• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 1 answer
  • Subscribers 31 subscribers
  • Views 861 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.

FDC2212: Fluid level sensor EMI rejection

rednblack
Prodigy 30 points
Part Number: FDC2212
Other Parts Discussed in Thread: FDC2214, FDC2214EVM

Hello,

 

We have an application using the FDC2212 to measure fluid level. The design is functionally OK but we are having problems with EMI rejection.

The design uses similar component values to those in the FDC2214 EVM. The sense electrodes are implemented on the bottom layer of the product PCB; there is a ground plane on an adjacent layer giving a capacitance offset of about 150 pF. The total change in capacitance due to fluid level is 2 – 3 pF. The LC tank uses a 22 uH inductor giving a resonant frequency of about 2.7 MHz.

 

The product has to pass EMI immunity testing to IEC 61000-4-6. This test injects a disturbance of 3 VRMS onto the product cable using a current probe, over a frequency range of 150 kHz to 80 MHz. The frequency is swept in steps of 1% of present value, with a dwell time of 1 second at each step.

 

The sensor PCB, incorporating the FDC2212, a small micro, linear regulators etc, is on the end of a 0.75 metre cable. The sensor is attached to the fluid vessel; the fluid in the vessel usually has a connection to ground but this is not well controlled and has an unknown impedance.

 

If I log the measured capacitance during the test sweep, starting at the second harmonic of the resonant frequency I get a series of deviations which are larger than my entire level measurement range – see attached plot.

 

The second channel of the FDC2212 is connected to an identical tank circuit and a pair of electrodes of the same dimensions as the sense electrodes, but implemented on a buried layer of the PCB with a ground plane above and below. This was intended to be used to cancel any changes in the LC tank values with temperature. This channel shows no disturbances at all during the test.

 

The FDC2212 is configured for differential operation on both channels. Amplitude measures 1.4V using a differential FET probe. 

The error flags in the status register are zero throughout the test.

 

The PCB layout and sense electrode design has been kept symmetrical as far as possible.

The problem is clearly that the common-mode disturbance is being converted to differential mode. I am curious why this behaviour only starts at the 2nd harmonic of the resonant frequency, there is a very slight disturbance at the fundamental but it is barely measurable.

 

We chose the FDC2212 for its advertised high intrinsic immunity to EMI. Are you able to offer any advice?

 

Thanks

 

  • 0
    TI__Mastermind 23810 points
    First you should try adding filtering to the sensor to improve performance. This guide focuses on inductive sensing, but you can start at section 2.2 for the FDC2212:

    www.ti.com/.../snoa962.pdf
  • 0 in reply to Clancy Soehren
    Prodigy 30 points

    OK, thanks.  I have the FDC2214 EVM on order, I notice this has pads defined to fit the common-mode filter components described in the app note.

    The trouble is that this is a very space-constrained design.  I have an area maybe 4 x 5 mm in which I could implement this filter.

    From the plot above I need a filter which gives at least 25 dB attenuation at 4.5 MHz, 36 dB @ 14.3 MHz, 30 dB @ 55 MHz.

    The largest filter capacitor I can use without losing too much sensitivity is maybe 33 pF.  So to give 25 dB @ 4.5 MHz I would need a CM choke of at least 670 uH.

    All the parts I have looked at from Bourns, Wurth Elektronik, etc in this inductance range, have a self resonant frequency of at most a few MHz, or are physically much too large.

    Do you have any recommended common-mode inductors for use with the FDC2214 EVM, or any other suggestions?

  • 0 in reply to rednblack
    TI__Mastermind 23810 points
    Would it be possible to try one of the larger inductors first to verify that it helps the problem? From there, you can scale back and try smaller footprint options.
  • 0 in reply to Clancy Soehren
    TI__Mastermind 23810 points
    Let me know if you still need help. Otherwise, I am closing this thread.
  • 0 in reply to Clancy Soehren
    Prodigy 30 points

    OK, I have had some encouraging results using the FDC2214EVM and a sensor assembly with the same dimensions as in the product.

    The response with no EMI filtering:

    (the vertical axis is the full range of level measurement).

    This is not a full compliance test, but is the closest we can manage in-house by injecting the disturbance using a function generator and a RF current probe.  The disturbance seen in a compliance test is 5 - 10X greater.

    The response with a filter comprising a Bourns SRF3216-222Y common-mode choke and two 33pF capacitors, fitted between the tank circuit and the FDC2214:

    As above, but using a Wurth Elektronik 744222 common-mode choke:

    This is certainly worth pursuing, and works a lot better than my initial calculations would suggest.

    Many thanks for your help.