For each test, we quickly stepped up the
dissipated power of the device and then sequentially stepped the power down. This is
achieved by sharply changing the supply voltage to increase the quiescent current of
the device, thereby altering the dissipated power. This fluctuation in power results
in a temperature change during the test, allowing us to calculate the system's
thermal parameters through constantly monitored internal temperature and power. In
these experiments, the temperature flow is opposite to the expected norm: the
temperature flow and changes are coming from the object, and the sensor follows the
object. Due to the small sensor thermal mass, there is a possibility to heat the
sensor quickly, creating a situation close to an ideal thermal step function,
allowing the use of Gaussian step equations under specific conditions. The following
steps outline the test setup:
- Thermal Head Setup
- Massive copper thermal
heads were attached under a stable and controlled temperature of +21°C
(close to room temperature) to minimize the influence of convection
airflows. This makes sure that the sensor's self-heating during the
tests does not affect the thermal head temperature.
- To avoid short circuits
of the CB on the bottom side of the PCB, the copper thermal head is
covered by 1-mil Kapton® tape to
electrically isolate the thermal head from the coupon board.
- Thermal Contact
Optimization
- A tiny layer of thermal
grease was applied between the coupon board and the thermal head to
improve thermal contact.
- To stabilize the thermal
contact further and prevent temperature leakage from the device under
test (DUT) to the surrounding air, the coupon board is pressed to the
thermal head by a porous rubber stick with controlled force.
- Test Chamber Conditions
- To avoid the influence of
room air movement, the thermal head with the attached coupon board was
placed into a closed test chamber and kept there for at least 15 minutes
before the test.
- DUT Setup
- Only one DUT is tested at
a time to make sure of accuracy.
- Each coupon board has a
0.1µF surface mount voltage supply ceramic capacitor populated on the
V+pin and referenced to GND.
- The I2C bus runs at
400kHz with a 3V pullup voltage, which does not change during the
test.
- The DUT supply changes
during the test in the following steps: 3V → 5.5V → 3V. This leads to a
device supply power consumption change: 0.4mW → 5.5mW → 0.4mW. The time
on each supply voltage step is 15s.
- The DUT is in continuous
conversion mode with 8 internal averaging and no pause between
conversions. The temperature data are transferred from DUT every
150ms.
Figure 3-2 shows the test
setup, containing the thermal head, an example connected coupon board, and the
rubber stick used in the experiments.