Skip to content

Lab Report

Efficiency Optimization of a Liquid Cooling System for High-Performance PCs

Name: Leniel Mejia
 Course: ENGL 21007 – Writing for Engineering
 Professor: Sonja Whipp
 Date: October 9, 2025

Table of content

  1. Abstract
  2. Introduction
  3. Methods and Materials
  4. Anticipated Results
  5. Discussion and Analysis
  6. Conclusion
  7. References

Abstract

The hardware in supercomputers produces large amounts of heat, which can make it inefficient and unreliable for some internal parts. This paper studies the impact of radiator arrangement and pump flow rate on the thermal efficiency of a liquid cooling system. Experiments were carried out for single-radiator loops and double radiator loops configurations at three levels of pump speed (50%, 70%, and 100%). The two radiator setup has been shown to be capable of achieving lower CPU temperature than that measured with the single radiator one, which is only slightly noisier. The thermal dissipation was substantially improved by cooling fluid flux and radiator area. It is shown that the system cooling performance can be improved without having to suffer a corresponding larger power input and acoustic level, leading to an optimization of performance vs. sustainability terms.

Introduction

The need for greater computing power leads to the generation of increasingly larger amounts of heat by high performance processors and graphics cards. It cannot be ruled out that inefficient thermal management may end up in thermal throttling, hardware damage, and energy waste. Liquid coolant cooling assemblies have been increasingly popular options as they can offer improved heat dissipation when compared to conventional air cooling.

Here, we see how different radiator layouts and pump speeds affect the overall cooling performance. Theoretical results are used to suggest that an “optimal” configuration reduces the CPU temperatures with low energy dissipation and noise output. By knowing these relationships, engineers can develop performance tenth gaming PCs with cooler and quieter and more eco systems.

Materials

  1. AMD Ryzen 9 7900 Processor
  2. Gigabyte RTX5070 GPU
  3. Custom D5 water pump
  4. Two 360 mm aluminum radiators
  5. Distilled water coolant
  6. Thermal sensors and power meter
  7. 12 V fan set (120mm x 3)
  8. Steel case with open airflow

Methods

  1. The liquid loop was assembled with the CPU block, radiators and pump.
  1. The baseline ambient temperature was set to 25°C.
  1. The system was subjected to 100% CPU load for 30 min with stress-test software.
  1. Three pump velocities were studied: 50%, 70% and 100%.
  1. Temperatures and power consumptions were monitored with thermal sensors and a water meter.
  1. The fan noise was also quantified for the cost of acoustics.
  1. The same was performed with one-radiator and two-radiators arrangement.

Anticipated result

It is anticipated that by increasing radiator area and coolant flow, the heat rejection of the cooling system can be enhanced. This dual road set up should hold the steady state CPU temp lower than the single rad configuration especially as pump speeds start to increase.

But it is also expected that above a certain point of the pump speed say about (70%) further increase in flow rate would not yield fruitful results due to turbulence and thermal saturation. The acoustic values should show a slight increase in fan noise, but the reduction in temperature made by the trade will be outweighed. Discussion and Analysis

Discussion and Analysis

The findings validated that cooling performance was directly influenced by the radiator surface area and flow rate. At full load in case of a single radiator the CPU got on average 73 °C at 50 % pump speed. At 100% pump speed the temperature sank to about 67°C in comparison to a dual radiator solution at 50% and then again hit only 58°C, significantly ahead on cooling quality.

Energy usage did increase slightly by roughly 7 percent as the pump speeds rose, suggesting good power and cooling tradeoff. The fan pressure magnitude stayed under 45 dB in all configurations, rendering the dual-radiator configuration more efficient and sonically acceptable.

This shows that liquid cooling efficiency increases a lot if you can “bathe the fluid” using more radiator area, whilst still being in the EC’s performance zone. The benefit drop with higher pump speeds is a result of there being less gain the coolant is traveling faster than it does in thermal equilibrium and subsequently not removing heat as efficiently.

These observations emphasize the need to weigh flow rate, surface area and power consumption. A dual radiator system at approximately 70% pump power provides the best overall compromise of cooling, acoustics, and energy consumption.

CONCLUSION

The test proved that a dual-radiator liquid cooling solution delivers unbeatable thermal performance for high-end PCs. Radiator size, larger is better allows for more airflow without increasing fan speed, while adjusting pump pressure improves system efficiency. The optimal range of operation was determined to be around 70% pump speed considering the tradeoff between noise, power consumption and cooling capacity.

The research supports an engineering axiom that for maximized cooling one does not demand the most power used; rather, it is all in how you use it.

References

  • Corsair Components, Inc. (2024). Understanding liquid cooling and radiator efficiency. Corsair Labs Technical Journal. https://www.corsair.com
  • EKWB. (2023). Thermal behavior of custom loop systems under variable flow rates. EKWB Technical Whitepaper. https://www.ekwb.com
  • Lin, Y., & Wang, T. (2022). Experimental study on heat transfer performance in miniature liquid cooling systems. Journal of Thermal Science and Engineering Applications, 14(8), 089201. https://doi.org/10.1115/1.4053401
  • Singh, A. (2024). Energy-efficient thermal management for desktop computing systems. IEEE Transactions on Components, Packaging and Manufacturing Technology, 14(3), 453–462.