Advertisement

Blog

Magnetic Field Sensors

There are many types of magnetic field sensors on the market. A brief list would include (but not be limited to) these, with the more exotic devices first:

  • SQUID (superconducting quantum interference device) magnetometer
  • Nuclear precession magnetic field sensor
  • Optically pumped magnetic field sensor
  • Electron tunneling MEMS (micro-electro-mechanical system)
  • AMR (anisotropic magneto-resistance) magnetometer
  • GMR (giant — or occasionally geometrical — magneto-resistance) magnetometer
  • Lorentz Force MEMS
  • MEMS compass
  • Hall effect device
  • Magneto-diode
  • Magneto-transistor
  • Magnetic tunnel junction magnetometer
  • Search coil magnetic field sensor
  • Fluxgate magnetometer
  • Inductive pickups

The MEMS solutions are gaining market share in the smartphone compass market, but in other markets such as automotive, other devices are more often used. Hall effect sensors (and some of the other inexpensive devices) are used as mechanical position and rotation sensors. In medical diagnostic research (e.g., brain research), the exotic sensors such as the SQUID device enable researchers to gain a better understanding of the brain functions in real-time.

In hard-drives, GMR and other sensors are used to achieve what now seems to be a phenomenal storage density. Fluxgate magnetometers have some applications in compasses where MEMS is not a fit. Many of the others have various research applications.

As we all recall from physics, magnetic field strength or flux density is measured in tesla (T) or gauss (G). T=G•104 . One tesla is one weber (Wb) per square meter where a weber is the measure of magnetic flux. Subjecting a one-turn coil to a linearly ramped down change of 1Wb across a 1s span would produce 1V at the coil's terminals — hence 1Wb = 1V•s.

(Source: Rockwell Automation)

(Source: Rockwell Automation)

Sensitivity of sensors can range from hundreds of mT for the most primitive sensors to 1•10-18 T for the SQUID type sensors. Sensitivity, accuracy, operating temperature, ruggedness, resistance to damage at high heat, and a host of other parameters come to play in selecting magnetic sensors for an application.

Hall effect sensors and simple inductive pickup coils are some of the most rugged types of sensors. The simple inductive device still gets used a lot in industrial applications — the classic setup uses a gear or toothed wheel with the pickup nearby.

Inductive pickup form factors can be large (i.e., a large aperture), though their sensitivity is relatively low compared to some of the other devices mentioned. MEMS sensors can be relatively small and quite sensitive.

A Google search for “magnetic sensor selection guide” can bring one to the different vendors' selection guides — the choices seem overwhelming.

Have you used magnetic sensors? If so, what type? How well did they work?

Related post:

17 comments on “Magnetic Field Sensors

  1. Davidled
    July 12, 2013

    Simply by using hall effect sensor, vehicle speed or engine fan speed is detected and controlled, and also, in the hybrid vehicle, hall effect current sensor is used for monitoring regenerated DC voltage as very crucial factor in order to expand HEV battery life including regenerative brake. So, I guess that in the future, hall effect sensor will be one of main sensors in HEV vehicle.

  2. BillWM
    July 12, 2013

    Hall effect sensors are a good fit where ever dirt and oil are found.

  3. Netcrawl
    July 13, 2013

    Interesting topic thanks for the share, hall effect sensor are used in industry to measure proximity of objects, speed, rotary positions, electric currents and intensity magnetic fields. These kind of  sensors are used in tachometers and electric motor. And companies like GE and Honeywell are good example of companies that perform well in this area. 

     

  4. antedeluvian2
    July 15, 2013

    I remember the National Semi (now TI) V/F and F/V converterLM2907/2917 as being quite useful in interfacing to magnetic sensors in dtetcting the rotational speed.

     

    I used regular proximity switches in a qaudrature arrangement to create a rugged linear displacement transducer. I wrote about it in the second half of my blog “Custom Sensors Enhance MCU Design” on MCC.

  5. Guru of Grounding
    July 17, 2013

    The choices do seem overwhelming. I'm looking into the feasibility of a sensor that will detect the magnetic field produced by a single conductor carrying 1 mA or less. Frequency range of 50 Hz to perhaps 5 kHz is sufficient. I could always use a Rogowski coil but I have a suspicion that a more recent type of sensor may be well-suited, yet smaller and less expensive. Any ideas to save me the tedious searching.

  6. BillWM
    July 17, 2013

    I would imagine some of the sensors used in some of the new hard-drive heads would be overkill, but possibly a GMR sensor of similar nature could be made to work.   Also some of the MEMS sensors can be small and sensitive.

  7. EMCgenius
    July 18, 2013

    There are 10^4 gauss per tesla.  The article gets this backward.  It should say T=G/10^4, not T=G•104

  8. musiklab
    July 18, 2013

    I once used Siemens SAS 231 W hall device for measuring magnetic flux in an application- it would output a voltage proportional to magnetic flux. 

    Anybody know of a similar device available today ? 

    NOW,

    I´d like to be able to measure a probe distance from a strong magnet, able to hold itself,  to measure   thickness in a range of 1-10 mm with a repeat accuracy of  0.1 mm or better, presented as a linear DC voltage,  scaled to the proximity law

    and calibrated in millimeters on a digital readout . 

    Any ideas ? Anybody?

    pls contact >>  michael@edinger.dk

  9. Brad_Albing
    July 18, 2013

    @EMCgenius – I'm confused about what you wrote. You said: There are 10^4 gauss per tesla which is true and matches what it says in the blog. But then you say It should say T=G/10^4 which is the inverse – and incorrect. A 1T magnet is a strong magnet. A 1G magnet – not so much.

  10. EMCgenius
    July 18, 2013

    We are agreed then: a tesla is 10,000 times stronger than a gauss.  Therefore divide gauss by 10,000 to get teslas.  Multiply teslas by 10,000 to get gauss.  The conversion in the article is bass ackwards.  Plug one gauss into it and see if you believe the result.  If the exponent had a minus sign it would work.

  11. Brad_Albing
    July 18, 2013

    Um… gosh – sometimes I just don't know what to say. Which for me is saying a lot. I'm sure at some point the fog will lift and all will become clear. Apparently today is not that day. Thanks for reading.

  12. EMCgenius
    July 18, 2013

    Dimensional analysis keeps us on the straight and narrow.  10,000 gauss per tesla times teslas causes the tesla unit to cancel (teslas divided by teslas is unitless), leaving only gauss.  Going the other way, 10,000 gauss per tesla is the same as saying there are 0.0001 teslas per gauss.  Multiplying teslas per gauss times field strength in gauss leaves teslas as the unit.

    In the article, the formula says teslas = 10,000 gauss per tesla times gauss, which calculates gauss squared per tesla — a nonsense result.

    Coulda happened to anybody.

  13. Brad_Albing
    July 18, 2013

    Ah… I think I see your point. The intent was to indicate that 1T = 10,000G.

  14. EMCgenius
    July 19, 2013

    In all fairness, the disconnect is simple.  If you had said 1T=10^4G, that stands alone as a fact, but without a numerical quantity on both sides, it reads that any number of gauss times 10^4 equals equivalent teslas, which is not correct.  That made all the difference.  Math is literal like that.

  15. Brad_Albing
    July 19, 2013

    Quite right. We'll be more careful with these mathmatical statements in the future.

  16. EMCgenius
    July 19, 2013

     

    All's well that ends well.  Other readers will appreciate the walk-through we provided.

  17. Brad_Albing
    July 19, 2013

    Well, they better!

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.