Hello!

I want to do impedance measurement on some sample using a 3 electrode configuration, DACs and ADCs. I was thinking of using a dual phase lock in amplifier to get the resistive and reactive parts of the signal. Everything will be controlled by a microcontroller (uC).

I was thinking of using a DDS (e.g. AD9850) for sine wave excitation (as it is better since the DAC on the uC has max. 1MSPS).

Does this sound like the way to go? Any suggestions on better/alternative solutions?

What I need help with if the chosen solution above stands is how to do the analog circuit design from the DDS output to the uC/uCADC.

I am currently stuck and have been searching google for many many hours so any help that can point me in the right direction or if someone wants to try draw some basic circuit/block diagram, it would be much appreciated!

Thank you in advance =)

 

Impedance measurement of what?

 

Thanks for answering!

I want to measure impedance on biological tissue (bioimpedance) by using non-invasive electrodes on the skin. So I will be measuring some impedance between two points. (I do not wish to use a 2 or 4 electrode configuration).

Ref. electrode configuration:

http://www.gamry.com/assets/Application-Notes/2-3-4-Electrodes.pdf


Hope this was what you were asking about.

 

I have built several 3 electrode skin conductance systems. Once I used a sound card. Should know a few things about this topic

 

I have done EIT (Electrical Impedance Tomography) on breasts in order to detect cancer. This is a safe, painless, non-intrusive technique that is very inexpensive and can be performed in a doctor's clinic.

You have to identify your frequencies of interest. You will need a very high impedance AC source and input instrumentation amplifier with high CMRR. Then you need a waveform digitizer that can measure both amplitude and phase.

 

I have done EIT (Electrical Impedance Tomography) on breasts in order to detect cancer. This is a safe, painless, non-intrusive technique that is very inexpensive and can be performed in a doctor's clinic.

You have to identify your frequencies of interest. You will need a very high impedance AC source and input instrumentation amplifier with high CMRR. Then you need a waveform digitizer that can measure both amplitude and phase.

You can do this easy with constant AC voltage source, and a current to voltage converter (inverting opamp) The Rin in this figure will be the skin conductance it self

The PDF file is something I made a few years ago. It will give you some idea about the principal

 

MrChips:
Frequency of interest are 3kHz-300kHz (Will this wide range give any challenges that I'm not seeing yet?)(I only want to excite with maximum 10 different frequencies in this range, do not want to do a sweep).

I want to excite with a small signal, approx. 30mV.

If I understand you correctly I will need a high impedance AC source to ensure that most of the current will flow to the pick-up electrode.

CMRR will help canceling similar signals that enter on the + and - terminal on the op-amp, this I am familiar with from before making a EKG.

Should I sample the output AC signal with the AC signal after the biological sample (+ current-to-voltage converter) and compare the signals to find the phase? I could write a simple C program to do this in the micro controller.

Is there a waveform digitizer that specialize in this and give the amplitude/amplitudedifference or phasedifference or something similar?

t06afre:

Thanks for joining!

Currently I fail to understand the 3 electrode system completely, at least how to make the circuit (I think). I understand the I/V converter and that I need something to excite (with high impedance ref. above). The last electrode is to measure voltage relative to e.g., the source/sink? This way I will measure the voltage as in a 2 electrode system except I avoid the DC potential on the exciting electrode? But how to make the analog circuit I think I dont understand. I will try make one soon and upload it here and any comment (I guess there will be a lot) would really help me fully understand this concept.

Thanks again for the fast and helpful replies!

 

t06afre:
Didn't see the PDF before now.. Will read it asap =)

 

I have been looking at your provided PDF and it was of great help! (if possible I would be very interested in the rest of the book/compedium if that is something you want to give away).

I provided a PDF with a hand drawn "schematic". I have some questions regarding this:

1. Is there any errors in the circuit?
2. Any suggestions that would improve this circuit further?
3. Is what frequencies I can measure mainly determined by the op amps slew rate for this circuit? (for my range 3kHz-300kHz).
4. Will the circuit work well for R = 100Ohms to 10k Ohms?
5. If the tissue was to be purely resistive would ideally the phase difference between ADC CH1 and ADC CH2 be zero for this circuit?

These are the questions that comes to mind at the moment. (I will try figure these out by myself, but any help is appreciated!)

I saw in your PDF that you had a capacitor in the feedback of the current-to-voltage converter. I did not understand the connection between the phase error and the capacitor value.

 

I have some questions. First you will be using quite high frequencies. What will you achieve with this? Then you use a low frequency say 80Hz you will for the most measure the properties of stratum corneum. As the frequency goes up. The deeper layers of the skin will contribute more and and more to the result. Common medical skin electrodes are not made for such high frequencies. And will for sure also contribute to your measurements results. Also if using 300KHz you will need a quite high sample frequency say 3MHz at least.
The skin will never be purely resistive. And hence you will always have to measure admittance. In my instruments which has always been low frequency. I have always only measured the conductance part only. And used synchronous demodulator technique. Both digital and analog approaches. The digital method is very simple. You simultaneously sample the reference signal and the signal from the current to voltage converter. The you multiply these to signals (sample by sample) after using a digital low-pass filter. You will have a number that represent the skin conductance part.

 

I have not been 100% correct about my project description. I will also do in vivo measurements and many of the tissues I want to measure has characteristics that lies 50kHz-300kHz. So I will not only use skin electrodes to measure non-invasive. I plan on making the measuring electrode very small so that what I measure are dominated at what is close to the measuring electrode. I was planing on using an external ADC so the sampling rate would not be a problem.

I know the skin will never be purely resistive, I was just trying to say that if we were to connect a resistor between the C and R our phase would ideally be 0? This might be a way to calibrate for any phase offset?

Synchronous demodulator technique is the same as digital lock-in I think?

If I would like to get the susceptance I could just phase shift the ref. signal by 90 degrees (in code) and multiply, given that the frequency is know: Number of samples to represent 90 degree phase shift:
TOTAL_NUMBER_OF_SAMPLES_PER_PERIOD * (1/4)

 

The synchronous demodulator is the heart in every lock-in also the analog ones. To understand the lock-in method take a look at the zip I posted here (post #2 or #3 I think)
http://forum.allaboutcircuits.com/showthread.php?t=74293&highlight=lock-in
I am not so sure how usefull your mesurement will be. What will you be looking for? Your type of measurement has been done before. But it has not resulted in much. But anyway if you not have done it take a look here. They have many free publications.
http://www.mn.uio.no/fysikk/english/research/projects/bioimpedance/
Also be aware of that the current density is important for then measuring the skin properties. Yamamoto and Yamamoto (1981) found the upper limit of linearity to be about 10uA/cm^2 at 10 Hz and 100 uA/cm^2 at 100 Hz.

 

Thanks for the lock-in PDFs. I will read through them thoroughly!

I am using some of the impedance data for tissue published by uio researchers. I want to distinguish between different tissue types like muscle, fat, blod vessels etc..

I found the AD5933 IC. It looks perfect for my application except for the limit of 100kHz. I see that it is built a lot like what I was thinking with a DDS, ADC, digital lock-in...

 

As a hint. If you can build a synchronous demodulator followed by a proper low-pass filter. You can use a much lower sampling rate. As the conductance values will be quite static. And in this case perhaps you can settle with somewhat lower precession. I guess you do not need to changes down to 0.02uS over the full measurement range. The best thing in your case would be a commercial lock-in. Fitted for your frequency range. But well I know. Those cost as much as a car
The opamp I found working best in the constant voltage part. Was the LT1115 opamp. For the current to voltage inverter do some experimenting. Be aware of the gain bandwidth product for the opamp. Then using such high frequencies it is easy to find your self outside the limits with this configuration. This is not an easy task. So you must not be afraid to discuss with your professor or any else that can help you.

 

i also doing on the bioimpedance analysis on human blood resistivity to determine the glucose level...can anyone suggest me some circuit design concept?