Question about FDC1004 +/-15pF range and OFFSET_CALn register

Thank you for very quick response. We have now done some testing and encountered a new problem.

Our FDC1004 offset calibration sequence flow is following:
1. CAPDAC is set to initial value and the capacitance is measured (fine tuning OFFSET value is set to zero)
2. Change CAPDAC value and measure the capacitance until we find CAPDAC value which produces measurement result as close as -16pF as possible (e.g. -15pF)
3. This CAPDAC value is then used in our application (now we have full measurement range of approximately 30pF available from FDC1004)

The problem is that when we get close to -16pF in the calibration sequence the change in the measurement value jumps much more than one CAPDAC value should (3,125pF according to the datasheet). As you can see from the data below the capacitance value changes around 3.1pF when we start to iterate the CAPDAC value. When the measured capacitance is near -7pF and we change the CAPDAC value the delta value between these two CAPDAC settings jumps from 3pF to around 9pF. This issue creates problems in some units because the jump is so big that the measured capacitance goes from -8pF to -16pF (overflow) and the usable measurement range is much smaller in those devices.

Can you explain this behaviour?

Calibration sequence log

measured capacitance, CAPDAC value = n
769820 (HEX value, single measurement)
14.824 (pF value, average ~50 measurements)

measured capacitance, CAPDAC value = n + 1
5DB7D5 (HEX value, single measurement)
11.715 (pF value, average ~50 measurements)

measured capacitance, CAPDAC value = n + 2
44EFBD (HEX value, single measurement)
8.617 (pF value, average ~50 measurements)

measured capacitance, CAPDAC value = n + 3
2CA067 (HEX value, single measurement)
5.578 (pF value, average ~50 measurements)

measured capacitance, CAPDAC value = n + 4
138C50 (HEX value, single measurement)
2.444 (pF value, average ~50 measurements)

measured capacitance, CAPDAC value = n + 5
FB9981 (HEX value, single measurement)
-0.550 (pF value, average ~50 measurements)

measured capacitance, CAPDAC value = n + 6
E15315 (HEX value, single measurement)
-3.834 (pF value, average ~50 measurements)

measured capacitance, CAPDAC value = n + 7
C905DF (HEX value, single measurement)
-6.872 (pF value, average ~50 measurements)

measured capacitance, CAPDAC value = n + 8
81C63E (HEX value, single measurement)
-15.778 (pF value, average ~50 measurements)

Are you using our EVM to evaluate the case above? Can you elaborate on the setup that you are using?

In terms of the capacitance that you are measuring, is it a known value or a static value as you change the CAPDAC value?

We made our own PWB so we are not using the EVM board.

Our setup is the same as described in my previous posts.

We have 2 metal rods (lenght ~120mm) which are ~25mm apart from each other. One rod is connected to GND and the other is connected to FCD1004 sense input.

First we calibrate the offset capacitance when there is air between the two rods. After the offset capacitance is calibrated we will add some substance between the the rods and basically measure the change in the dielectric constant.

The issue described in my previous post is happening in the calibration phase (air between the metal rods) where we try to calibrate our system so that the capacitance measurement result is as low as possible (the smallest value FDC1004 reports back is -16pF)

As the dielectric constant of the substance we are measuring is varying over several measurements we would like to have full measurement range from the FDC1004 available. As the maximum reported capacitance from FDC1004 is +16pF so we would have 32pF measurement range if we succeed to calibrate the FDC1004 to -16pF.

Other question in my mind is regarding the OFFSET_CAL register value because there is no clear explanation in the datasheet how OFFSET_CAL tuning affects to the measurement range of FDC1004. Does the OFFSET_CAL value only digitally subtract/add the programmed value from the result or does it actually change the measurement range?

Lets assume following case: If the FDC1004 reports -13pF after coarse tuning (with CAPDAC) and then we use OFFSET_CAL register to fine tune the reported capacitance value from FDC1004 by -2.5pF to -15.5pF (= -13pF - 2.5pF). If we then add load capacitance of 31pF does the FDC1004 measure the load capacitance correctly and report +15.5pF (= -15.5pF + 31pF) or does some overflow occur and the result would be 13.5pF  (-13pF + 31pF = 18pF which is over the FDC1004 range thus it would report +16pF. Then the OFFSET_CAL value of -2.5pF would be subtracted from the result -> 16pF - 2.5pF = 13.5pF )?

There is something off with your setup just from intuition. What is the change in capacitance from air to filling the container with the material? What kind of substance are you detecting? What application is this for, liquid level sensing?

I understand that you are offsetting the initial values so that you can use the full input dynamic range of the part but from what I understand, it seems like you are relying on absolute values...The CAPDAC and offset registers correct the data digitally. Your last example is correct. If you add a 31pF load capacitance, it should measure it correctly as long as there is enough margin in the input capacitance range.

If you can explain what the application, I can give you a better use on how to look at the data.

Thanks

-David Wang

Capacitive Sensing Applications