# AD5933 measuring unknown impedance and setting proper resistors

Question asked by greyesgt on Jul 1, 2016
Latest reply on Apr 13, 2018 by Mnzars

Hello, I am new to working with the AD5933 and impedance measurements. I am interested in measuring impedance between two 3M medical electrodes placed on the surface of the arm/forearm. I am interested in cross sectional impedance with opposing electrodes on the arm in one setup similar to impedance tomography. In another setup, I am interested in neighboring impedance with electrodes next to each other on the same side of the arm. I am not using a 4-electrode setup similar to bioimpedance analysis.

I have started out by using the AD5933 evaluation board and two electrodes. In particular, I’m interested in measuring (potentially small) relative impedance fluctuations when certain arm exercises/movements are performed. I’m not necessarily interested in what the absolute values of impedance are, but capturing the change in impedance. In particular, I am interested in maximizing the relative impedance changes I'm measuring at a given range. Most of the fluctuations I've seen are only 10-20 ohms, making it difficult to distinguish my movements.

(1)

The first part is about determining the unknown impedance range I should be focusing on. Related work places arm/hand skin impedance anywhere between 1kohm to 100kohm, or lower for dry skin and such. In other words, what is the appropriate impedance range (Zmin – Zmax) I should be focusing on for my experiment? Based on my understanding, the impedance range determines the resistor choice for Rfb and Rcal (or vice versa?).

My eval board configuration is: vdd=3.3V, internal clock=16.776MHz, freq sweep 1kHz to 100kHz every 500Hz, output range 0.4Vpp, settling=0. I followed the post here: https://ez.analog.com/message/43586 and set Rfb=Rcal=200kohm. The result is shown in Figure 1. According to the measured output, my unknown impedance is in the range of 18.25kohm to 18.75kohm for the sweep.

However, based on the AD5933 theoretical software provided by musach, the minimum impedance with this setup should be Zmin=23.2kohm. See Figure 2. Based on this, I am saturating the ADC and should not expect valid data. I tried lowering both resistors to 150kohm with the theoretical software suggesting a Zmin~=17.5kohm. Now my impedance readings are roughly 13kohm, again below the Zmin I am expecting from the theoretical software output.

* How should I adjust my Rfb and Rcal to find my Zunknown range? What values do I trust?

* Why are my measured values always lower than the expected Zmin?

(2)

Is it possible to focus only on a narrow range of impedance values to achieve a higher resolution in that range? For example, focus on the range from 8kohm to 10kohm only. I am using the evaluation board with the AFE. However, I have the option to not use the AFE with a separate hardware bench top setup.

Based on the AD5933 datasheet and using the following configuration with the AFE and no sweep (vdd=3.3V, vpk=0.4vpk, vdcoffset=vdd/2, freq=40kHz, gain=x1) for a desired Zmin=8kohm, I found Rfb~=30kohm. This is based on the equation for Rfb on page 5 of AN-1252 (http://www.analog.com/media/en/technical-documentation/application-notes/AN-1252.pdf).

From that same page 5 of AN-1252, to achieve Zmax=10kohm we set Rcal = (Zmin+Zmax)/3 = ~6kohm. However, I understand the data sheet suggests Rfb should be equal to Rcal to achieve unity gain voltage at the output of the receive side I-to-V amplifier in the AD5933. I can’t verify these numbers in the AD5933 theoretical software because it appears to assume Rfb=Rcal. See Figure 3.

The values I'm measuring with this setup on the evaluation board are roughly 2.8kohm (Figure 4), below the approximate Zmin=3.48kohm based on 40kHz from Figure 3.

* How is Zmin calculated in the theoretical software? The result does not seem to match the Rfb equation from AN-1252.

* Is it possible to narrow the impedance range (i.e., increase resolution) to capture small variations in a particular range?