Tool/software:
Hi,
I would like to use the TMP117 temperature sensor for my high accuracy temperature sensing application. I would like to know the thermal response time to 63% of final value in still air for the TMP117 sensor through its top case. I am more interested to know the response time through the top case of the sensor. In my application I cannot use the bottom dia-attached pad for my measurement.
Thanks
Hello Amal Sivaji,
For thermal response there are two things we need to consider to provide the most relevant data (mentioned below).
We might not be able to exactly match your PCB specs, but we could see whether we have data that we can share, or potentially do some new testing.
1. Thermal Mass
Specifically, the effective thermal mass, including both the sensor and the surrounding PCB. To minimize the thermal mass (of the sensor), you can choose a smaller sensor; to minimize the thermal mass of the PCB, you could use a small PCB, or a flex PCB.
For example, in another a newer and smaller device we recently released, TMP118, we show a much better thermal response in the flex PCB versus a typical rigid PCB (0.11 seconds vs 1.40 seconds), since the thermal mass of the device, and the medium (stirred liquid) isn't changing, the only variable is the thermal mass of the PCB, this shows the PCB dominates the difference in thermal response.
2. Thermal Resistance
Specifically, the thermal resistance between the sensor and the 'temperature target' - in your case, is the Thermal Junction-to-Ambient Resistance.
Amal Sivaji,
I forgot to mention, we have an application note with some comparisons on thermal response time which is actually in still air, using both rigid and flex PCB (snapshot of the results below). Here is the application note:
Let us know if this answers your question, otherwise, it would be useful to have some more details regarding your PCB.
Note: this application note uses 3 tau (95%), versus the 1 tau (63.2%) specification you might see in some of of our datasheets.
Amal,
Thermal response time is a function of mass that has little to do with the sensor IC. It will be slower on a thicker PCB, it will be slower when there is copper fill and it will be slower for larger IC packages. Rest assured all of our sensor ICs are fast enough to capture real temperature change, and are instead limited by the physics of the mass around them. Similarly, the temperature stability (and typically accuracy) will be better when the sensor is surrounded by a significantly larger mass. Mass is your thermal capacitor.
We have previously recommended that customers take advantage of the thermal pad on the bottom of the TMP116 and TMP117 in the DRV package. This thermal pad, which is normally intended to dissipate heat away from a high-power IC, provides a low thermal resistance (theta j) path to the IC sensor itself. This thermal pad doesn't eliminate the issue of mass, and it doesn't mean the sensors in the DRV package are not capable of measuring temperature at the top of the package.
Since the thermal response is entirely an issue of mass, the YBG package is a superior choice. The YBG package does not have plastic encapsulation around the IC, so it has a minimum amount of mass. The IC sits directly on the BGA solder balls and there is no thermal pad.
In summary, heat transfer into the sensor is more constrained by the materials around the sensor that would draw that heat away. The sensor itself will absorb heat from all directions and digitize that temperature change fast enough to see it occur. Even when submerged quickly into a large mass of fluid, the sensor is able to report several samples as the temperature settles.
thanks,
ren
