• State
  • Locked Locked
  • Replies 6 replies
  • Subscribers 32 subscribers
  • Views 633 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

LDC proximity reading confusion

Valentin Dimitrov
Prodigy 60 points

I have made some trial with the EVM and custom PCB to evaluate the LDC1000 for a possible use in proximity
sensig application. I need to measure the distance between the pcb coil and a metal target with 1µm
(or better) resolution. The maximum distance is about 0.2cm. The design parameters for my PCB coil are:


So far i have managed to get about 1V peak-to-peak at the CFB, set up RP_min and RP_max so that my proximity
readings are between 3000(max. distance) and 20000(min. distance). My headache at the moment is that when
the proximity reading reach the value 10922 it stay at that value for the next 10-15µm then jump
to 11000 and above. I am observing similar behavior with the EVM - when the proximity reading is 10922 the
device is "frozen" for the next 30µm. I guess i am doing something wrong here, but I'm running out of ideas
what to do next :)

  • 0
    Prodigy 10 points

    I am having the same problem. I am using the LDC1000 EVM and reading data directly (serial IO). All default settings as far as I can tell, but the EVM documentation is not clear on how chip settings are handled. As my slowly moving target approaches the coil the reading gets "stuck" at 10922 for a while and then continues. I am no seeing it jump ahead after this however; it just continues upward from 10922. My overall plot runs from about 10000 to 28000 and the part that is stuck at 10922 is about 5% of the total plot.

    Two clues about what is going on:

    1) In all other data blocks of 512 readings there is some variation from reading to reading. For the blocks that read 10922 there is not variation: it is really stuck.

    2) 10922 base10 is 10101010101010 in base2. Can this possibly be coincidental?

    Having spent a few days working with this device I now don't know if I can trust any of my readings. The rest of the curve looks generally correct but has a number of small odd features that repeat from test to test. Are these true measurements or artifacts of the sensor?

    Actually I don't know if the problem is with the LDC1000 or the programming on the EVM. I see that the previous post has not gotten a reply in over two weeks. This really needs to be addressed. Thanks for any light that can be shed on this problem.

  • 0
    Prodigy 195 points

    I can also se the 10922 artifact on my evm with standard coil. Here is an example graph. Just tried to hold the coil steady at 10922 and some wiggeling to get larger and smaler sample values.

    Settings

    0x01,0x13,,Byte,Rp_MAX
    0x02,0x3B,,Byte,Rp_MIN
    0x03,0xBE,0x94,Byte,Sensor Frequency
    0x04,0x15,0x17,Byte,LDC Configuration

    Regards Torgeir Skomedal

  • 0 in reply to Michael Wollowitz
    TI__Genius 9300 points

    Could you reduce RpMAX setting by 1 or 2 codes (1-2 steps up in resistance) and see if it helps?

  • 0 in reply to Evgeny Fomin
    Prodigy 195 points

    The same result with RpMax = 38785ohm

    I use 10k uC samplerate. Xaxis is sample number.

    0x01,0x11,,Byte,Rp_MAX
    0x02,0x3B,,Byte,Rp_MIN
    0x03,0xBE,0x94,Byte,Sensor Frequency
    0x04,0x15,0x17,Byte,LDC Configuration

    Regards Torgeir

  • 0 in reply to Torgeir Skomedal
    TI__Genius 9300 points

    Hello Torgeir,

    There are two ways to fix it.

    1st, is to redesign the sensor (LC tank) to achieve higher Q value.

    If that is not possible, for the final application, when you have a well-defined range of motion, adjust Rp parameters so that 10922 is excluded from the useful range of Rp values.

    It is also worth taking a look at inductance measurements - perhaps they give you enough resolution?

  • 0 in reply to Evgeny Fomin
    Prodigy 195 points

    Thank you for the suggested solutions.

    For my applications it is not a blocking issue.

    The issue is such that testers wonder what it is, so it is OK to get it documented.

    As you see the glitch at 10922 is not that big compared to normal noise in the sample value region.

    And i have to say the noise at close distance is impressive low.

    Regards Torgeir