• State TI Thinks Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 32 subscribers
  • Views 1068 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.

FDC2112: Reducing radiated emissions from sensor plates

Hadyn van der berg
Prodigy 100 points

Part Number: FDC2112

Hello

We are using the FDC2112 in a liquid level application. We are using a differential measurement, 2 plates inside a vessel and they have direct contact with the liquid, and the device and application is working.

However, we are having problem when trying to pass CE EMI certification since we get radiated emissions above the allowed level. Our investigations show the noise is coming from the sensor plates attached to FDC2112 and the noise disappears when FDC2112 stops it's measurements.
In addition, the harmonics are 16.8MHz multiples, roughly double the resonance frequency. Interestingly, the radiation is large when the sensor is dry. When we wet the sensor (ie, place liquid into the vessel), the noise reduces significantly , under the limit (and hence we can pass in that state).

I see in another forum post (Chris Oberhauser66 e2e.ti.com/.../2432400 there are 2 suggestions:

  • ferrite beads in series between the FDC2112+LC tank, and the sensor
  • And caps to ground from IN0x pins.

Our design aim was to reduce any parasitic capacitance or inductance attached to the circuit, but here I see additional inductance and capacitance being added.

Our circuit is as below :

We use the internal oscillator, but calibrate our measurement including the internal oscillator frequency on the production line.

1a/ Can you explain what purpose the ferrite beads perform?

1b/ Can you explain what purpose the cap to ground perform? 

1c/ How should we chose the ferrite beads.


2/ Presumably our ESD clamp diodes will be functioning similar to the caps to ground? Although these components have around Cjunction = 0.5pF , so not 1/2 the nominal sensor capacitance (as mentioned in that post).

Our measurement range is typically 2-25pF sensor capacitance after calibration (including offset) . However, the offset capacitance is around 30pF.

2b/ Do I understand it correctly then that our nominal sensor capacitance is 30 + ((25-2)/2) = 41pF? 

3/ Do you have any suggestions on how to reduce the radiated noise (in particular in our dry sensor case)?

4/ Can we adjust any registers to reduce the noise? I'm wondering if slowing down the sampling will reduce the average power output and therefore reduce the spikes? I guess I should also recheck the Vpk amplitude.

5/ Regarding the drive strength and oscillation amplitude, is it possible to increase the drive strength from small to large via SW, and monitor the Amplitude High and Low Warnings, and then select a mid-point drive strength between those? Algorithmically by SW that is possible, but in practice will that give a correct /sufficient drive strength?


Many thanks,
Hadyn van der Berg

  • 0
    TI__Genius 17440 points

    Hello,

    Thanks for reaching out. I'll be able to provide more detailed answers to your questions next week.

    In the mean time, some of the suggestions in this EMI application note may be helpful to you. The coil shielding will only be effective for inductive sensing, but the other options are suitable for the FDC221x as well as the LDC devices.

    http://www.ti.com/lit/an/snoa962/snoa962.pdf

    Best Regards,

  • 0 in reply to Kristin Jones93
    Prodigy 100 points

    HI Kirstin,

    Thanks for the reply. Just to let you know that we retested and there is no correlation between the plates being wet with liquid, or being submerged in liquid, and reduced EMI emissions. We get nearly the same level of EMI emissions whether the sensor plates are dry or wet. 

    Apologies, an incorrect measurement.  If the FDC2112 is taking measurements, we get EMI emissions above the allowed limit.

    Thanks,

    Hadyn

  • 0 in reply to Hadyn van der berg
    TI__Genius 17440 points

    Hello Hadyn,

    Thanks for your patience. I've answered your questions below:

    1a) The ferrite beads can help mitigate emissions above the resonant frequency of the LC tank. If you have a frequency that is especially problematic, select the ferrite beads to minimize emissions at this frequency.

    1b) The capacitors to ground help shunt high frequency signals to ground, which can be generated when the driver switches between INxA and INxB. 

    1c) See 1a

    2) I do recommend adding additional filter capacitors that are sized appropriately for the sensor capacitor.

    2b) Actually, your nominal sensor capacitance would be just the capacitance of the C in the LC tank.

    3) In addition to the passive filtering options above, you can also reduce the Idrive settings. If you do so, ensure that the sensor oscillation amplitude is always between 1.2V-1.8V after the waveform has settled. If you do not need a very fast sample rate, you can also put the device in sleep mode, which prevents all of the channels from oscillating.

    4) See (3). Slowing down the sample rate by changing RCOUNT may actually increase your emissions, because the sensors will be driven for longer. If you slow down the sample rate by putting the device in sleep mode, this could help.

    5) It is much more accurate to measure the oscillation amplitude using an oscilloscope. The changing sensor capacitance can also change the oscillation amplitude, so it's best to monitor the amplitude across the full range of the sensor capacitance.

    Best Regards,