Accelerometers can be used to measure not only acceleration but also vibration. Historically, accelerometers have seen widespread use in manufacturing production lines and automotive electronics. They are now widely used in portable electronic devices such as cell phones, video game controllers, and cameras. Medical devices increasingly incorporate them to add capabilities to measure motion such as the rate and depth of chest compression.
Choosing the Right Material and Methods
The first and most important way to avoid significant errors when using an accelerometer is to understand the strengths and weaknesses of the two most common types: piezoelectric and MEMS. This will allow you to determine which type is the best suited for your application.
These accelerometers are typically made of a quartz crystal. When accelerated, a mass squeezes the quartz and generates a voltage that is proportional to its acceleration see PiezoAccelTheory.gif.
Piezo Accel Theory source Wikipedia
Piezoelectric accelerometers are accurate: their response is quite linear, usually having a linearity error no larger than 2%. They have a very large dynamic range because their offset, drift, and noise is very small. They can be used over several orders of magnitude. Piezoelectric accelerometers are used in many different industries and applications. These devices measure shock and vibration in labs and on manufacturing production lines. Recently, Voler Systems designed a tester for Phillips Medical to measure the baseline vibration levels at a potential site to help with installation planning and preparation for new MRI facilities.
A piezoelectric accelerometer cannot be used to measure static acceleration, such as the earth’s magnetic field. The lowest frequencies they can measure are usually between 0.1 and 1 Hz.
They can measure frequencies as high as 10 KHz or even 1 MHz for special models.
MEMS (Micro Electro-Mechanical Sensors) usually measure a change in capacitance. A small mass mounted on a tiny lever causes the lever to move when accelerated, changing the capacitance. The mass and lever are built on a tiny Silicon chip. A circuit, usually built on the same chip, turns the capacitance into a voltage.
MEMS accelerometers have the advantage of being able to measure static acceleration, such as the earth’s magnetic field. They cost about 1/10 to 1/100 as much as the piezoelectric types. The MEMS accelerometer voltage output is easy to measure, but they are not as good as piezoelectric types in some applications. MEMS accelerometers typically do not respond to frequencies as high as piezoelectric types. They are usually suitable for frequencies up to 100 Hz or 1 KHz. Because they have large offset and drift, they are suitable for use over one or two orders of magnitude of vibration or acceleration.
Voler designed a MEMS accelerometer into a disposable medical device used inside the body, where the cost of a piezoelectric accelerometer would have made the device too expensive.
Aliasing, The Most Commonly Misunderstood Aspect of Accelerometers
The second biggest source of error is often aliasing. Aliasing is a type of sampling error in which frequencies that do not actually exist appear in the data. Since nearly all data acquisition devices are digital, they have a sample rate: the rate at which samples are collected and stored. If you use the wrong sampling rate the sampled waveform will differ significantly from original.
For example, if the sample rate is 1,000 Hz and the data is a 950 Hz sine wave, the difference is 50 HZ. A 50 Hz signal would actually appear in the data, even though it did not come out of the accelerometer. Any signal that is higher than half the sample rate will be aliased. Aliasing produces a new frequency that is the difference between the sample rate and the measured frequency. When the difference in the frequencies is small the resulting frequency is low and very noticeable.
There are two ways to avoid aliasing: increasing the sample rate and anti-aliasing filtering. Increasing the sample rate takes faster and more expensive data acquisition equipment, and it creates more data, which takes more storage space and more time to analyze. In most cases using an anti-aliasing filter is the most practical way to remove the unwanted noise from the signal. The selection of filters is covered in the whitepaper with a link at the end of this article.
Impact of Time Delays
The third biggest source of error is not accounting for the impact of the time delay between sampling the data and the application software processing that information. Each application will have unique requirements but this factor directly impacts cost and latency.
Some applications require small latency for example a gaming controller must have a quick response time from a changing input to processing in the application, or it will confuse the operator. Control applications require very small latency depending on the response time of the control. Other applications like a pedometer tolerate longer latency, since the data is only being recorded.
These requirements affect the buffer and pipeline design. Interlacing multiple channels can also increase the complexity of the design. Many systems use a multiplexer to sample one input channel after the other in sequence. As a result the channels are not read simultaneously.
Incorporating accelerometers into your design can provide valuable capabilities. But to deploy them effectively you need start out by selecting the right type of accelerometer for your application and follow careful sensor design practices so as not to induce errors or latency issues. This will allow you to meet the application requirements for your design. For more detailed background see our Data Acquisition Basics Whitepaper at https://www.volersystems.com/blog/design-tips/122-data-acquisition-basics.
About Voler Systems
Voler Systems is a full-service consulting firm that helps emerging and established firms commercialize novel and innovative devices. Our MEM accelerometer and instrumentation consultants design products for various consumer and industrial applications.
Software: Testing Motion Detection for Wearable Devices
Active Mind Technology provides golfers with stats and trending information about their golf game. Their product, Game Golf, uses GPS to track location, accelerometers to track swing and related information, and Near Field Communication (NFC) tags on each golf club to identify the club being used. They approached Voler Systems after early prototypes exhibited erratic behavior. Because of their commitment to quality they wanted every measurement to be accurate. We helped improve some of the software, but the most important thing we did was to provide a way to test it after every new software release.