Gaming laptop shuts down under load: heat, current or logic board

The machine runs perfectly for half an hour. Then, in the middle of a match or an export, it cuts out cold: no warning, no blue screen, as though somebody pulled the cable. A while later it boots again quite happily.

That symptom is one of the most common on high-end laptops — MSI, Legion, ROG, Victus, Omen, Alienware, Predator — and it has three very different possible causes. It is worth separating them before spending a euro, because the fix for one does nothing for the other two.

An abrupt shutdown is a protection, not a fault

First of all, to take the drama out of it and put some precision in.

When a computer shuts down cold without closing anything, a safety mechanism has almost always acted. The processor has an absolute thermal limit and, on reaching it, cuts the power rather than carrying on. The power circuit has the same: if a voltage sags below a threshold, everything switches off instead of running badly.

So the shutdown is telling you something has gone out of range. The useful question is not “why does it shut down” but what went out of range. And there are only three answers.

First cause: the heat is not getting out

This is the most common one and the most rewarding, because it nearly always has a fix.

A gaming laptop packs the power draw of a desktop into three centimetres of thickness. The whole system depends on heat travelling from the chips out to some tiny radiators, and that journey degrades in three ways at once:

• Dust packed into the radiator. Not a layer of dirt: a felt blanket bonded to the fins, made of hair, ash and dust. The fan keeps spinning exactly as before and the air no longer gets through.
• Spent thermal paste. Between the chip and the heatsink sits a very thin layer of paste. Heating and cooling cycles gradually push it out towards the sides — the pump-out effect — and where there was paste there ends up being air, which is the opposite of what was needed. There is no fixed lifespan: it depends on the machine, the compound and your use.
• Hardened thermal pads. The graphics memory and the power transistors do not touch the heatsink: they reach it through silicone pads that dry out and crack over the years.

The sign that confirms this cause is a progressive climb before the cut: the fans spool up, performance drops, the chassis gets hot and then it shuts down. If you have a temperature monitor, you will watch the curve rise. If the cut is instant and the machine is lukewarm, this is not your cause.

Second cause: the power supply cannot hold the peak

This is the one almost nobody looks at, and on laptops more than three or four years old it is very common.

Modern gaming laptops run on hybrid power. When the CPU and the graphics chip together ask for more than the charger delivers, the machine makes up the difference by pulling from the battery, even while plugged in. That is not a fault: the manufacturers themselves document it as intended behaviour, and it is what allows more manageable chargers without giving up performance.

The problem shows up as that battery ages. An old cell loses capacity, but above all it loses the ability to deliver current in a burst. And then this happens: the machine holds up perfectly on the desktop and shuts down the instant the first serious peak arrives. With the battery at 100 %, plugged in, and with no temperature rise at all.

The same thing happens by another route if the charger isn’t the right one. A brick with less power than the original fits and charges, but it won’t hold the peak. And an original charger that’s already old can have lost part of what it used to deliver, with nothing on the outside to give it away.

Three things that point at this cause:

• It shuts down with the machine lukewarm, with no preceding temperature climb.
• It shuts down sooner with an old or swollen battery, or only shuts down when it is not plugged in.
• The shutdown is instant, with no stutter or drop in performance first.

If on top of that you notice the machine only runs with the charger attached, or that it has stopped charging altogether, the starting point is a different one and it is at a laptop showing no signs of life.

Third cause: the logic board

When the cooling is fine and the external power is fine too, what is left is the circuit that distributes current inside the machine.

The CPU and the graphics chip are not fed directly from the charger’s 19 volts: there are conversion stages on the board — transistors, coils, capacitors and a controller — that step that voltage down and hold it steady while the draw changes a thousand times a second. Those stages run hot and are among the first things to age in a laptop that has gamed hard.

When one of them starts to fail the symptom is exactly the same: the machine holds the desktop, asks for a current peak and the voltage sags just enough for the protection to trip. It also appears through corrosion, a fatigued joint under a chip or a coil that has lost contact.

None of that needs a whole replacement board if it can be avoided: the component is located and that point is rebuilt. It is what the microsoldering lab is for, and on a PC laptop it is more comfortable work than on a Mac, because there is more space between components and schematics circulate with less secrecy.

Liquid metal, which not everyone knows they have

An important note before anybody opens your laptop.

Part of the recent gaming range — ASUS documents this openly on its own support site — does not use thermal paste on the processor but liquid metal: an alloy with far higher thermal conductivity. It is held in place by a barrier that stops it escaping.

What you need to know is that the compound conducts electricity. The manufacturer warns about it without hedging: if a droplet reaches components on the board, it can cause a short circuit. It also attacks aluminium, so not just any heatsink or any procedure will do.

Translated: a strip-down done without knowing that machine uses liquid metal, or a paste change made with the best intentions and the wrong material, turns routine maintenance into a board fault. That is not a reason to leave it unopened; it is a reason for whoever opens it to know beforehand what they are going to find.

What not to do

• Do not keep gaming “until it shuts down”. Every cold cut is a loss of power with nothing closed cleanly, and that is where corrupt file systems come from. Sustained heat does the solder joints no favours either.
• Do not block the fans trying to cool it. Cooling pads with a misaligned grille hurt the airflow more than they help.
• Do not blow compressed air at the fan with the heatsink still fitted. You pack the dust inwards, which is where it is worst, and you spin the fan far above its rated speed, which takes the bearing with it.
• Do not add “a bit of paste” without removing the old. And do not put liquid metal in a machine that never had it, or ordinary paste in one that did.
• Do not swap the charger for a generic higher-wattage one assuming it will hold up. The voltage and the connector have to be the manufacturer’s.
• Do not strip the machine “to look at the dust” if you are not changing the paste. Lifting the heatsink ruins the layer that was left.
• Do not assume it is the graphics chip. It is the first thing said on forums and the last thing it usually turns out to be.

How one cause is separated from another without guessing

The order we follow, in case you want to know what will happen to the machine.

First the shutdown is reproduced with the machine instrumented: CPU and graphics temperatures, real draw at the input, and battery behaviour as the load climbs. That alone separates the thermal scenario from the electrical one without opening anything, because a temperature curve rising until the cut and a voltage sagging on a cold machine look nothing alike.

If it is thermal, the whole cooling system comes off, the radiators are cleaned, the compound appropriate to that specific model is renewed and the pads are replaced. If it is electrical, the machine goes to the board bench: the sagging power stage is measured and the component located. And if what no longer measures up is the charger or the battery, we say exactly that, because it is the simplest of the three repairs and it would be absurd to charge you for a larger one.

We tell you what it has and what it costs before you authorise anything, and what we do carries a 12-month warranty on parts and labour. How we handle a Windows laptop or an Android phone is covered separately, the rest of the workshop guides come off the same bench, the full route is on how the process works, and if you would rather start by telling us exactly when it shuts down, write to us.