• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 30 subscribers
  • Views 1633 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.

Urgent. problem on using FDC2214 to do people approaching detection

zhao Steven
Prodigy 90 points
Other Parts Discussed in Thread: FDC2214

Hello,

I'm making feasibility study of FDC2214 used for our people proximity detection, based on the EVM got from you. And, get some problems or something not yet understood why it is like this. If convenient, would you pls give some advice or comments? Many thx.

Q1, too small variation of capacitance after hand approaching

just around 0.5pF around  increasing, or much less. distance is 5cm away from EVM, If use bigger copper pad, the only change is from the base capacitance without approaching, eg, from 45pF to 52pF, but have no improvement on the variation range during approaching----still around 0.5pF.

this variation is too small, compared to the one during our EMC test mentioned in Q2 below. How to increase this value? or how to make our algorithm design based on this value, to guarantee the good performance to withstand electromagnetic interference outside?

Q2, EMC test----immunity to conductive disturbance (IEC61000-4-6)

We performed this test on the EVM board. Inject the noise into only 5Vdc and GND of USB cable. GUI shows that the capacitance variation would be up to 2pF around at resonant frequency (5MHZ around). This variation is far more beyond the 0.5pF variation during normal hand approaching. So, I don't know how to use our sensor in the real product. 

  • 0
    TI__Expert 8050 points
    Hi Steven,

    To improve sensitivity, you may take off the capacitor in the LC tank (C17 for Ch0). For more details on proximity detection of a hand, please refer to this application note: www.ti.com/.../snoa940

    You may consider using common-mode chokes to reduce noise coupling on the USB side.

    Regards,
    Yibo
  • 0 in reply to Yibo Yu
    Prodigy 90 points

    Hello Yibo,

    Many thx for your quick reply.

    There is no good on improvement after taking off C17, even the red LED on board is switched on after that.

    Also, the calculated capacitor is far away from the actual value----33pF. It shows that the calculated value could be up to 50pF around, which is 17pF high than the actual one.  All data is got from the EVM board FDC2214 of TI. Why? Meanwhile, the variation of capacitance, during hand approaching & away, is just around 0.03 ~ 0.06pF.  I can't use this sensor if the error is so big and variation is so small from hand approaching.

    The copper pad size is aournd 3~4cm^2, distance is 5cm between hand and copper pad.

    Urgent. Many thx.

  • 0 in reply to zhao Steven
    TI__Expert 8050 points

    Hello Steven,

    The red LED simply indicates the sensor frequency is different from the default value. You can ignore it.

    The difference is due to the parasitic capacitance of the PCB. That's also why the LC tank can still oscillate after taking off the 33pF cap. We were able to get around 15-20cm detection range with a 3-4 cm^2 sensor after taking off the 33pF cap. Please ensure you have a solid ground connection in your system, and the setup is mechanically stable. Again, if your noise is much more significant than that in the application note, please consider using common mode chokes to suppress the noise. The 0 Ohm resistors near the LC tank (R33, R23 for CH0) are meant for that purpose. You can replace them with a common mode choke, e.g. SRF3216-222Y. On the bottom are place holders for additional filtering caps. For example, you can populate C37, C38 in case of EMI susceptibility issues with long wires, or C39, C41 in case of EMI emission issues with long wires.

    Regards,

    Yibo

  • 0 in reply to Yibo Yu
    Prodigy 90 points
    Hello Yibo,
    I need to confirm with you on two points, before further consulting you.

    Q1, what's your definition of "sensor detection distance", from capacitance variation point of view? 5 femto-fara? 10ff? or several hundreds of ff level?
    I think, practically, there is no good if it's lower than 50ff (better >100ff), since it could be up to 20ff during the EMI immunity test (showed in the GUI, using the FDC2214 EVM board, 3Vrms interference injection @ 5Vdc & GND wire). Otherwise, algorithm in MCU is too difficult to design, even impossible.

    Q2, what's the calculated capacitance value showed in the "CONFIGURATION" sheet of your GUI?
    just consider this configuration: FDC2214 EVM board without any modification (33pf//18uH); using USB cable to connect with laptop;
    laptop is energized by common isolated power supply adaptor, which is connected with AC network (Line, Neutral, Earth 3 wires).
    In my side, it's showed 50pF around. whatever the PCB is directly placed on the desk, or above paper box.

    Now, let's go back to your last reply. I performed tests based on your advice----add common choke & C37 & C38, even remove the C17, but have no big change on the capacitance variation----still the 30ff variation @ 5CM distance. Of course, the calculated total capacitance is changed to 62pF around, which is due to the newly added C37 & C38, I think.

    Further, if convenient, could we have a short phone call? typing like this is too low efficient. If ok, pls send me by email: nudt302@sina.com
    Appreciation.
  • 0 in reply to zhao Steven
    TI__Expert 8050 points
    Hi Steven,

    1. We define the minimum detectable distance to be a change of the peak-to-peak noise. It is less meaningful to define an absolute number because what really matters is the signal-to-noise ratio.

    2. Different laptops and cables do vary in terms of noise injection. In any system, there are always various sources of noise which need to be carefully considered.

    We always want to help our customers as much as we can. Please understand that there's certain procedure we need to follow for a large company like TI. Would it be possible for you to discuss with your local TI sales rep/Field apps your system requirements and schedule?

    Regards,
    Yibo