I have used EVAL-AD5933 EBZ evaluation board to measure noninvasive bio-impedance from the radial artery.
I have configured the frequency sweeping range from 3 kHz to 11 kHz with a sampling rate of 3 Hz. Is it possible to detect the radial artery pulse from the device??
Not enough specifics to understand what your setup is with the demo board. What kind of electrodes do you use to connect to the subject's skin and what would be the expected impedance range between your electrodes? Maybe it would make sense not to sweep, just output one frequency, say 10KHz or so, and set the number of increments to zero. Or setup a sweep with some non-zero number of increments, but zero frequency increment.
To record a heartbeat waveform, perhaps you want at least 10-20 readings per period, which is typically less than a second, so it would make sense to read every 50 milliseconds or so. You can achieve that by either setting up a sweep or, with some coding, reading repeatedly at a single frequency.
Thanks for the comment.
I have used BIOPAC-ECG electrodes placed on the wrist area on top of the pulse of the radial artery.
I have kept an inter-electrode distance of 7 cm.
I have used the calibration resistor as 1k ohms and I'm getting readings in between 1k Ohms to 10k ohms.
The time plot doesn't give a pattern and I couldn't figure out the radial pulse wave.
Still, not clear what your setup looks like. The only thing to suggest is checking the voltage waveform at the RFB pin of your AD5933 with the oscilloscope when running measurements to make sure that it is a clear sine wave without any distortion or clipping at the power supply rails. You might want to review similar projects like https://hackaday.io/page/3071-biomed-shield-for-arduino-101 .
Thank you for the feedback. I have measured the voltage waveform at RFB pin with the oscilloscope.
I have obtained a sine wave at 10 kHz as per my coding was altered to perform frequency sweep only at 10 kHz.
However, when I perform the frequency sweep with 1k Ohms resistor, I'm getting repetitive cycles of following wave pattern.
When I connect the electrodes to the radial artery skin, I'm getting repetitive cycles of following wave pattern.
Normally you want to see the waveform just like the one on your first picture. If your gain setting is 1x - you should have the correct reading of your complex impedance. If your gain setting is 5x - the second internal OPAMP in the AD5933 is clipping the signal at the rails (which you cannot see) and the results are not going to be correct. What is your gain setting and AD5933 VCC and VDD supply voltage?
Do you know for sure the nature of the transient between 0V straight line and the waveform on your second and third picture? It looks like it is the AD5933 coming out of standby mode in the beginning of the sweep, but impossible to tell not knowing more about your setup. What is your number of settling cycles? Depending on the duration of that transient (the scope probe is AC –coupled? or perhaps resulting from your AFE circuit? or both?) you might consider increasing number of settling cycles so that the early distorted cycles are not processed by the DFT.
What is the difference between the first and second pictures? Second picture was taken with 1k resistor, but what about the first one?
The third picture indicates that your skin/electrodes (or, less likely, your circuit) for some reason respond non-linearly to the applied excitation voltage. What voltage do you apply between the skin electrodes?
The good news as all the signals you shown ion the pictures seem to be within the dynamic range of the AD5933 internal ADC if your gain setting is 1x.
I have set the gain setting to 1x and VDD supply voltage is 3.3 V.
I don't apply a voltage on the skin separately apart from the device electrodes. I am using the bipolar configuration and I thought AD5933 EVAL board is operating with bipolar configuration.
Do I need to apply a separate voltage on the skin to measure Impedance?
I am pretty sure that all AD5933 eval boards are unpolar: everything is measured with respect to virtual ground at VDD/2 and that is the DC level applied to the skin. So it might be useful to ensure that the subject is well-isolated from the circuit ground. Is this the broad you are using: https://www.analog.com/media/en/technical-documentation/user-guides/UG-364.pdf?
I do not think you need to add any additional voltage to the skin. What is your excitation voltage? At 2V p-p one needs to be mindful of the electrode-skin interface behavior - it can be non-linear and hysteretic.