Before you begin
Temperature needs context. A CPU during a demanding render is not comparable to an idle GPU, and a brief peak is different from sustained heat with falling performance. There is no single temperature limit that fits every gaming PC.
- Record the exact component model, sensor name, temperature, workload and monitoring app.
- Compare readings from the same sensor. CPU package, individual cores, GPU core and GPU hotspot are different measurements.
- Look for a change from the PC’s usual behaviour: louder fans, sustained slowdown, a new warning or shutdowns.
Stop testing if there is a burning smell, visible liquid, a thermal warning or repeated shutdowns. Shut the PC down, disconnect power if safe and contact support. Do not run stress tests to reproduce a suspected cooling failure.
Step-by-step checks
-
Confirm what the reading represents
Use the component maker’s monitoring software or a tool recommended by support. Record both the current value and whether it remains high. Compare it with the specification for your exact CPU or GPU, rather than a limit quoted for a different model. Some processors deliberately boost towards their thermal target under heavy load, so a high reading alone is not enough to diagnose a fault.
Model behaviour varies; see AMD’s explanation of Ryzen 7000 thermal targets. Its figures should not be applied to every processor.
-
Check where the PC is positioned
Place the case on a stable, firm surface with its intake and exhaust vents clear. Avoid an enclosed cupboard or a position pressed tightly against a wall. Keep fabric, papers and curtains away from vents. A bottom intake should not sink into thick carpet. Follow the case manufacturer’s clearance guidance and consider whether the room is unusually warm.
-
Check accessible dust filters
Shut down, switch off the PSU if fitted and unplug the mains lead before cleaning. Remove only filters designed for user removal, following the case instructions. Clean them using the method the manufacturer recommends; any washed filter must be completely dry before refitting. Do not spray liquid into the PC or use a household vacuum on internal components. Ask support before removing panels or cleaning inside if you are unsure.
-
Observe the cooling without touching moving parts
From outside the case, check whether the expected case and CPU-cooler fans operate. Do not put fingers or tools near them. A stopped GPU fan at light load can be normal on models with a zero-RPM mode; compare with that model’s guidance. A grinding noise, a fan obstructed by a cable or a CPU fan that unexpectedly stops should be reported.
Example of normal fan-stop behaviour: ASUS graphics-card fan start and stop conditions.
-
Check for an unexpected background workload
Close your games and demanding apps, then look at Task Manager. Updates, rendering, downloads and background scans can keep a PC busy when it appears to be idle. Let legitimate work finish and compare again. If a game runs at a very high uncapped frame rate, an appropriate in-game frame cap can reduce unnecessary load. Note the change so it can be undone.
-
Consider recent changes
Tell support if you recently moved the PC, added a component, installed fan-control software or changed a performance profile. Return only settings you deliberately changed and know how to restore. Do not lower thermal protections, alter voltages, flash BIOS or copy somebody else’s fan curve as a generic fix.
For the relationship between airflow, fans, slowdown and shutdowns, see Intel’s overheating troubleshooting. Internal cooler work should be discussed with ElitePCS first.
-
Compare a normal workload after the simple checks
If there are no warning signs, repeat the same ordinary game or task and record whether temperatures, noise and performance improve. Stop if a warning or shutdown returns. For liquid-cooled systems, report any new unusual noise, visible leak or missing pump reading; a software reading can also be unavailable because of how the system is connected. Do not open a sealed liquid cooler or remove it to investigate.
What the symptoms mean
A short peak when opening a game
A brief boost in activity can cause a temperature spike. Record sustained behaviour and performance as well as the maximum number.
Temperature rises while FPS gradually falls
This can be consistent with thermal limiting. Check ventilation and record the readings for support instead of repeatedly pushing the system harder.
Fans are louder than before
Higher workload, a warmer room, dust or a changed control profile can all matter. Note whether the noise is normal airflow or a new grinding or rattling sound.
The PC switches itself off
Heat is one possible cause, but power and other faults can also cause shutdowns. Stop repeated tests and report the conditions to support.
Common questions
What temperature is too hot?
Use the manufacturer’s specification for the exact component and sensor. Tell support the model, workload and reading if you are unsure. A blanket “safe below this number” rule can be misleading across different CPUs and GPUs.
Is warm air coming out of the case bad?
Warm exhaust can be normal when the cooling system is removing heat. Look at component readings, airflow and stability together rather than judging only by how the exhaust feels.
Should I replace thermal paste?
Not as the first step. Removing a cooler introduces other variables and requires correct refitting. Contact support before removing the cooler, especially if the PC is new or the problem appeared suddenly.
Can I leave the side panel off permanently?
That is not a substitute for diagnosing the cooling. It changes the intended airflow and exposes moving parts and electronics. Keep the system assembled as intended and ask support if a controlled test is needed.
When to contact support
Contact ElitePCS support for repeated thermal warnings, shutdowns, a cooling component that appears faulty or sustained abnormal readings after the external checks. Include:
- Your order number, CPU and GPU models and the monitoring app used.
- The exact sensor labels and readings at light use and during the affected task.
- The game or workload, how long it takes for the issue to appear and any performance change.
- Photos of the PC’s position and vents, and a short recording of any unusual noise.