---|---|
| Input | System rail present |
| Enable | At its active level |
| Controller supply and internal regulator | Present |
| Switch node (meter, DC) | Roughly the output voltage, because the meter averages the pulses |
| Output | The expected voltage, steady |
| Power-good | High once the output is in range |
An oscilloscope on the switch node shows the actual square wave and its duty cycle, which confirms the converter is really switching.
The inductor and the capacitors
The inductor's value and current rating, together with the output capacitors, decide how much ripple remains on the rail and how well it copes with sudden load changes. Replacing an inductor with one of a different value, or leaving out output capacitors, can make a converter unstable: the rail may oscillate, fail under load, or trip protection. When repairing, use the same inductor value and a current rating at least as high, and restore every output capacitor. Polymer bulk capacitors on converter outputs are also a wear item on older boards; they slowly lose capacitance, which raises ripple long before anything fails outright.
Multi-phase converters
Processor and graphics cores need far more current than one converter can handle, so several buck converters — phases — run in parallel, each with its own power stage and inductor, switching in turn. A multi-phase controller balances the current between them. Diagnosis works phase by phase: compare switch nodes and power stages between phases, and look for one that runs hotter or reads differently from its siblings.
Common failures
- Low-side MOSFET shorted — switch node shorted to ground; output missing; the controller may refuse to start or hiccup.
- High-side MOSFET shorted — switch node shorted to the input; the output can rise to the input voltage and damage everything on the rail. Treat any rail measuring close to the system voltage as an emergency.
- Output capacitor shorted — output shorted to ground; hiccup mode.
- Bootstrap capacitor failed — high side does not switch; output missing or weak.
- Feedback resistor drifted — output at the wrong voltage.
- Controller damaged — no switching despite correct supply and enable.
Example: a hiccuping memory rail
A memory supply rail reads a flickering few hundred millivolts on a meter. The regulator's enable and supply are correct. With power removed, the output reads under an ohm to ground, while the switch node reads normally. The regulator is in hiccup protection, trying to start into a short on its output. Injecting a low voltage into the output warms a ceramic capacitor next to the memory. With it replaced, the rail comes up steady at its nominal voltage.
FAQ
Why does the switch node read about the output voltage on a meter?
The meter averages a fast square wave whose average equals the output voltage.
What is hiccup mode?
A protection mode where the controller repeatedly tries to start into a short or overload. The output flickers at a low voltage.
Why use a MOSFET instead of a diode for the low side?
A MOSFET wastes far less power at high currents, improving efficiency.
Can a buck converter output the full input voltage?
If the high-side MOSFET shorts, yes. That can damage everything on the rail.
Why do some converters skip pulses at idle?
To save power at light load.
Bottom line
A buck converter is a high-side and low-side switch, an inductor and capacitors, driven by a controller that adjusts the duty cycle. Check input, enable, controller supply, switch node, output and power-good. Recognise hiccup mode as a sign of a short, treat a rail at input voltage as an emergency, and suspect MOSFETs, capacitors, bootstrap parts and feedback resistors before the controller.