Figures 24 and 25 show the temperature change of the MOSFET in the absence and with the top heatsink. At higher currents, a low-side MOSFET without a heatsink is hotter than a MOSFET with a top-side heatsink. The low-side MOSFET is at 20Using a 60mm heatsink at 0A is about 30°C lower than when not using a heatsink. Similarly, a MOSFET with a 25mm heatsink is 22°C cooler compared to a MOSFET without a heatsink.
Figure 24Low-side MOSFET temperature variation with and without top heatsink.
The high-side MOSFET is at 20Using a 60mm heatsink at 0A is about 33°C lower than when not using a heatsink. Similarly, the MOSFET is 26°C cooler when using a 25mm heatsink compared to a MOSFET without a heatsink.
Figure 25High-side MOSFET temperature variation with and without top heatsink.
Figures 26 and 27 show the temperature change of the MOSFET in the absence and with a backside heatsink. At higher currents, a low-side MOSFET without a heatsink is hotter than a MOSFET with a backside heatsink. The low-side MOSFET is at 20Using a 60mm heatsink at 0A is about 29°C lower than when not using a heatsink. Similarly, a MOSFET with a 25mm heatsink has a 23°C lower temperature compared to a MOSFET without a heatsink.
Figure 26Low-side MOSFET temperature changes with and without underside heatsink.
The high-side MOSFET is at 20Using a 60mm heatsink at 0A is about 31°C lower than using no heatsink. Similarly, a MOSFET with a 25mm heatsink has a 25°C cooler lower than without a heatsink.
Figure 27High-side MOSFET temperature variation with and without bottom heatsink.
For a 60mm heat sink, the following measurements are at 20Recorded at 0A load currents using 3W (m·k) and 6W (m·k) gap pads to understand the effect of gap pad thickness on thermal performance. The two different gap pads are KeraFlol 86 300 Softtherm and 86 600 SoftTherm, as shown in Table 2 at the beginning of this ***.
When the gap pad was changed from 3W (m·k) to 6W (m·k) (thermal resistance was about half of the original) with the top heatsink, a 1-minute decrease in the temperature of the low-side MOSFET was observed6%, the temperature of the high-side MOSFET is reduced by 35% (table 15).
Table 15Gap pad with top heatsink.
With the bottom heatsink used, the temperature of the low-side MOSFET is measured to decrease by about 76%, the temperature of the high-side MOSFET is reduced by about 66% (table 16).
Table 16Gap pad with underside heatsink.
As highlighted, the PCB is optimized for good thermal conductivity and heat dissipation and acts as a very effective heatsink for MOSFETs. This approach is often undesirable in real-world applications where multiple heat sources are present and the heat dissipation capacity of the PCB is limited. The preferred method of heat dissipation is through an ECU enclosure that is thermally connected to the PCB. MOSFETs in a "top cooling" package achieve the lowest thermal resistance between the heat source (MOSFET) and the heatsink (housing), allowing a direct thermal connection between the exposed pad and heatsink of the top MOSFET while minimizing heat inflow into the PCB.
MOSFETs with the same die but in different packages are required to directly compare their thermal performance. All previous measurements used the NVMFS5C460NL, but this MOSFET is not available in the "top cooling" package variant. Therefore, the NVMFS5C450N (SO-8FL Exposed Pad) and NVMJST3D3N04C (Top Thermal Package, Top Exposed Pad) were selected for the following measurements.
The NVMJST3D3N04C is only available as a standard-grade device, while the NVMFS5C460NL is a logic-grade device. In this application, the efficiency of the standard-grade device is expected to be slightly lower than that of the logic-grade device. Still, since the losses are not significant, only the differences in thermal performance, NVMFS5C450N, and NVMJST3D3N04C can be compared.
Table 17Package overview.
The plastic surface area of the top surface of NVMFS5C450N in SO-8FL is 317mm2, with the plastic surface of the bottom surface of NVMJS3D3N04C in LFPAK10 TC (270mm2). Both devices have approximately the same size of exposed pads.
A heat sink with a height of 25mm is used for the following measurements to avoid any limitations of the heat sink and to maximize thermal performance to optimize any differences when heating.
nvmfs5c450n)
Tables 18 and 19 show the temperatures of the high-side and low-side MOSFETs (NVMFS5C450N) with and without heatsinks. The heat sink is mounted on the top surface of the MOSFET (plastic housing).
TABLE 18NVMFS5C450N - No heatsink.
Table 19NVMFS5C450N - 25mm heatsink on the top side.
Figure 28NVMFS5C450N - Low-side MOSFET temperature with and without heatsink.
Figure 29NVMFS5C450N - High-side MOSFET temperature with and without heatsink.
Figures 28 and 29 show the use of a heatsink mounted on the top surface of the MOSFET plastic to improve the heat dissipation of the low-side and high-side MOSFETs.
In 5At 0A load current, the temperature of the low-side MOSFET is about 6°C lower than that without a heat sink, while the temperature of the high-side MOSFET is about 8°C lower. And in 20At 0A load current, the temperature of the low-side MOSFETs is about 40°C lower than that without a heat sink, while the temperature of the high-side MOSFETs is about 37°C lower.
The thermal performance of both measurements is within the expected range, and the heat sink can significantly reduce the MOSFET temperature. In general, the NVMFS5C450N switches at a slower speed due to the higher gate charge and therefore has a higher temperature than previous measurements with the NVMFS5C460NL. Even the on-resistance is slightly lower.
nvmjst3d3n04c)
Tables 20 and 21 show the temperatures of the high-side and low-side MOSFETs (NVMJS3D3N04C) with and without heatsinks. The heatsink is mounted on the top surface of the MOSFET (exposed pad).
TABLE 20NVMJST3D3N04C - No heat sink.
TABLE 21NVMJST3D3N04C - 25mm heatsink on the top surface.
Figure 30NVMJS3D3N04C - Low-side MOSFET temperature with and without heatsink.
Figure 31NVMJS3D3N04C - High-side MOSFET temperature with and without heatsink.
Figures 30 and 31 show the use of a heatsink mounted on an exposed pad on the top side of the MOSFET to improve heat dissipation of the low-side and high-side MOSFETs.
In 5At 0A load current, the temperature of the low-side MOSFET is about 8°C lower than that without a heat sink, while the temperature of the high-side MOSFET is about 10°C lower. And in 20At 0A load current, the temperature of the low-side MOSFETs is about 40°C lower than that without a heat sink, while the temperature of the high-side MOSFETs is about 37°C lower.
In addition, in this measurement, the thermal performance is as expected. Since the NVMJS3D3N04C and NVMFS5C450N use the same die, the losses and heat generation are higher than previously measured with the NVMFS5C460NL, which is due to higher switching losses due to the higher gate charger.
Figure 32 compares the low-side MOSFET temperatures of the bottom (NVMFS5C450N) and top (NVMJS3D3N04C) exposed pads with a heat sink on the top side.
Figure 32NVMFS5C450N and NVMJS3D3N04C - Low-Side MOSFET Temperature with Heatsink
Figure 33 compares the high-side MOSFET temperatures of the bottom (NVMFS5C450N) and top (NVMJS3D3N04C) exposed pads with a heat sink on the top side.
Figure 33nvmfs5c450n and nvmjst3d3n04c
High-side MOSFET temperature (with heat sink).
In general, the thermal performance of this particular PCB and setup is very similar, regardless of whether a MOSFET with an exposed pad on the bottom or top side is used and the heatsink on the top side of the MOSFET package. For low-side MOSFETs, the underside exposed pad package performs slightly better than the top-exposed pad, and vice versa for high-side MOSFETs.
For MOSFETs with exposed pads on the backside, a large amount of heat flows into the PCB, optimized to be an effective heat sink. A heat sink on the top plastic surface of the MOSFET also helps to reduce the MOSFET temperature.
MOSFETs with exposed pads on the top side have relatively poor thermal coupling between the PCB and the plastic surface on the bottom surface. However, the leads soldered to the PCB can also allow heat to flow into the PCB. The exposed pad on the top side of the MOSFET is connected to the heat sink and efficiently dissipates heat.
Both configurations are thermally dissipated through the bottom and top sides of the MOSFET package. For a bare bottom package, the thermal resistance between the MOSFET and the PCB is lower than that between the MOSFET and the heat sink. For top bare encapsulation, and vice versa; The thermal resistance between the MOSFET and the heat sink is low. This results in similar thermal performance being achieved with completely different configurations, and effective heat dissipation can be implemented for both types of packages.
Different measurements and comparisons show the effect of a heat sink connected to a power supply on the MOSFET temperature. Based on the results, the following conclusions can be concluded that are valid for a given setting:
If a MOSFET with an exposed underside pad is used, and a thermally optimized PCB is used for thermal conduction and heat dissipation, the difference between the MOSFET temperature of the heatsink is less than 3°C regardless of whether it is the underside of the mounting PCB or the top surface of the MOSFET.
The MOSFET temperature depends on the heat sink size.
In 20At 0A load current, the MOSFET temperature using a 60mm heatsink is approximately 30°C lower than a setup without any heatsink.
Compared to a setup without any heatsink, the temperature of the MOSFET is reduced by approximately 15 to 20°C when using a 25mm heatsink.
With a 10mm heatsink, the MOSFET is 10°C cooler than a setup without any heatsink.
This temperature change is directly proportional to the thermal resistance of the three heat sinks. It also shows that if a thermally optimized PCB layout is used, the heat sink needs a certain mass and thermal conductivity to significantly reduce the temperature.
25mm and 60mm heatsinks at 20The MOSFET temperature difference of 6°C at 0A load current is lower than initially expected.
In no more than 15At 0A load current, the MOSFET temperature difference between the 25mm and 60mm heatsinks is relatively low, only about 2°C. The load current is higher than 15At 0A, the temperature difference increases by a maximum of about 6°C.
This points to the need for a merit-based heat sink selection to better balance cost and thermal performance improvements.
MOSFETs with an exposed top pad and heatsink achieve similar thermal performance to a MOSFET with an exposed bottom pad mounted on a thermally optimized PCB with a heatsink on top of the package. If you want to minimize the amount of heat flowing into the PCB, MOSFETs with exposed pads on the top side are the right choice because they have minimal thermal resistance to the heatsink mounted on the top side of the package.
All measurements are consistent, repeatable, and in line with general theoretical expectations. This proves that both the electrical and mechanical settings are functioning properly and reliably.
Of course, the test setup is far from real-world applications, such as inside a custom-made aluminum enclosure with a heat sink, where the power supply is part of a complex ECU. However, it explains the influence of different parameters, such as the thermal resistance of the heat sink or the effect of gap pad thickness on the MOSFET temperature. It also makes it clear that similar performance can be achieved by mounting the heatsink on the top side of the heat source (in this case, the MOSFET) or on the other side of the PCB (assuming the PCB layout is thermally optimized, with heat vents on all layers and a larger copper area to allow heat dissipation to flow through the PCB).
If you want to minimize the amount of heat flowing into the PCB, you should use a MOSFET connected to the heatsink with an exposed top pad.
If you need data sheets, sample testing, procurement, BOM matching and other needs, **Customer Service WeChat: 13310830171.