Magnetometers

One of the earliest sensors that humans made to aid us as we navigated the world, was the compass. Long before the accelerometer, the gieger counter, or the frog leg oscilloscope, nearly one thousand years ago inventors in China noticed that a lodestone (a natural magnet) would align itself in a consistent direction if floating on a still pool. From there the compass was born.
 
The magnetometer is the descendant of those early compasses and is now a sensor that is hugely important to the modern world and used across a wide array of technologies and devices.
 
Now, often in a Transducer Thursday article, I would break down the core effect that underpins a sensor and then go through the conceptual steps to turn that effect into a useful sensor. However, there are near countless ways to build a magnetometer, from a magnetised mass on a spring (see accelerometers for more details on how this kind of force sensing-based approach would work) all the way through to sensors that utilise quantum effects. So instead of giving you brief rundowns on how all these sensors work, we’re going to take the opportunity to talk about another development in sensor technology, micro-electromechanical sensors, or MEMS.
 
MEMs technology has enabled so many advances all from developments in miniaturisation. Accelerometers, gyroscopes, and pressure transducers have all been shrunk down to smaller than a grain of rice allowing them to be used on printed circuit boards alongside micro-processors. Typically these sensors have an A to D (analogue to digital converter) incorporated with them which allows them to communicate directly in digital buses with other micro-electronics.
 
But how do we actually miniaturise a sensor such that it can be mounted on a circuit board? Let’s dive into two example technologies.
 
MEMS tech was and is so groundbreaking because it has miniaturised mechanical systems within the sensor which means that we can utilise moving parts and strain and deflection effects. However, we don’t always utilise moving parts.

The limits of miniaturisation mean that some parts are quite limited, like bearings and complete rotating parts which are seldom if ever possible. But it gives us a lot of options for moving pieces and allows any components that have no moving parts so long as the structures can be created on micro-metre to nano-metre scales. Because of this MEMS tech quite often uses capacitance changes in response to the small movements of minuscule parts, this is fundamentally the main way that MEMS accelerometers work, for example.
 
We have a number of effects we can take advantage of for magnetometers. One we can use is the Lorentz force. This principle states that a force is generated by a current-carrying conductor in a magnetic field. Useful right? All we have to do is measure the force on the wire and we have a sensor. However, we have an issue, other effects can cause deflection in the wire. Gravity and other accelerations can also cause deflections, we can work with these limitations, but the Lorentz force provides us an opportunity to filter these effects out. If we switch the current on and off the Lorentz force will disappear and re-appear, whereas the deflection due to gravity or acceleration will remain in place allowing us to isolate the force due to magnetism. This also gives us a way to filter out deflection due to heating, as deflection and strain due to heat effects will persist and the Lorentz force will not. We can miniaturise multiple structures to take advantage of these effects.
 
However, moving structures have disadvantages, they have response times due to the time it takes for the structure to deflect, this means that whilst we can filter out constant gravity and accelerations as we can allow for settling time, rapid changes are more challenging to filter. Moving structures also have a frequency response which means that there will be some exciting frequencies that provoke extreme non-linear responses. The challenge and ingenuity of sensor design, of this and similar types, is tailoring the structure such that the settling time and the frequency response are desirable. See our further reading section for examples of Lorentz force magnetometer designs.
 
What if we had a design of sensor that has no moving parts? These are still easy to miniaturise and will avoid some of the problems we just described for moving structures. One example is the tunnel magnetoresistance sensor or TMR magnetometer. TMRs take advantage of quantum tunneling which allows electrons to ‘tunnel’ through very short distances of an insulating material. In a TMR two conductive magnetic plates are held apart by a very thin film of an insulator, a matter of nano-metres thick.

Due to complicated quantum effects which my poor engineer brain struggles to understand, electrons are increasingly likely to successfully tunnel from one magnet to another the more closely aligned their magnetic fields are. If electrons find it very hard to tunnel, there is no overall change in charge carriers to carry current and so the resistance of the structure is very high. If the magnetic fields are aligned however, electrons start to tunnel in one direction more than the other, this causes a drop in resistance of the material as a current can more easily flow. If one of the ferromagnetic plates is a permanent magnet, and the other is allowed to respond to external magnetic fields then the resistance will vary depending on the response of the second plate. Based on the varying resistance of the combined material a judgement can be made about the magnetic field strength. TMRs have several excellent qualities, their resistance change is extremely strong, such that it is easily detectable, and they have very limited response to other influences, such as acceleration or heat. They also eliminate moving parts that would produce mechanical wear.

These advantages mean that they are used extensively in magnetic encoders, their fast and strong response being extremely desirable in this circumstance. One particular use case is in mechanical hard drives where they are used to read the magnetic ones and zeros stored on the disk but with a non-linear correction, they can be used to sense field strength rather than just 1’s and 0’s.
 
MEMS magnetometers of various kinds are used in our phones, many phones have a rudimentary compass which can be computed from the readouts of a 3-axis magnetometer and a model of the earth’s magnetic field.
 
Here at Flintmore, we have been working with MEMS technology, including magnetometers, for years. Part of our work has been comparing MEMS sensors to the industry-grade sensors we also work with extensively. With constant development in the field, we are excited to see a point where high-performance high-fidelity sensors can fit onto a circuit board the size of a thumbnail. For some instrumentation types, this is a world we already live in.
 
Do you have a complex instrumentation or test task you need input on? Flintmore has extensive experience in getting the best from sensors and tailoring systems to meet challenging requirements. Get in contact today.

Further Reading

Two papers that present designs for lorenzt force MEMS sensors giving an excellent impression of the complexity of the arrangements of these devices to get the behaviour we want:

https://transducer-research-foundation.org/technical_digests/HiltonHead_2010/hh2010_0045.pdf

https://iopscience.iop.org/article/10.1088/0960-1317/26/10/105021/ampdf

The wikipedia articles for magnetometers, and for TMR are decent, but unfortunately I cannot recommend them as strongly as the articles on other sensor types. But they do provide good overviews and alternate explanations:

https://en.wikipedia.org/wiki/Magnetometer

https://en.wikipedia.org/wiki/Tunnel_magnetoresistance

This article on TMRs from TDK is brilliant, I would check it out for a more in depth explanation of the function of TMRs.

https://product.tdk.com/en/techlibrary/productoverview/tmr-angle-sensors.html