A laptop that restarts endlessly — logo, black screen, logo again, or simply fans spinning up and down without ever showing anything — is in a boot loop. Operating system problems cause some loops, but when the loop happens before the operating system even starts, the cause is in the hardware or firmware. The pattern of the loop is the most useful clue.
Software loop or hardware loop?
Answer this first, because the fix is completely different:
- Loops after the manufacturer's logo, during "preparing automatic repair" or similar — usually an operating system or storage problem. Test with a bootable USB stick; if the laptop runs from it, the board is probably fine.
- Loops before any logo appears — power-on aborts and restarts. Firmware, memory or power problems.
- Loops at the logo, before the boot menu works — firmware initialisation fails: memory training, BIOS or ME problems.
- Restarts after minutes of use — heat, power under load or firmware watchdogs.
The rest of this guide covers the hardware and firmware side.
Cause 1: memory
Memory is the most common hardware cause of a loop before the logo. The firmware trains memory at start-up; if training fails, many machines retry, restart and try again.
- Reseat and swap removable modules; test with a single known-good module in each slot.
- Check the memory power rails (VDDQ, VTT, VPP) for correct voltage during the attempt.
- On boards with soldered memory, look for liquid damage around the memory chips and check their rails.
- Watch for beep or blink codes that report memory errors.
Cause 2: BIOS and ME firmware
A corrupted BIOS or a mismatched ME region produces loops that look identical to hardware faults.
- After a BIOS update — a partially written image can loop indefinitely. See Laptop Dead After a BIOS Update: Recovery Methods.
- After a BIOS write with a foreign image — the Intel ME may fail to initialise and cause restarts or long delays; the fix is usually a clean ME region, as described in The Intel ME Region Explained for Laptop Repair.
- CMOS state — a corrupted settings area can cause loops on some models. Disconnecting the coin cell for a minute (with all power removed) clears it.
Cause 3: power-good and regulator problems
If a rail drops briefly during the power-on sequence or during memory training, the platform resets:
- a regulator that starts but cannot hold its output under load;
- a failing output capacitor causing excessive ripple on a core rail;
- a power-good line held near the threshold by a leaky part;
- a weak battery or underrated adapter that sags as load increases.
Use an oscilloscope on the suspect rails and the power-good signals, triggered on the restart, to see which one dips first. The role of these signals is explained in Power Good Signals and Why They Matter.
Cause 4: CPU and GPU
Cracked solder joints under the CPU, GPU or chipset can produce loops that depend on temperature or on pressure on the board. A discrete GPU that fails during initialisation can make a board restart at the same point every time. These are harder to prove; rule out memory, firmware and power first.
Cause 5: peripherals
Faulty peripherals can trigger resets:
- a shorted keyboard or touchpad;
- a failing SSD that hangs the bus during initialisation on some systems;
- a damaged USB device or daughterboard.
Disconnect everything that is not needed to boot and add items back one at a time.
Cause 6: heat
A laptop that restarts after a few minutes of use may be overheating: blocked fans, dried thermal paste, a failed fan or a fan control problem. Check temperatures from the BIOS setup or a live USB system if the machine runs long enough.
Reading the loop with a power meter
A USB-C power meter or a bench supply turns a loop into a pattern you can read. Watch one full cycle and note:
- How high the draw climbs before the restart. A loop that never rises above standby levels never got the main rails up; one that climbs to normal running levels got as far as the CPU.
- How long each attempt lasts. A restart after a second or two points at the power-on sequence; after ten to thirty seconds, at memory training or firmware; after minutes, at heat or a firmware watchdog.
- Whether every cycle is identical. Identical cycles suggest a fixed failure point; varying ones suggest something marginal, such as a rail sagging under load or a cracked joint.
Writing these numbers down for one cycle often narrows the cause before you take a single measurement on the board.
Watchdogs and timed resets
Some restarts are deliberate. Firmware and controllers on the board include watchdog timers that reset the platform when something stops responding, and on some Intel platforms a non-working ME leads to a shutdown or restart after a fixed period. A loop with a very regular interval — the same number of seconds or minutes every time — is a strong hint that a watchdog is acting, which points at firmware rather than failing hardware.
A diagnostic order that works
| Step | Action | What it rules out |
|---|---|---|
| 1 | Boot from a known-good USB stick | Operating system and storage |
| 2 | Test with one known-good memory module | Memory modules and slots |
| 3 | Disconnect peripherals | Keyboard, touchpad, SSD, USB boards |
| 4 | Clear CMOS (coin cell out briefly) | Corrupted settings |
| 5 | Measure rails and power-good during the loop | Power delivery |
| 6 | Read and check the BIOS chip | Firmware corruption, ME problems |
| 7 | Check temperatures and cooling | Overheating |
Example: a loop after a memory upgrade
A laptop loops endlessly after the owner installed a larger memory module: fans spin up, the screen stays dark, the board restarts after a few seconds and repeats. The original module boots fine; the new one boots fine in another laptop.
The technician measures the memory rails during the loop: VDDQ is correct, but VTT sags noticeably with the new, higher-density module installed. The VTT regulator's output capacitor turns out to have lost much of its capacitance. With a new capacitor, the rail holds steady and the larger module trains and boots normally. The original module had simply been light enough not to expose the weak part.
FAQ
Can a bad SSD cause a boot loop before the logo?
Usually the firmware tolerates a missing or faulty drive and shows an error. Some faulty drives hang the initialisation though, so disconnecting the drive is a quick test.
Why does the laptop loop only when cold or only when warm?
Temperature-dependent loops point at cracked solder joints, a marginal component or a regulator near its limits. Gentle heating or cooling of suspect areas helps localise it.
Does a boot loop damage the laptop?
Not usually by itself, but a loop caused by a short or overheating can make the underlying fault worse. Diagnose rather than letting it loop for hours.
Can removing the coin cell fix a loop?
Sometimes. If the loop comes from corrupted settings, clearing them solves it. If not, nothing is lost by trying.
What if the loop has a fixed number of attempts and then stops?
Many embedded controllers retry the power-on sequence a set number of times before giving up. Count the attempts and look at the step where each one fails.
Bottom line
First decide whether the loop is software or hardware. Then test memory, disconnect peripherals, clear the CMOS, watch the rails and power-good signals during the loop and check the firmware. Most hardware boot loops trace back to memory, firmware or a rail that cannot hold up under load.