5.2 Thermal Resistance Measurements
The following thermal resistance measurements were carried out in an enclosure with controlled air flow (0 m/s) and controlled temperature conditions.
Respecting the JEDEC measurement method, which implies creating a delta power and measuring a delta junction temperature, the ambient temperature variations are considered negligible:
Rth=delta_TJ/delta_power
The two states of measurement are produced under Linux by changing the governor mode from Power Save to Performance. This leads to two power measurements and two junction temperature measurements.
To increase the measurement accuracy and get the correct thermal resistance, it is necessary to repeat the measurement more than 50 times and to average the results. In this case, the average value is 20.5 °C/W, which is close to the simulated target of 21 °C/W on a JEDEC board.
The measurements standard deviation is about 1.5°C/W. It is mainly dependent on the power measurement dispersion on the CPU power line. It is important to strongly average the power measurement when the temperature settles to a stable value.
Compared to the SAMA7G5 Series device, the SAMA7G5 Series SiP device offers improved thermal performance through a lower thermal resistance.