FDC2214: Remote Sensor Penetration, Sensitivity and Shield

I recommend doing a few prototypes to test different options for the sensors. You should be able to create prototype PCB sensors cheaply and quickly since they are just copper pours in different shapes. You can even do some experiments with copper tape.

1.) and 2.)  Within a few cm would count as "close range", so trying a differential sensor would make sense. Have you seen the test data from our remote sensing app note?

http://www.ti.com/lit/an/snoa935a/snoa935a.pdf

3.) We do not have data on how closely the shield should track the excitation waveform. Your waveforms look close enough for a prototype. We often have customers that just use ground as the shield, so be prepared to test that option with your prototypes as well.

4.) To improve the performance of the shield driver, a rail-to-rail op amp would be helpful. You will need to experiment to see how effective your current design is. Possibly a higher slew rate op amp would be helpful as well. I suggest posting a similar question focusing on the EL8101 to the forums to directly ask the op amps team for a better recommendation.

A few other notes:

- I recommend using a reference sensor since you need high sensitivity.

- I also suggest using an external clock. This will reduce your noise floor.

- Make the RCOUNT setting as large as possible. This will also help reduce your noise floor.

Thanks Clancy,

That app note is what started all this :)

As advised, I've made a new post to discuss the op amp. https://e2e.ti.com/support/amplifiers/f/14/t/873118

I think the FDC2214 dev kit already has an external clock and by default is set to the maximum RCOUNT.

One of the problems is that the results don’t seem to be repeatable.

Moving from air, to freshwater, to saltwater, there is a distinct difference in capacitance within a single test. However, when I empty the reservoir and repeat the test, while the second test produces distinctly different results for the 3 mediums, they aren’t different enough from the first test. e.g. The result for saltwater in one test might be too close to the result for freshwater from the other test, regardless of how distinguishable they are when excluding other tests. The salinity in both the seawater and saltwater is controlled.

We can’t really implement a reference sensor. As an example, if the product is submerged in freshwater, it won’t have any contact with air or saltwater.

The oscilloscope is showing a Vpk of 2.2V whereas the datasheet says it should be between 1.2V and 1.8V. Have I misinterpreted?

How do I change the excitation frequency and voltage in the TI Sensing Solutions EVM GUI? (for FDC2214EVM)

Kind regards,

You can change the excitation voltage by adjusting the sensor drive current This is the IDRIVE setting in the GUI. See Figure 26 in the user's guide. Reduce the code value to reduce the current and the voltage.

Changing the excitation frequency requires changing the hardware. You can adjust either the sensor itself, or the parallel sensor capacitor or inductor. 

One issue you will run into is that there will be temperature drift. This is one of the reasons we suggest using a reference sensor. You could try experimenting by using the same discrete parallel capacitor and inductor with no extra capacitance sensor connected as the reference on a different channel. These components would need to be high quality with low temperature drift.

Thanks Clancy,

Frequency Target

Having to swap out components makes it a lot harder to test different frequencies. The datasheet shows how to calculate the frequency from the value of the parallel capacitor and inductor. (1) However, is there a way to calculate what frequency we need in order to select the right value for the capacitor and inductor? The sensing plates we are testing are significantly larger than the plates on the FDC2214EVM.

Our most promising design is a parallel plate configuration where the test medium becomes the dielectric between the plates (sense plate and ground plate).

These are the configurations we would like to do more testing on, with various plate shapes and positions, with and without a shield. (2) Do they all seem viable?

Single-ended sensor with a ground plate, undershield and a guard ring.

Differential sensor with two plates, undershield and a guard ring.

Single-ended sensor with an undershield and a guard ring.

Temperature Drift

We were hoping that we could use the products internal temperature sensors to compensate for temperature drift. (3) Do you think that the drift is consistent with temperature change?

Kind regards,

1.) I believe I answered this on a different thread that you started. Unfortunately, we do not have much guidance on this. I suggest choosing a frequency near the lower end of the operating range and on the higher end of the operating range and testing to see if there is much of a difference in performance. This will save you from doing too many iterations.

2.) Yes, these all seem viable. The parallel plate sensors are what we usually see for liquid sensing, so I suggest starting with those versions.

3.) There will be a correlation between temperature and drift, so this should help you. Using a reference sensor is the best solution, but a temperature sensor is second best. 

Thanks Clancy,

1)

Yes you did. Thank you. I did a frequency sweep between 10kHz and 10MHz and found that the difference in impedance between saltwater and freshwater increased as frequency increased (for the sensor design tested). What is the maximum frequency we can run the FDC2214? The datasheet says 10MHz but other E2E threads say 6MHz. At 6MHz there was still an uncomfortable amount of overlap in the data.

2) and 3)

Thank you.

Thanks Clancy :)