Thermosyphon cooling technology solves the GPU server heat problem

      With the development of deep learning, simulation, BIM design and AEC applications in all walks of life, with the support of AI technology and virtual GPU technology, powerful GPU computing power analysis is needed. Both GPU servers and GPU workstations tend to be miniaturized, modular and highly integrated. The heat flux density often reaches 7-10 times that of traditional air-cooling GPU server equipment. 

     Due to the centralized module installation scheme, there are a large number of NVIDIA GPU graphics cards with large heat generation, so the heat dissipation problem is very important. In the past, the commonly used thermal design has been unable to meet the use requirements of the new system. The traditional liquid cooling GPU server or liquid-cooled GPU server is inseparable from the blessing of the fan. The thermosyphon cooling technology is gradually widely used in server heat dissipation.

       At present, the thermosyphon cooling technology in the market mainly uses the column or plate radiator as the body, penetrates the heat medium pipe at the bottom of the radiator, injects the cooling medium into the shell, and establishes a vacuum environment. This is a normal temperature gravity heat pipe. 

    The working process is as follows: at the bottom of the radiator, the heating system heats the working medium in the shell through the heat medium pipe. Within the working temperature range, the working medium boils, the steam rises to the upper part of the radiator for condensation and heat release, the condensate flows back to the heating section along the inner wall of the radiator and is heated and evaporated again. The heat is transferred from the heat source to the heat sink through the continuous circulating phase change of the working medium to achieve heating Purpose of heating.

       From the original aluminum extrusion heat sink to the newly air cooling heatsink, it is still a good choice to use moer fins for better cooling performance. You may think that since some small fins are so easy to use, is it better to use more and larger fins? However, the farther the fin is from the heat source, the lower the fin temperature, whick means limited cooling effects. When the temperature drops to the temperature of the surrounding air, no matter how long the fins are made, the heat transfer will not continue to increase.

      Unlike the heat pipe, the thermosyphon heat dissipation uses the pipe core to bring the liquid back to the evaporation end, but only uses gravity and some ingenious designs to form a cycle, which uses the liquid evaporation process as a water pump. This is not a new technology and is common in industrial applications with high heat release.

  Generally speaking, the refrigerant inside the GPU will boil, flow upward to the condensing end, change back into liquid and return to the evaporating end. Theoretically, there are two advantages:

1. Avoid heat pipe drying up and can be used for overclocking and ultra-high performance chips.

2. Because there is no need for water pump, the reliability is better than the traditional integrated liquid cooling.

       The most important point of thermosyphon cooling now is that its thickness will be reduced from the traditional 103 mm to only 30 mm (less than one third). It is relatively small in shape and will not damage the performance. In order to facilitate processing, most manufacturers use aluminum materials at present. Copper is also used, and the temperature may be further reduced by 5-10 degrees. It is only for GPU servers with high heating capacity ,with the technology developed, more and more thermosyphon thermal  solution will be used in other applications in the future.