i7-10870H CPU thermal throttling?
i7-10870H CPU thermal throttling?
Laptop Model: Acer Predator Helios 300 (315-53)
i7‑10870H
RTX 3060
Recent repairs have led to significant thermal throttling, even when the system is idle. Benchmarks indicate the GPU operates normally, but CPU tests reveal substantial throttling, and Cinebench scores are low, with a multi-score of 400 which is unusually poor for this processor.
In the ThrottleStop Limit window, an “EDP Other” warning persists under the Ring column, accompanied by multiple yellow and red indicators for the core and ring during light to heavy workloads. Under stress, PL2 warnings also begin flashing for the core, GPU, and ring within TS Limits.
HWiNFO logs confirm thermal throttling in both CPU and Ring domains.
Previous actions attempted:
- Verified and reinstalled BIOS to the latest version (original from repair).
- Restarted Windows and Acer PredatorSense settings.
- Limited CPU turbo via Windows power settings (capped at 99 %), with minimal impact.
- Cleared dust from fans and vents; no obstructions found.
- Reapplied thermal paste and repositioned the heatsink correctly.
Possible firmware or EC adjustments that might have been altered during repair—such as lower PL1/PL2 power caps or PROCHOT offsets—could be contributing to premature throttling, even at idle. To investigate further, check the current power limits and offset values, and consider adjusting them if necessary.
For more details, see: https://flic.kr/ps/46DeQ4
Thank you! The service center suggested replacing the motherboard after a power issue. Since it's out of warranty, the repair cost was nearly 90% of the laptop's value. I took it to a chip-level repair shop, which fixed the problem. They mentioned a blown component, probably a MOSFET. I also ran the benchmarks (CPU and GPU) and they looked normal. Here is the image of FIVR's options: https://imgur.com/OymhUtv
When all settings return to their defaults, the situation changes. If temperature readings remain stable despite adjusting those sliders for CPU vcore, cache, iGPU and discrete GPU, it suggests the technician likely connected a wire between point A and point B, which supports my belief that the processor is overheating. I agree with this since I've seen laptops diagnosed by disassembly, finding components hardwired to deliver power. The customer was told to turn it up. Regarding your unvoltage question, an unvolted processor should typically be between -80 and -100, so doubling that would be unusual. Was this the same approach used in your first undervolt?
It might be wise to briefly open the laptop and inspect the heatsink, motherboard, and VRM area to determine if thermal repasting is necessary. There’s no barrier preventing you—just decide if it’s time. Ensure you’re using premium thermal paste and pads. Also, verify that all vents are free from dust and debris. Finally, confirm your undervolts cover not only the CPU core but also the CPU cache. Reducing the voltage by 50mV on the iGPU and exploring mobile RTX3060 undervolting could improve performance.
Considering the laptop’s cooling system is integrated, applying undervolt to both heat dissipation components should yield better outcomes overall.
Good luck!