This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

LDC1000EVM Linear Position Sensing Target Requirement

Other Parts Discussed in Thread: LDC1000EVM, LDC1612

Hi,

I am planning to use LDC1000EVM to measure linear distance with a range of 0-30mm. I have a few questions as follow,

1) Do I have to use a specific type of metal for the target, knowing the inductance,resistance of the target too?

2) What is the best shape for the target? Is rectangle okay?

3) With a designed target, how can I get the sensor output versus distance curve? Will the curve be linear?

Thank you!

  • Hello Huichen,

    The LDC1000 is only recommended for metal type identification applications. For all other applications, I recommend using one of our newer devices; In your case, I recommend the LDC1612, which offers many advantages. The high-resolution multi-channel LDC1612 is an L-only devices which offers greatly improved measurement accuracy, better part-to-part variation, lower power consumption, easier design integration, and multiple channels.

    Here is an apps note which covers linear position sensing in detail: http://www.ti.com/lit/an/snoa931/snoa931.pdf

    Here is more information on the sensor design: http://www.ti.com/lit/an/snoa930/snoa930.pdf 

  • Hey Ben,

    Thank you for your recommendation. In fact in my application, the linear position sensor will be used to sensing a quasi-linear motion of the end of a rotational arm under small angle approximation. So if LDC1612 lateral linear displacement measurement technique is used, the dx variation is 0-30mm, dz variation is 0-1.1mm. May I know what I can do to reduce errors and increase the accuracy of measurement?

    Thank you!
  • Hello Huichen,
    the LDC1000 can be used for linear position sensing, but our newer devices such as the LDC1612 offer superior performance, a second channel, and easier system design integration. A linear position sensing system with the LDC1612 will have better system accuracy than the LDC1000. As an added bonus, you will find that the LDC1612 is less expensive than the LDC1000.

    A dz variation can be reduced by using a second coil of identical geometries, which either arranged in inverse direction to the first coil, or is placed on the opposite side of the rotational arm. Does your system allow for placement of a second coil in such fashion?
  • Yes it allows a placement of a second coil, however I am not sure about the configuration, so I am attaching a photo here for you to evaluate.
    www.dropbox.com/.../AAADv8SRTTJ-TKNsQ85kORd3a

    In the photo, the part including target 1 will have dx variation of 0-30mm, dz variation of 0-1.1mm. And the part including the target 2 will hence have a dz variation of 0-30mm, dx variation of 0-1.1mm.

    I am not exactly sure how to compensate this dz variation through two dimensional lookup or curve fitting. Could you elaborate a bit on that?

    And at the same time, my rotational arm is made from aluminium. Is there any kind of insulation required between the target and the aluminum? Will aluminum affect the sensor performance?

    Thank you!
  • Hello Huichen,

    I don't have access to dropbox; can you please add the picture to the forum thread?

    I don't know what your system looks like, but here is a generic example of how a second coil in opposing direction can be used to improve the z-axis dependence.

  • Hi Ben,

    The information you provided is very helpful! I have come up with two design options, could you please take a look at it? The first design is pretty much similar to the example you provided. The second design uses circular coil, placed at two sides of the rotational arm.

    And in the mean time, could you advise if any insulation between the aluminum target and aluminum balance arm is required? Do I need to carry out a calibration process as the example you suggested in your post before I conduct the real measurement? Which coil shape do you think is better for my application?

    Thank you so much for your help so far!:)

  • Hello Huichen,

    The triangular target (similar to what you've drawn in figure 1) gives better resolution than the stretched coil approach from figure 1. However, the better approach is typically the one with more controlled mechanical tolerances. I would expect that it is easier to control tolerances with the first approach because it does not require a second target.

    In both drawings, your target moves closer/further away from the coil as the arm spins, which was not the case in the compensation information that I sent previously. We don't have simulation or measurement data on the suitability of the approach in this case.

    The aluminum arm will certainly be sensed by the coil if it is within approximately 2 coil diameters away. If both coils see the arm at equal distance at all time, this will simply result in a fixed offset and a reduction in dynamic range, which may not be a problem. If it is not seen by both coils equally, then it would interfere with the measurement and would likely need to be dealt with. Here's some information on shielding from interfering metal: e2e.ti.com/.../inductive-sensing-how-to-shield-from-metal-interference

  • Hi Ben,

    If I understand correctly, what you meant by "target move closer/further away from the coil" actually refers to the fact that target is not parallel to the coil surface as it will be in a certain angle as arm rotates? It is a small angle rotation though, typically around 1 degree, so I am assuming a quasi linear motion here, and the dz variation maximum should be 1.1mm.

    And do you have any way to over come this in my application of sensing end displacement of the balance arm? What if I sense the displacement using axial measurement, the center shift should be only 1.1mm while my measurement range is 30mm. Let me know if you have some ideas. :)

    Cheers,
    Huichen

  • Hello Huichen,
    correct, I referred to the z-axis variation due to the angular rotation.

    As you correctly pointed out, an alternative method for improving z-axis variation is to do a distance measurement with one coil and a linear position measurement with another coil. Both compensation techniques are possible and should result in higher system accuracy compared to an uncompensated system.
  • Hey Ben!

    I decided to try with the two stretched coil method as I described previously in my post. I am looking for some sources for me to design the stretched coil but I cant find it. It is said in the LDC linear position sensing notes that, there should be PCB layout scripts available on TI website. May I know if you can point me to some link that helps design the stretched coil? The design tool seems to only be able to design circular, square, hexagonal, octagonal coils.

    Thanks!
  • Hello Huichen,

    here are the scripts that I used to generate the coils. The scripts can be used in Altium directly. They may need some adjustment in different CAD tools. 

    LDC coil scripts.zip

  • Hi Ben!

    I have actually decided to use circular coil and circular target, and measure the axial distance, because it seems that TI has more resources on it, including the WEBENCH design tool.

    Currently I have LDC1612 EVM Module. It has two sample coils ( 13.9mm in diameter, 19 turns per layer, 2 layers,0.15mm for both trace width and trace spacing, 330PF for C). Since my measurement distance is about 28mm. I am wondering, with the above sensor information, how can I generate the inductance vs distance plot, and resolution vs distance plot as the WEBENCH tool give us when we design a coil? The target material is selected as copper.

    And another question is, even a theoretical curve is obtained, when using the sensor together with the coil, do we have to do some calibration before hand? Such as moving the target in 0.5mm step and plot the inductance vs. distance curve? So when sensing the axial distance, we can obtain the distance value from the inductance reading on the GUI.

    Thank you!