The Evolution Of Circuits For Stretch Sensors
StretchSense is well known for our soft, stretchy capacitive sensor technology, but what many people don’t realize is that sensors are only half of what we develop. We actually develop sensing systems, and we devote as much time and energy to our circuits as we do to our sensors.
Sensors need a circuit to turn the analog signal they produce into useful digital data. It’s this data that enables us to create new and exciting wearable applications. For any questions regarding our sensors or circuits, you can get in touch with us.
We know from experience that developing stretch sensing systems is not a smooth process; the individual and interconnected design of algorithms, hardware selection, firmware implementation, embedded drivers and systems synchronization are all critical to high-quality sensing. In this article, we’d like to offer some examples of the associated challenges from a circuit development perspective.
First, though, we’d like to start by clarifying another common misconception; this one is about soft capacitive sensor technology. Due to the ubiquity of capacitive sensing touch screens, you might assume that it’s easy to make a circuit to run our sensors, but the term capacitive sensing has been misused.
Conventional capacitive sensing is more accurately described as time constant sensing.Capacitive touch screens rely on the change in capacitance that occurs when your finger touches the screen. Crucially, the resistances of the sensing circuits are easily predicted, and knowing this makes it easy to correlate sensor readings with capacitance. Also, the task is simplified significantly because a touch screen is effectively a sealed system with a constant shape.
This eliminates durability challenges from repeated large deformation and prevents environmental factors such as humidity and sweat from having much influence. These two core assumptions — having a predictable resistance value and a highly constrained environment — are the reason capacitive touch screen sensor methods are fundamentally incompatible with our very soft, very elastic sensors.
So why do we care so much about measuring capacitance? Capacitance is governed by the dielectric constant of the stretchy dielectric material, the thickness of this dielectric and the overlapping area of the electrodes. The dielectric constant of our capacitive stretch sensors is very stable and insensitive to changes in temperature and moisture. Capacitance, therefore, is governed by area and thickness.
This is exactly what we want in an accurate and reliable stretch/deformation sensor — an output that directly relates to geometry. If we can measure capacitance accurately and quickly, we can measure deformation accurately and quickly. To learn more about capacitive strain sensing check out our white paper.
Our capacitive stretch sensors are excellent for measuring deformation, but they’re not simple devices:
- They have a highly variable resistive component that is significant, unknown and unpredictable.
- The resistance places a limit on the current that can be driven through the sensor and makes it susceptible to noise if not controlled appropriately (but is great for low power consumption!).
- They do not conform to simple lumped parameter models that form the foundation of standard capacitive sensing methods. This is especially the case when the sensor is in motion.
- Without appropriate system design, they will proximity sense the human body, moisture, fluorescent lights, nearby switched mode power supplies and all manner of unavoidable things to be expected in our daily environment.
- All of the above are amplified when you want to make the sensor thinner, softer or lighter.
And this is just on the sensors’ side. If you consider the electronics and firmware, you get further problems:
- If there is any lack of synchronicity in your data acquisition system, or if it is interrupted by other processes, sensor noise increases significantly.
- You need robust algorithms for calculating capacitance that use all available sensor data to increase accuracy. Many standard algorithms are fragile. If the interrogation waveform is even slightly non-ideal due to poor sensor design, circuit design or simply because the sensor has changed shape during the measurement window, it will generate an error.
- Without smart optimisation, complex calculations are painfully slow and power hungry on wearable-embedded systems with limited resources. This creates substantial latency that reduces the mechanical bandwidth you are able to sense, and punishes battery life.
- To get the very best performance, the analog front end needs to be tuned to the sensor properties, including cables, connectors and anything else that connects the sensor to the electronics.
- The electronics are the bridge between the analog and digital domains. Digital signals wreak havoc on analog signals if this bridge is not tightly controlled.
- Without careful design, effectively connecting to multiple sensors requires switching circuits that will disrupt the system and add significant noise.
This seems like an overwhelming list of things to solve, but over the last ten years we’ve managed to do just that. Our sensors and circuits have co-evolved, and every part of the system — from mechanical to electrical to software — has been optimized to work together. The mechanical sensors and connections are soft, stretchy and comfortable, and the physical signal integrity is assured right from the sensor excitation through to raw signal digitization.
Our algorithms are also fast and robust, giving you the option to maximise the sampling rate for high performance or minimise CPU time for low power. Lastly, we provide precise, repeatable outputs in SI units that directly relate to real physical properties, rather than relative measures with an unknown reference point.
StretchSense has done all of this because a good product is much more than just sensors — it is the integration of many elements into a form and function that makes your customers’ lives better. When you have a new product to design and build, you know that the sooner you put the pieces together and get it into your customer’s hands, the more opportunities you will have to learn and the better the product will be.
Using our sensing circuit as a component on your board — along with the rest of your electronics — means fewer headaches as you try to solve all the challenges listed above. It also increases your speed to market. The product development cycle cannot (nor should it) support the extended development time required to make one subset of components not only work, but work reliably for all conditions — and all for a scale-friendly price.
If you’d like to learn more about how our sensing systems can work your wearable tech or VR project, please get in touch here.