Addressing the Challenges of Traditional Thermometers

date

Feb 2, 2025

slug

addressing-the-challenges-of-traditional-thermometers

status

Published

Addressing the Challenges of Traditional Thermometers - Measuring Temperature Over a Distance Using Ultrasound

Thermometers often present a few challenges. For example, with traditional thermometers, we must move to the location where we want to measure the temperature. Additionally, environmental factors such as sunlight can affect the accuracy of readings.

Our goal is to solve these issues with a device that measures the flight time of sound through air, or other gases and calculates the temperature based on that and other measured properties of the medium.

The First Step: Building the Device

The initial phase of the project involved creating a device capable of producing and receiving ultrasonic sound. Our primary objective was to measure ultrasonic sound with high precision over short distances to validate the concept.

There are off-the-shelf modules available for distance measurement applications. These modules consist of two ultrasonic transducers placed next to each other, facing the same direction, and soldered onto a common board. However, since we needed to measure over a greater distance, separating the two units posed hardware challenges, although it would optimize precision.

Designing the Units

Initially, we considered ordering pairs of transmitters and receivers. We knew that we would need to design two circuits for the units:

For the Transmitter (T unit): We needed to design a circuit that could switch the polarity. A simple H-bridge is perfect for this task.

For the Receiver (R unit): The voltage output from the receiver was too low to be interpreted by the processing unit, so we had to amplify it.

Both units are connected to a single system — an Arduino board running straightforward software. This software sends a 40kHz square wave signal to the transmitter unit, which acts as a filter by only operating at that frequency, ensuring that the transmitted and received sound signals are pure sinusoids. By measuring the time it takes for the sound to reach the receiver over a fixed distance (currently 10 cm), we can calculate the speed of sound, and from there, approximate the temperature.

Challenges and Inaccuracies

Unfortunately, the measurements were so inaccurate that we had to break them into blocks of thousands and calculate the average.

At first glance, the device seems to react inconsistently to different heat sources. However, this variation is due to the fact that sound velocity is influenced not only by temperature but also by factors like the composition of the medium. For example, when we placed a lighter under the measurement section, the temperature seemed to drop. Lighters contain butane, where the velocity of sound is about 60% of that in air. On the other hand, using a hairdryer increased the temperature, as expected.

Software

Here is the core part of the software:

We ran into some issues trying to get the PWM (Pulse Width Modulation) timer to work. PWM is a dedicated peripheral built into most microcontrollers that allows us to offload the task of sequential pulse generation. This would free up the main processing unit to focus on other critical tasks, such as data analysis. Ideally, the sum of the two delays should amount to 25 microseconds in the absence of other system delays. However, due to inherent timing discrepancies in the system, we arrived at the current values (9 and 8 microseconds), which produce a square wave signal that approximates 40 kHz.

In my experience, implementing PWM is much easier on low-level microcontrollers, which are typically programmed in Assembly. (Once the project reaches its optimization phase, we plan to use such devices. For now, using an Arduino speeds up testing.)

Here is an embarrassingly ugly function to calculate the temperature from the time difference:

The numbers are derived from the following formula:

where T is temperature, measured in degrees Celsius, and the time unit is in microseconds.

A Quick Workaround

We originally planned to use a lighter for demonstration during the upcoming open day. However, since the change in medium composition caused a drop in the temperature graph, and given our time constraints, I decided to reflect the function over the time axis and shift it along the temperature axis. I know it's a terrible solution, but we found it hilarious.

First Build, Optimize Later

After reaching the first milestone of this project, the result still demonstrates that the idea we envisioned is workable — and that’s what matters for now.

There are several improvements needed for the circuit:

Impedance matching

Recalculating component parameters for more efficient power consumption and optimal component selection

Eliminating electromagnetic interference

Reducing audible, acoustic noise

Reducing the size of the device

Reducing the size again

These are part of the optimization phase of the project that I mentioned earlier. We’ve put this stage aside for now and opted to use off-the-shelf components, so we can shift focus to a more exciting area: The 2D Heat Map.