2 Procedure

Most modern LDOs include a thermal shutdown feature to protect the device from excessive damage at high junction temperatures. For a given level of power dissipation, an LDO with thermal shutdown has a maximum ambient temperature it can operate at before thermal shutdown is triggered and the device shuts off. Equation 2 shows the substitution of the thermal shutdown temperature for the junction temperature and the rearrangement of Equation 1 :

Equation 2.

The Measuring the thermal impedance of LDOs in Situ application report explains how this equation provides a way to determine θ_JA without needing direct access to the junction of the device. First, a small amount of power dissipation is chosen so that the maximum operating ambient temperature is essentially the thermal shutdown temperature of the LDO. A hot oven is used to set the ambient temperature and the LDO is allowed to soak for five minutes. The hot oven must then be turned off to stop any airflow as JEDEC standard models assume no forced convection. An oscilloscope is then used to monitor whether the LDO is shutting off the output. This behavior indicates that thermal shutdown has been triggered. If the LDO does not enter thermal shutdown, the ambient temperature is increased, and the procedure is repeated to determine the maximum operating ambient temperature. This procedure is repeated for increasing levels of power dissipation to provide a sufficient linear regression when calculating θ_JA by using Equation 2.

There are some limitations to the accuracy of this procedure as the ambient temperature and power dissipation are susceptible to change while making the measurement. Turning the hot oven off to ensure natural convection causes the ambient temperature to gradually decrease. Removing any cooling due to the convection created by the hot oven simultaneously increases the junction temperature of the device. Due to the bandgap of the LDO reference drift over temperature, the output voltage decreases, increasing the power dissipated in the pass transistor. The accuracy of the measured θ_JA is also reduced by the measurement accuracy of the hot oven, which is typically ±2°C. To address these limitations, a wide range of power dissipation levels must be chosen to illustrate a wide maximum ambient temperature range. Equation 2 shows that θ_JA is defined as the slope of the trendline between these two variables. As such, verifying that linearity is maintained across a wide range of temperatures and power dissipation increases the credibility of the θ_JA measurement.

The major advantage of this procedure lies in its relative simplicity. It can be used to measure θ_JA on any board since it does not require modifications to the PCB or LDO to measure the specific board or junction temperature, allowing for a more functional PCB layout and testing environment that is more applicable to system designers at the slight expense of accuracy. Because the goal of this application report is to aid designers by investigating general trends between PCB layout and thermal performance across multiple packages, the creation of a functional test setup is prioritized.