Designing With an Inductance to Digital Converter

Most of us are familiar with inductive detection circuits. There are standard devices used to detect a gear (toothed wheel).

These sensors are mounted in close proximity to the rotating gear. As each tooth moves by the coil and its permanent magnet pole-piece, a small current is induced into the coil. The gear speed can be measured easily and reliably.

There are coin validators (part of the mechanism in a vending machine) that check the denomination of the coin you dropped and whether it is real or counterfeit. The technology is similar. A coil wound on a permanent magnet core detects the coin's properties as it rolls by. Though the coin is (probably) not made of an iron alloy, eddy currents induced into the coin as it rolls by will produce a detectable counter-magnetic field.

There are inductive loops buried in the road to detect when vehicles drive over them.

These work reasonably well for detection of cars, OK (mostly) for motorcycles, and poorly for bicycles unless designed specifically for that purpose.

All these devices use simple circuitry to monitor the inductor. With the permanent magnet version described above, the AC voltage is amplified and used in subsequent processing circuitry. Other versions have an AC current applied to the inductor (referred to as an excitation current). Then metallic objects in proximity affect current flow in the inductor. The traffic loop detector uses this method.

For an inductive sensor, you could measure the inductance of the sensing coil and use that data to draw conclusions regarding what's in its vicinity. This process is generally more complex. (See: Z Meter on a Chip? Impedance Meter Phase Detectors and Z Meter on a Chip? Impedance Meter Bridge Circuits.) It doesn't work very well when the inductor's quality factor (Q) is low. Q indicates how pure the inductor is — how much series or shunt resistance is associated with the inductor.

Texas Instruments has devised a way to measure inductance that simplifies this picture. It has released what it calls an inductance to digital converter. TI uses the acronym LDC, with “L” for loop instead of “I” for inductor. Presumably, this is so you don't think the company is selling insulation displacement connectors. (Don't confuse TI with Tyco.) The LDC1000 can be used in various proximity-sensing applications such as those described above.

The device is accurate enough that it can measure the inductance of a coiled spring through a large range of compression. That means you could use it as part of a weigh scale. The inductance is measured to within a fraction of a percentage point using a 24-bit delta-sigma ADC, so the inductance of a spring in compression over a small range actually is a quite practical and realistic application. Similarly, you could measure the amount of compression of a spring that's part of a vehicle's suspension. This could find use in an active suspension system — the type that adjusts the spring force and the damping factor of the shock absorbers for various loads and road conditions.

The cost is low enough (less than $3 in small quantities) that it lends itself to some related rotary or slide applications. You could use it where you might otherwise use an A-quad-B encoder or potentiometer. You'd need to fabricate some additional metal parts, plus a shaft or slide mechanism assembly, for such applications.

TI also has an Eval board with a nice user’s guide available. That should get you thinking about other uses. I may procure one of the Eval boards and try the following experiment.

Some of you probably know how railroad signaling systems work, so you can ignore this brief review. Track occupancy is detected by applying low-voltage DC or AC across insulated sections or blocks of track. With rolling stock present, the rails are shorted, a relay drops out, and appropriate signaling occurs. The more sophisticated systems don't simply look for that short (threshold detection). They actually measure the resistance and the rate of change of the resistance. This is useful for grade crossings — the warning time can be adjusted depending on the train's speed.

I wonder what the inductance of a section of track is — the approach circuit for a grade crossing. This is an oddly shaped inductor. It's a rectangle that is 4 feet, 8.5 inches wide and (in one example I used) 550 yards long. Exacerbating this is its low Q, complicated by parasitic capacitance to ground (literally earth, not circuit common) of varying conductivity. I estimate it's somewhere between 600μH and 1.5mH. If you could measure that inductor and monitor the slow changes that occur between rainy and dry weather and null them out, you might have a more reliable track occupancy detection system. I'll try it and see.

Do you have any applications for an IC like this? What comes to mind?

Related posts:

25 comments on “Designing With an Inductance to Digital Converter

  1. eafpres
    October 10, 2013

    Hi Brad–interesting thoughts on detecting inductance variation of a coil spring in an automotive suspension.  Of course, similar to your rail application, there are a host of environmental factors to account for.  Presumably the large temperature variations, presence of water, snow, and ice, along with mud and other contaminants, would affect the measured inductance.  Perhaps these challenges could be features, however,  If the coils springs around the car could be sensed in this way, it would be additional data for the environment in which the car is being operated.  Combined with other sensors in the car, possibly signatures could be developed from the inputs of multiple sensors, allowing more awareness of the likely road conditions (hot/cold, wet/dry, ice/snow etc.).  Such data could be used by traction control in addition to adaptive suspension.

  2. eafpres
    October 10, 2013

    @Brad–regarding your oddly shaped and sized inductor.  I'm not clear on how you get a circuit when no train is present.  Are you suggesting a connection shorting the two rails together at the far end of the section, therby forming a large single loop with a high aspect ratio and all the other problems you mentioned?  Or is it only an inductor when rolling stock come along?

  3. Netcrawl
    October 10, 2013

    Aside from environmental considerations  I think there's something how about crosstalk between loop? this one could be the most serious inhibiting factor.

  4. samicksha
    October 10, 2013

    TI is continiously gaining grip on analog and if we talk about LDC, focus of the technology is on detecting moving parts and i guess much more suited for automotive industry.

  5. Brad_Albing
    October 10, 2013

    @eafpres – Certainly it's a hostile environment, but like you said, I think a signature could be developed. I hadn't specifically carried that line of thought out to accomodate traction control, but that's the next logical step.

  6. Brad_Albing
    October 10, 2013

    @Netcrawl – Environmental problems are alwas difficult to overcome. Regarding the crosstalk issue, between which loops do you expect there will be crosstalk? I can probably think of a way to work around the problem, but I was wondering which part you meant specifically.

  7. Brad_Albing
    October 10, 2013

    @eafpres – D'oh! I omitted that detail in my simplified explanation – forgot that not everyone has an image of these ckts in their head. Yes, there is functionally a shunt at the far end of the loop (in my example, the end farthest away from the grade crossing). In actual present day applications, there are insulating rail joints and a (large value) capacitor across the rails. Then the excitation current is around a few hundred Hz to perhaps a few kHz.

  8. Brad_Albing
    October 10, 2013

    And then add the rolling stock as a sliding shunt and the inductance steadily rises as the train approaches the grade crossing.

  9. eafpres
    October 10, 2013

    @Brad–so, you have a willing rail crossing nearby?  Be sure to watch for trains before you go hook your gear up.

  10. Brad_Albing
    October 10, 2013

    I have friends at a local tourist railroad that I can talk to and get access and run some tests.

  11. Netcrawl
    October 11, 2013

    Trying would be a good idea, just to find out how it works, well I got friends in railroad industry, hope to meet them and discuss this one. 

  12. Davidled
    October 12, 2013

    ->interesting thoughts on detecting inductance variation of a coil spring in an automotive suspension.

    I thought that there are two topics regarding on suspension application. Firstly, wire engineer might review the bracket location near to vehicle suspension to mount the sensors that used to detect gear speed. Lastly, before applying the algorithm, simulation engineer might need to review all electric current data in the different road condition.

  13. Brad_Albing
    October 13, 2013

    @Netcrawl – yes – let me know what they say too.

  14. fasmicro
    October 14, 2013

    @samicksha >>focus of the technology is on detecting moving parts and i guess much more suited for automotive industry.

    Why do you think it is more suited for the automotive industry? The auto industry is extremely regulated and not a place for a lot of experimentation

  15. fasmicro
    October 14, 2013

    Generally, the guys in the biomedical world should take a look. There has been the idea of having inductive coupling to power some of the wearables but at the end, it offers nothing readily useful. There are some good ideas in this article they can pick.

  16. RedDerek
    October 14, 2013

    @B_Albing – There are groups out there that run miniature rail roads. These could be another resource for testing. On the Electronics101 Yahoo group there are a few members.

  17. Brad_Albing
    October 15, 2013

    @RedDerek – I thought of that, but the scale is so much smaller that it wouldn't be a fair test. The parasitic capicitance and leakage resistance is so much lower on the scale models that it would produce results that would look much better than what you'd get at prototype scale.

  18. PieterBKK
    October 21, 2013

    I started to design a guitar pick up

    (actually two)





  19. Brad_Albing
    October 21, 2013

    @PieterBKK – you'll want to operating with a suitably high resonant frequency (sensor freq.) and with a suitably short response time (settling time) on the digital filter (response time set to 010 or 192). I have concerns about interaction between the pickup coils – the excitation frequency of one may cause the adjacent one to “pull” towards it. Just my 2 cents worth….

  20. yalanand
    October 27, 2013

    Inductance-to-digital converter (LDC) allows inductive sensing – a magnet-free, contactless, sensing technique that could measure the position, composition or motion of a metal or conductive objective as well as detect the extension, compression or twist of a spring. 

  21. Brad_Albing
    October 29, 2013

    @yalanand – yes – I think some of the applications with springs look pretty interesting – like weigh-scales and vehicle suspensions.

  22. SunitaT
    October 29, 2013

    The LDC1000 allows inductive sensing, by using springs and coils as inductive beams to deliver better reliability, better performance and greater suppleness than existing detecting solutions – at poorer system cost and power.

  23. etnapowers
    January 7, 2014

    I wonder if the inductance variation may be measured with a wien bridge's tecnique.

  24. etnapowers
    January 7, 2014

    This system looks like to work in a similar way to the MEMS, in both cases the mechanic stress in one direction is trasformed in a electrical pulse, it is a very interesting application

  25. Brad Albing
    January 7, 2014

    @etnapowers – you certainly could use a conventional bridge technique (a Wheatstone bridge is the one you're thinking of). The TI part is probably more accurate than what you could build from discrete parts to implement the bridge. And it's far more convenient.

Leave a Reply

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