Wonderful – the device is fully developed, the prototype is running and doing what it should. Well, for the sake of completeness, it is tested in a climate chamber for increased ambient temperatures – and promptly fails due to overheating.
Challenges in thermal management
Although most developers are certainly not so naïve in their approach, thermal management for a medical device is always a topic that keeps engineers busy. For most electronics and mechanical engineers, thermodynamics is at best a minor subject and so there are always unpleasant surprises.
With a methodical approach, most thermal problems can be narrowed down in the early project phase and a suitable solution can be developed.
For various reasons, attempts are made to construct a device with a closed housing and without a fan. These include protection against interference, dirt ingress, noise and costs. But how can I estimate beforehand whether it will work without a fan?
Conditions and calculations
The first requirement is to know the ambient conditions during operation. In the temperature range, unfavourable situations such as direct sunlight or cold air flow from an air conditioning system must also be considered. The ambient pressure, depending on the weather and altitude, is also relevant: Air at low pressure and at higher altitudes is less dense and can therefore absorb less heat per unit volume. The future environment also plays a role: where could excess heat be dissipated? Other parameters can play a role depending on the specific case.
Secondly, the conditions in the appliance must be determined. In particular, the temperature range for which the installed or planned components are approved. Furthermore, whether a certain temperature range must be maintained for functional reasons (examples: 40°C due to protein coagulation or 10°C to 50°C due to calibration of a sensor, …). Space conditions in the appliance, the position of possible ventilation openings and the functional arrangement of components play a role in many cases.
Calculation methods and optimisation
The amount of heat to be dissipated can often be based on the electrical energy requirement.
Pel = Q’v
Pel is the electrical power consumed, Q’v the heat flow to be dissipated, both typically in W.
To avoid overdimensioning, a high continuous load Pel is usually selected as the basis for calculation instead of a one-off short-term maximum power.
Now it’s time to do a little maths.In most cases, a few formulae are sufficient:
For the heat transfer through a flat wall – and that is what most enclosures are – the following applies:
Q’ = k*A*Δθ
Where Q’ is the transported heat flow and k is the heat transfer coefficient. It is mainly dominated by air movement.
It ranges from around 5 W/m2K for plastic housings and still air to over 50 W/m2K for metal housings and strongly moving air in and around the appliance. A is the surface area of the appliance that is available for cooling and the temperature difference
Δθ = θimax – θumax
with θimax the maximum permissible or desired temperature inside the appliance and θumax the maximum ambient temperature to be assumed.
If we find
Q’ > Q’v
then we can be happy, because the heat loss is “simply” dissipated via the outer shell without any further measures.
Unfortunately, the world of appliance development is not always so friendly and the power loss is greater than the heat flow emitted through the wall. Now it is time to consider whether and how it can be done without extra ventilation.
It is very effective to reduce the power loss as such (get to the root of the problem!), which is possible in many cases, but often not.
The variables hidden in k (heat conduction and thickness of the wall material, heat transfer coefficients on both sides of the wall) provide powerful tools, as it is directly incorporated into the formula as the factor k. It is primarily determined by the media (in this case air) on both sides of an enclosure wall and their velocity as well as their thickness and the thermal conductivity of the wall material.
Since in most cases the wall thickness of an enclosure is rather low, the air velocity determines how well the heat can be transported. Forced air movement can significantly improve heat transfer. If the air movements inside the appliance and in the surrounding area differ significantly, k must be calculated separately; the above estimates are not sufficient.
Increasing the surface area of the housing (A) is also an option. Only in rare cases is a slight enlargement of the housing sufficient. As this is usually a cost-effective solution, you should still do the maths. How many per cent larger would the surface area have to be to dissipate all the heat? With a cuboid housing, extending each edge by a factor of x leads to an increase in the total surface area by a factor of x2.
Significantly more additional surface area can be achieved with cooling fins. The transport paths for the heat within the fins are considerably longer than the original wall thickness. This is why the heat conduction properties of the material play a major role. Aluminium conducts heat around 200 times better than solid plastics and around 20 times better than stainless steel. Cooling fins can also be attached to the inside of an enclosure. Regardless of whether they are on the inside or outside, it is initially favourable to mount the fins vertically to allow convection flow. Pin heat sinks, which often have even better heat transfer values, can be arranged more freely.
Further or other measures may be possible inside the device, e.g. blowing directly onto the surface in question. In this case, a fan only circulates the air inside the appliance. This still has advantages over openings in the housing wall.
Localised heat sources can be attached directly to an external cooling surface. For a more precise calculation, the heat flow must be divided between the individual surfaces of the enclosure, whereby each partial surface will have individual k, A and possibly also Δθ.
If it is possible to install ventilation (direct air exchange with the environment) due to the boundary conditions, this is usually more effective than optimised heat transport through the wall.
Convection ventilation can be considered for vertically installed appliances. The “chimney effect” is utilised: the heated air column inside the appliance generates a lower gravitational pressure on its underside than the cooler ambient air. If there are openings near the base of the housing and the top of the appliance, cool air will flow in at the bottom and heated air will escape at the top. The low noise level is an advantage. The amount of heat that can be dissipated depends heavily on the design of the appliance. For example, it is favourable if the appliance is as high as possible, has large air inlet and outlet surfaces, the components emitting the most heat are located at the bottom of the appliance and the permissible temperature difference to the surroundings is large. Well dimensioned, heat dissipation can quickly be 10 times better than with a closed housing.
If none of this helps, active ventilation will probably be necessary. At least we now know beforehand and don’t just realise it during the test.
Please note that all details and lists are not intended to be exhaustive, are not guaranteed and are for information purposes only.
