Home
Thermal overload is still one of the most common failure reasons for modern equipment and systems. Due to the increasing performance density of electronic components, the equipment becomes more efficient and at the same time smaller. Therefore the problem of heat dissipation becomes an important part of the development as a whole. The thermal simulation of heat development and the heat flow by way of a simulation programme has now become essential.
Software for thermal analysis
Nowadays we have programmes that enable thermal analysis of a complete system. Schroff work with the simulation software Flotherm during the development of their cases, subracks and cabinets. As special service, Schroff offer the use of their climate control laboratory, including heat simulation and the know-how of their climate control experts, to their customers for development of own equipment.
Das The principle of the simulation software is simple. The user graphically configures a geometric model of his equipment from various components. These data can also be taken from a CAD system. If the programme is used purely for the development of an enclosure, the internal assembly consists of a normal test assembly with those standardised components (boards, power supplies, etc.) which are most likely to represent the future application. If the simulation is for a customer, of course, the components stipulated by the customer would be “built in”. Following that, he will determine the physical parameters such as material data, the performance losses of the installed components, the performance and characteristics of the fans, the environmental parameters and the like. Now the software calculates the linked parameters from convection, controlled airflow, radiation and heat conduction on a 3D grid. This is done with similar equations as for the weather forecast, but through the installed mechanisms it is much simpler. The software calculates the entire airflow in the enclosure (air speeds and pressures) as well as the temperature distribution of the air and the solid items in the system.
The result is visible
As a result the calculated airflow field is analysed and graphically illustrated. Now the developer can undertake alterations with for instance other fan performances, another configuration of the fans or the air inlets, the assembly of air baffle, other materials and material thicknesses or with another configuration of the installed components. Sometimes very simple changes, such as the increase of the fan speed, do not always lead to a solution of the cooling problem. It can even happen, with certain cinfigurations, that the increase of the fan performance compounds the problem. In most cases a number of variants are calculated before a satisfactory enclosure or enclosure construction has been found. The construction of the optimum version is then realised and the real heat distrubtion is tested in the laboratory. Through the simulation and subsequent measurements the critical points in the system and the position of the temperature sensors are established. On the other hand the test results are used for later simulations, with the aid of suitable correction factors, to continually improve the balance between theory and practice.
Additionally to the application at system level the software programme can be used for detailed analysis on a smaller scale, for instance for components. Often calculations of this kind are carried out for critical components with high performance density and serve to evaluate the performance of cooling fins positioned directly on the components and to improve their design.
Example: Temperature distribution in a customised enclosure
An example illustrates the advantages and effective use of simulation software. A customised enclosure (IP 54) with assembled subrack and various plug-in units plus a power supply is tested. An ambient temperature of 55 ºC is stipulated. The performance losses of the assembled plug-in units lie around 5 W, 3 W, 4 W and 9 W, the power supply generates a performance loss of 17 W. The power supply is positioned directly on the side panel, on which also a cooling unit is fixed. The subrack is connected directly with the rear enclosure wall, onto which, too, a cooling unit is fixed. The first simulation of the temperature distribution (Picture 1) showed that with pure convection alone air circulation could not be achieved, which would be necessary to transport the heat from the components to the cooling outside panels. Therefore two fans were installed inside the enclosure and their performance loss taken into account with the rerun simulation. The fans were positioned behind each other in line with the boards for the first test. Through this an even temperature distribution was guaranteed (Picture 2). The achieved internal temperatures were still around 5K higher than the maximum temperature aimed for. In the next test other fans with a lower maximum pressure were used and they were located in an offset manner. The power supply was still being cooled by the outside fan. To guarantee that the air stream of the enclosure would not pass the power supply, an air baffle was fixed between the power supply and the subrack. The result was better now, but still not sufficient. In the last test yet other fans with higher air volume and maximum pressure were applied and positioned in an offset manner below the subrack, although this entailed a higher performance requirement. Furthermore the enclosure dimensions and the positon of the subrack in the enclosure were changed slightly. With these measures an optimum temperature distribution could be achieved inside the enclosure (Picture 3).
Software for thermal analysis
Nowadays we have programmes that enable thermal analysis of a complete system. Schroff work with the simulation software Flotherm during the development of their cases, subracks and cabinets. As special service, Schroff offer the use of their climate control laboratory, including heat simulation and the know-how of their climate control experts, to their customers for development of own equipment.
Das The principle of the simulation software is simple. The user graphically configures a geometric model of his equipment from various components. These data can also be taken from a CAD system. If the programme is used purely for the development of an enclosure, the internal assembly consists of a normal test assembly with those standardised components (boards, power supplies, etc.) which are most likely to represent the future application. If the simulation is for a customer, of course, the components stipulated by the customer would be “built in”. Following that, he will determine the physical parameters such as material data, the performance losses of the installed components, the performance and characteristics of the fans, the environmental parameters and the like. Now the software calculates the linked parameters from convection, controlled airflow, radiation and heat conduction on a 3D grid. This is done with similar equations as for the weather forecast, but through the installed mechanisms it is much simpler. The software calculates the entire airflow in the enclosure (air speeds and pressures) as well as the temperature distribution of the air and the solid items in the system.
The result is visible
As a result the calculated airflow field is analysed and graphically illustrated. Now the developer can undertake alterations with for instance other fan performances, another configuration of the fans or the air inlets, the assembly of air baffle, other materials and material thicknesses or with another configuration of the installed components. Sometimes very simple changes, such as the increase of the fan speed, do not always lead to a solution of the cooling problem. It can even happen, with certain cinfigurations, that the increase of the fan performance compounds the problem. In most cases a number of variants are calculated before a satisfactory enclosure or enclosure construction has been found. The construction of the optimum version is then realised and the real heat distrubtion is tested in the laboratory. Through the simulation and subsequent measurements the critical points in the system and the position of the temperature sensors are established. On the other hand the test results are used for later simulations, with the aid of suitable correction factors, to continually improve the balance between theory and practice.
Additionally to the application at system level the software programme can be used for detailed analysis on a smaller scale, for instance for components. Often calculations of this kind are carried out for critical components with high performance density and serve to evaluate the performance of cooling fins positioned directly on the components and to improve their design.
Example: Temperature distribution in a customised enclosure
An example illustrates the advantages and effective use of simulation software. A customised enclosure (IP 54) with assembled subrack and various plug-in units plus a power supply is tested. An ambient temperature of 55 ºC is stipulated. The performance losses of the assembled plug-in units lie around 5 W, 3 W, 4 W and 9 W, the power supply generates a performance loss of 17 W. The power supply is positioned directly on the side panel, on which also a cooling unit is fixed. The subrack is connected directly with the rear enclosure wall, onto which, too, a cooling unit is fixed. The first simulation of the temperature distribution (Picture 1) showed that with pure convection alone air circulation could not be achieved, which would be necessary to transport the heat from the components to the cooling outside panels. Therefore two fans were installed inside the enclosure and their performance loss taken into account with the rerun simulation. The fans were positioned behind each other in line with the boards for the first test. Through this an even temperature distribution was guaranteed (Picture 2). The achieved internal temperatures were still around 5K higher than the maximum temperature aimed for. In the next test other fans with a lower maximum pressure were used and they were located in an offset manner. The power supply was still being cooled by the outside fan. To guarantee that the air stream of the enclosure would not pass the power supply, an air baffle was fixed between the power supply and the subrack. The result was better now, but still not sufficient. In the last test yet other fans with higher air volume and maximum pressure were applied and positioned in an offset manner below the subrack, although this entailed a higher performance requirement. Furthermore the enclosure dimensions and the positon of the subrack in the enclosure were changed slightly. With these measures an optimum temperature distribution could be achieved inside the enclosure (Picture 3).
 |  |
Picture 1. Temperature distribution with normal convection (X cut - horizontally/Z cut - vertically through the enclosure)
 |  |
Picture 2. Temperature distribution of two fans arranged behind each other (X cut - horizontally/Z cut - vertically through the enclosure)
 |  |
Picture 3. Temperature distrubiton after optimisation (X cut - horizontal/Z cut - vertically through the enclosure)
The correlation between the calculation and the practice is surprisingly good. Therefore it is all the more important for professional developments of electronic equipment or systems, that a climate control specialists advises the developers and, if required, carries out the calculations. Moreover through the use of simulation tools the cooling concept can be optimised at an early stage, before time consuming construction and production of samples. Development times and costs are reduced considerably.
