Who, What, Where
Authored by Matt Hankinson
Temposonics, MTS Sensors
Cary, N. C.
Edited by Robert Repas
• A single linear sensor monitors 20 positions.
• Time to read 20 positions matches that of a single position.
• Strokes over 5,000 mm, accuracy of ±20 μm, and resolutions of 1 μm.
MTS Systems Corp.
Foolproofing embedded sensors
Magnetostrictive Sensors For Bottle Filling
How Magnetostrictive Sensors Work
It is often advantageous to have a single sensor monitor more than one feedback parameter. This is now possible in the case of a linear-position sensor that can track as many as 20 moving positions along its axis. The sensor uses magnetostriction as a feedback mechanism.
Sensor electronics send out an interrogation signal. The magnetic field of this signal interacts with position magnets located along the sensor axis to generate sonic waves that travel along a wave guide. The device captures the return signals over a time period corresponding to the entire length of the sensing element.
It turns out this return time corresponds to the worstcase situation for a single magnet sensor. Thus it takes no more time to sense multiple magnets on a magnetostrictive sensor than to sense a single position. And a multiple-position sensor uses the same basic magnetostrictive sensing technology as a single-sensor device, so the two cost about the same. For example, MTS Temposonics sensors can sense up to 20 magnet positions simultaneously within limitations using a single sensor.
One limitation of a multiple-position sensor concerns the spacing between any two magnets. If magnets are close together, their return signals literally overlap and distort from constructive wave interference. Therefore, multiposition magnetostrictive sensor applications need at least a 75-mm space between magnets. This also means any motion that brings a magnet too close to its neighbor causes a loss of signal validity. Fortunately, most applications have enough room to properly space the magnets. Or it might be possible to orient the moving surfaces to keep the 75-mm minimum spacing. Clever design and logical process steps typically mitigate this physical sensing limitation.
It can be challenging to get the individual magnet position data to the instrumentation or controller. It is possible to give each magnet its own separate signal channel. For example, some dual-output analog sensors provide two separate channels corresponding to the two magnet positions. But with more magnets, analog wiring becomes complex and costly.
Industrial fieldbus networks are candidates for transmitting position data on two or more magnet positions. Typically, network-enabled magnetostrictive sensors send position data via the Profibus DP protocol. However, CANbus is another viable network option. Both networks are widely used, robust, and handle a relatively high data rate for fast communication.
The point of a single linear-position transducer with multiple-position feedback is to take the place of multiple sensors. It not only saves money, but also simplifies installation. Each magnet can deliver absolute position information as well as velocity information while in motion.
How Magnetostrictive sensing works
Magnetostrictive position sensors are essentially sonic wave sensing devices. A high-resolution clock measures the time a sonic wave takes to travel the distance between a fixed reference point and a moving magnet. By knowing the speed of the sonice wave, elapsed time is used to calculate the absolute position of the magnet. In addition, the magnet does not touch the waveguide, so there are no parts to wear out.
At their core, magnetostrictive position sensors have four basic components: the position magent, a waveguide, a pickup (also known as sonic-wave converter), and the driver and signal-conditioning electronics.
The conductive "waveguide" wire, made from nickel-based magnetostrictive alloy, carries a short burst of electrical current. The current is called the Interrogation pulse. As it travels along the waveguide, it creates a concentric magnetic field that surrounds the waveguide along it axis.
When the waveguide-magnetic field reacts with the permanent-magnet field from the position magnet, the magnetostrictive effect results in a strain on the waveguide that creates a pressure wave along the guide.
The strain wave travels at the speed of sound in the waveguide, approximately 2,850 m/sec in both directions away from the position magnet. One wave is absorbed at the far end of the waveguide by a damping mechanism. This helps prevent reflections from the end of the waveguide which could cause interference. The other wave travels to the pickup.
The complete waveguide, damping module, and pickup assembly is commonly referred to as the sensing element (SE).
The magnet position is determined by measuring the time it takes the wave to travel along the waveguide. The time is measured from the initiation of the SE interrogation pulse until the return signal is detected by the pickup.
The drive and signal-conditioning electronics generate the precisely timed interrogation signal and convert the internal timing measurement to the desired output. The output can be either an analog voltage, a current value, digital pulses in the form of serial data, or even high speed industrial-network bus communication.
The position (X) is proportional to the time between the two pulses (T) by the speed of sound (S) or X = T x S.
The accuracy of the output is ensured with a precise scaling factor calibration - the speed of sound in the waveguide - that is found using a laser interferometer for each sensor at the final stage of production.
Multiple tool, platen, or cartridge positions coming back from a single sensor lowers per-tool cost of feedback. Tool or cartridge positioning automation significantly reduces changeover time to boost machine productivity.
High-performance control systems need the measurement performance now possible with internal linearization, such as ±20-m accuracy with 1-m resolution. Magnetostriction provides this performance while providing shock and vibration resistance. The superior resolution and accuracy of the magnetostrictive position feedback mean a more precise “cut,” better process quality, and less wasted product.
Magnetostrictive sensors come in stroke lengths over 7,000 mm. There is also a flexible housing option to simplify shipping and installation.
Faster response times with industrial Ethernet protocols, such as EtherCAT, make possible tighter control loops. A typical system can now hit sensor update times of 100 sec for motion-control applications.
One area where multiple-magnet magnetostrictive sensors have taken root is in paper slitters. Slitters have several cutting units positioned using magnetostrictive sensors and closed-loop control systems. Results from the use of multiple-magnet position sensors include more accurate cutting and shorter setup time. It’s often possible to build a smaller and less-expensive machine by controlling multiple knives with a single sensor.
Some slitters measure as much as 10 m across and can have as many as 60 knives along the same axis. Several 20-magnet sensors are used together for such machines.
A position-monitoring system lets operators change the setup of the machine for the next slitting run. Prior to automation, operators manually moved and locked the knife cartridges in place. They positioned them by a tape measure or gauge blocks. Now, the multiple-magnet feedback approach lets setup take place in seconds rather than hours.
Winders present another candidate for multiple-magnet position sensors. Winders slit adhesive tape from a master roll. Like with slitters, operators once positioned the knives and the individual winders by hand. Use of multiple magnets and magnetostrictive technology can let the position of any number of winders be continuously controlled.
A final application example is in plastic injection machines. A plastic injection molding machine needs two sets of high-speed, multiposition sensors. One sensor covers the injector ram and carriage while the second sensor monitors the mold platen and ejector. Speed is critical in this application for high-performance control of position, velocity, and even acceleration.
Fifteen years ago, many injection molding machines used proximity sensors to detect and control the injection, carriage, mold, and ejector during an injection cycle. Operators repositioned the proximity sensors to adjust the cycle. It took a lot of setup time for the adjustments and performance was often marginal. Today, linear-position sensors replace the old “array” of proximity sensors and speed things up considerably. Magnetostrictive sensors are particularly well suited for this application, especially for larger, longer stroke machines. Multiposition magnetostrictive sensors can measure stroke ranges greater than 5,000 mm with resolutions down to 1 m.
All in all, multiposition linear sensors can reduce per-axis-feedback costs while drastically cutting downtime.