Sensors: Know Thyself
August 4, 2003
Off-the-shelf sensors come with a lot of information that helps users make accurate measurements. That information, usually found in a data sheet, may include linearity, temperature drift, output ranges, calibration, and other characteristics. But paper data sheets get filed and forgotten, so when it comes time to use a sensor, engineers often scramble to find needed specifications.
To combat such problems, manufacturers of sensors that produce analog outputs have led a movement to embed information within sensors. This embedded information, stored in a small memory chip, can tell a data-acquisition system what sensors are attached to it and how they work. To make it easy for electronic equipment to extract information from a sensor, manufacturers agreed to a standard data format. That format, as embodied in IEEE Standard 1451.2, defines a transducer electronic data sheet, or TEDS. So far, companies involved in supplying sensors for vibration and acoustic testing have led the way in adding TEDS data to their products.
Class Act: Class-1 smart sensors use existing connections to communicate TEDS data to an instrument or card. These smart sensors work with older systems. Class-2 smart sensors provide separate communication lines for the TEDS data, which lets manufacturers retrofit TEDS memory chips in older sensor types. |
The outputs of some accelerometers, for example, vary slightly from one device to the next due to the nature of the sensing elements. Although accelerometers of one type might generate an average output of 100 mV/G, individual sensors might produce slightly different outputs for the same acceleration. So instead of specifying a nominal output of, say, 100 mV/G on a datasheet, the sensor manufacturer calibrates individual sensors and places information about each sensor's characteristics in its memory chip. This sort of sensors-specific information helps engineers rapidly set up a system to make accurate measurements.
Typical TEDS data can include: Basic Information-manufacturer's ID, model number, serial number, etc.; Standard and Extended Information-sensitivity, response time, measurement range, etc.; and User information.
Of course some of the information will vary from one type of sensor to another. A strain gauge and a pH sensor may have little in common beyond TEDS fields for manufacturer ID, model number, and serial number. But the P1451.2 standard lets the TEDS structure accommodate data for a wide spectrum of sensor types.
A complementary standard, IEEE P1451.4, provides for a mixed-mode (analog and digital) interface between sensors and data-acquisition systems. This interface specifies the connection between a data-acquisition system and the TEDS data stored in a sensor's memory chip. The standard also specifies how manufacturers can build TEDS-enabled sensors as either Class-1 or Class-2 devices.
A Class-1 sensor use the same two- or three-wire connection engineers would find on older sensors that lack built-in memory. Devices in this class share the analog-output line with TEDS data transmissions. Under control of a data-acquisition card that complies with the IEEE 1451.4 standard, a sensor will transmit its TEDS data to the card, which passes the data to instrument-control software. The presence of the memory chip does not affect normal sensor operation. Thus the Class-1 devices provide for legacy operations. Users can employ Class-1 sensors in older data-acquisition systems that cannot read TEDS data. Likewise, TEDS-enabled data-acquisition cards will work with sensors that lack a memory chip. In this case, software will not find TEDS data in a sensor and may ask the user to enter that information from a paper datasheet. In short, the use of Class-1 "smart" sensors neither forces a replacement of nor disrupts an existing data-acquisition system.
On the other hand, Class-2 device relies on separate connections for the sensor output and communications with a sensor's memory chip. This separate connection lets sensor manufacturers provide Class-2 capabilities for any type of sensors, including load cells, humidity sensors, and strain gauges. And the Class-2 standard makes it easy to retrofit older sensors with TEDS information. Manufacturers simply add the required memory chip and two dedicated digital-communication lines. Both Class-1 and Class-2 devices use the 1-Wire communication protocol supported by many standard integrated circuits. (See "Overview of 1-Wire Technology and Its Use" at www.maxim-ic.com.)
To take advantage of the TEDS data embedded in a sensor, engineers will need newer data-acquisition cards, modules, or instruments that can access the information in Class-1 or Class-2 sensors. And those types of products are now available from several manufacturers. But few engineers will tear out and replace older systems just to take advantage of smart sensors. The use of smart sensors can pay off, however, if you routinely reconfigure test systems and often change sensors, sensor types, or the mix of sensors used in a test system. This situation occurs in multichannel structural testers that use hundreds of sensors. The extra cost of having to repeat a lengthy test due to improperly connected sensors or due to manual input of incorrect sensor data may justify the cost of replacing older gear with newer "smart" data-acquisition products. Also, any engineer specifying a new system that will use many analog sensors should evaluate smart sensors and compatible data-acquisition products.
When engineers can't justify the cost to upgrade to smart sensors and cards, they still can take advantage of TEDS data, even for older sensors. Manufacturers have started to offer virtual TEDS data users can download. These online resources let engineers access information about a specific sensor or sensor family in a TEDS format. Of course, the software must "understand" the TEDS format and know how to use the TEDS data.
The overarching IEEE 1451 standard includes other subsections that specify wireless links and multi-drop buses that can connect sensors and data-acquisition systems. (See http://ieee1451.nist.gov.)
As always, commercial developments frequently bypass standardization efforts. Instead of embracing a new standard, sensor manufacturers and instrumentation users may opt for a simple, existing standard: Ethernet. Earlier this year, Lantronics of Irvine, CA (www.lantronix.com), introduced a Web server, called XPort, which fits into the housing of a standard RJ-45 Ethernet connector. The small module includes an embedded controller, operating system, and electronics that make it easy to integrate into equipment.
Although the XPort may seem like overkill for inexpensive sensors, it's conceivable that Ethernet connections and Web servers will become standard in high-end sensors. It's a lot easier to connect several Ethernet cables to a computer than to connect hundreds of individual sensor wires, whether or not they comply with the IEEE P1451.4 standard. Still, sensors such as RTDs and thermocouples will remain firmly in the analog camp because it's just too expensive to add digital capabilities to such simple devices.
About the Author
You May Also Like