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.

LDC1614EVM: LDC1614 EVK use with custom coil design

Part Number: LDC1614EVM
Other Parts Discussed in Thread: LDC1614

Hi TI team,

I am working on one product design where we need to identify four different objects on a 2x2 coil array (similar to the keypad design). I explored and found that this can be achieved with inductive sensing and the LDC series is the right choice for this application.

I started with the LDC1614EVK and my custom-design coil. I designed the coil from the coil design tool on the TI website. Following is the 2x2 coil array with coil details.

With this I am facing the following issues:

  • I connect only a single coil on CH0 and it is detecting perfectly on the Sensing EVK GUI and no variation in inductance in GUI data.
  • When I connect the 2x2 array on CH0 and CH1, inductance starts getting vary and not stable on the GUI data. It has a huge variation. Can you please let me know how can I make it stable?
  • Also, my application is to detect 4 different objects, so I am planning to use different wire Inductor values inside the object i.e. 100uH, 200uH, and 300uH, etc. My object and coil distance is less than 10mm. Due to variations in the inductance of the, I am not able to detect changes of inductance when I place an object on the coil.

Can you please guide me on how to make 2x2 array coil inductance stable? Also, with different inductor value, Can I able to differentiate the objects?

    

Total PCB thickness=0.8mm and FR4 material.

Looking forward to your valuable feedback to solve the issue.

Thanks,

Hitesh

  • Hello Hitesh, 

    I have some clarification questions about your application: 

    • What drove the decision to use a 2x2 array? Are you sensing the 4 different objects in the four different spots or does the object cover all four sensors? 
    • Your coils seem to be arranged in a combination of series and parallel configurations. Is there a reason you did this instead of having four separate coils broken out to four separate headers? 
    • What material are your objects made of? Is there anything that differentiates them besides what you are putting inside them? (like size or material)

    Thank you, 

    Justin Beigel

  • Hi Justin,

    Thanks for your quick response. Please see my inline response:

    What drove the decision to use a 2x2 array? Are you sensing the 4 different objects in the four different spots or does the object cover all four sensors? 

    - Yes, I am using 4-different objects in the four different spots.

    Your coils seem to be arranged in a combination of series and parallel configurations. Is there a reason you did this instead of having four separate coils broken out to four separate headers? 

    -  This is POC after that I need to build a product where I can detect 10 different objects on a total 4x4 (16-square) array at that time I will use 2-analog mux switcher and one LDC chipset (similar concept to Keypad EVM).

    • What material are your objects made of? Is there anything that differentiates them besides what you are putting inside them? (like size or material)

    - Material is plastic and inside I am planning to add an inductor coil or metal piece (0.5mm to 1mm) at bottom of it. No, the object size will remain the same.

    Thanks,

    Hitesh

  • Hello Hitesh, 

    If you have different objects over each sensor, then it will be better to separate out each coil to a single channel. In your customer board to the EVM, this would look very similar to the EVM's coils: 

    When scaling to the 4x4 matrix, you would just need to MUX 4 of the sensors to each channel. 

    Best Regards, 

    Justin Beigel

  • Hi Justin,

    I am wondering why the 4x4 keypad matrix thing won't work. In the future, if they want to increase the matrix then have too many MUX ICs. Just trying to understand such solution work for object detection application or not.

    Thanks,

    Hitesh

  • Hello Hitesh, 

    If they wanted to keep with the parallel sensors approach, it can be done but you do not get information about which sensor the target is over for sensors that are connected to the same channel. It would also require changing the connection in the sensor layout. With the layout you shared, the top and bottom layer are not consistent with which sensors are in parallel. You could change the layout so that the two top  sensors are parallel to each other then connect that to one channel of the device. This will mean that any change over either of the two sensors will be identical to the LDC1614 but would increase the area of sensitivity to cover the larger grid without as many MUX in the circuit. 

    However, if they can use 4 MUX ICs, then it would give them more information about the target position. 

    Best Regards, 

    Justin