• State TI Thinks Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 30 subscribers
  • Views 326 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

LDC1614: Using multiple coils to remove external interference

Jordan Berke
Prodigy 75 points
Part Number: LDC1614

My LDC1614 implementation uses three coils to sense a single target. Since we have multiple coils, we can compare them to remove ambient interference. I'm wondering if there is a best practice approach for this, as I can see drawbacks in the methods that we're currently using. I'm attaching pictures pre and post-filtering.

Graph 1:

Graph 2:

Here are my questions:

1. Is this an expected amount of interference? I've done extensive prototyping on the EVM and I never saw as much as I'm seeing in graph 1. The EVM essentially looked exactly like graph 2 without special settings or math. That makes me think there's some functionality on the LDC side that I'm unfamiliar with, that is already filtering some or all of this interference.

2. If there is no internal interference subtraction with reference coils, is there a best practice approach to this? If we use LDC1 as a baseline and do a diff, we're essentially operating such that the system is at most as reliable as LDC1. If we do the following:

LDC1filtered = LDC1raw - LDC2raw
LDC2filtered = LDC2raw - LDC3raw
LDC3filtered = LDC3raw - LDC1raw

Each sensor is only taking advantage of one reference sensor at any time. In both calculations, as the target approaches any one coil, that signal will be subtracted from the other coils as well. That's the source of the big difference in the shape of each of these curves between graphs. 

  • 0
    TI__Genius 17440 points

    Hi Jordan,

    It would help me answer your questions if I had some additional information about your setup: 

    • Are all of your coils the same size? Are they they same distance from the target? 
    • What is the target? How does it move in relation to the coils?
    • Are you using the internal oscillator or an external oscillator?
    • Are you trying to sense the absolute position of the target?

    General information about your application would also be helpful.

    Best Regards,

  • 0 in reply to Kristin Jones93
    Prodigy 75 points

    Hi Kristin, thanks for your reply. I sent you a DM with some specifics but I'll put the general answers here.

    1. Yes, all 10mm 4 layer coils. The target moves closer and further from each of the coils as its position moves. Almost like triangulation.

    2. The target moves at an angle closer and further from the coils.

    3. Internal

    4. Trying to sense the angle of the target, so essentially yes the position of the centroid of where the target interacts with the inductive field.

    But more generally, we're seeing a very clear noise signal that impacts all coils the same. You can see essentially each peak in the graphs above are shared between all signals. That's why when we do a simple diff, we get such clean signal. What I'm trying to figure out is:

    1. How should I go about diagnosing the source of this noise? Where would you personally begin? My coils are relatively simple, and do not result in this interference pattern when connected to the EVM. The noise exists when the target is stationary. 

    2. What methodologies can be used in firmware to remove noise that is so clearly impacting all sensors? It all seems like very low frequency noise, so I'm not sure if we can identify and remove similar frequencies.

  • 0 in reply to Jordan Berke
    TI__Genius 17440 points

    Hi Jordan,

    Thanks for providing the requested information. I'll respond to your private message as well.

    I think the biggest source of your noise is the internal oscillator. We use an external oscillator on the EVM that has less jitter, and we recommend that all precision applications use an external oscillator. The noise in your first measurement looks fairly typical of the internal oscillator, which is why I asked about it.

    If you have a footprint available for an external oscillator, I suggest populating it. If not, I suggest continuing your testing with the EVM until you can create another iteration of your board.

    The best way to remove this noise is to use an external oscillator. If you need to remove noise due to changing temperature as well, I suggest using a ratiometric approach instead of a differential approach with your reference sensor.

    Instead of LDC1filtered = LDC1 - LDC2 you would use LDC1filtered = LDC1/LDC2

    Best Regards,