Post Go back to editing

Problem in measuring impedance at lower frequency from 10 Hz to 1k Hz using AD5933

Hello everyone, I am developing an impedance measurement system using AD5933 with a microcontroller. I can measure impedance from 47 ohms to several mega ohms from frequency range 5k to 100k Hz. Now I want to measure impedance from frequency range 10 Hz to 4k Hz. I have uploaded my AFE picture. I know there are many threads related to lower frequency excitation, I have gone through them but I did not get an appropriate idea.

I have uploaded a picture below. According to my understanding after going through this table, we can measure impedance at a low-frequencies by simply giving an external clock to AD5933 pin. I have tried but it's just working for 16Mhz and 4Mhz external oscillator clock. For lower frequencies from 10Hz to 1K Hz how can we measure impedance, do we need any calculation for settling cycle register?


  • Quite an elaborate AFE you got there... The suspect would be the circuit R8R9R10C3 that acts as a low-pass filter with a time constant (R8||R9||R10)*C3 = 1.41630901288kOhm*1μF = 1.4 ms, so operating frequencies lower than a few KHz start noticeably leaking through. The suggestion would be to increase the resistors R8, R9, R10 (and, of course, R13 to keep R13/R10 ratio the same) or C3 or all the above. And prepare for long sweep duration.

    If your current number of settling cycles works for your measurements at high MCLK, there should be not need to change it at lower MCLK.

  • Thank you for your reply, I got this circuit from chabowski paper [1],  In this paper page num. 16 it is discussed that (ICA and ICB are for DC bias canceller. the voltage detectors are averaged on R8 and R9 and its correspond to DC bias voltage. 

    I think it's not an AFE problem because now I am trying to measure lower frequencies without AFE. I have measured 20kohm resistor with RFB 10kohm, it gives correct measurement from frequency sweep (5k-100k) by using 16Mhz external clock (settling cycle 0x64) and from 1khz to 5khz by using the 4Mhz external clock (settling cycle=0x0F), but when I am connecting 2Mhz external clock it acts as I have shown in video (running infinitely). but if I change the settling cycle from 0x0x0F to 0x05 it gives the result that I have shown in the image (not correct), I have measured 20kohm with 10kohm RFB. 



         Konrad Chabowski, Tomasz Piasecki, Andrzej Dzierka, Karol Nitsch

         Metrology & Measurement System, Vol. XXII (2015), No. 1, pp. 13–24.

  • It would be helpful if you posted the circuit without the AFE, Otherwise it is difficult to comment.

    You may be bumping into the artifacts of the "DFT" type the chip is performing. Those artifacts tend to show up at the lower end of the frequency sweep regardless of MCLK frequency. Basically, at Frequency Codes below about 250000 (decimal) the “DFT” errors start growing (in oscillating manner) quite substantially, so generally there is quite sizeable systematic error in the measurement unless you measure at a handful of specific Frequency Codes where this oscillating error is zero (the so-called coherent sampling, where the sampled sequence encompasses the integer number of signal periods). If you want more details, tfsoft posted reference to a paper where it is discussed.

    Other people in this forum have bumped into this effect and worked around it by reducing MCLK. To see how far they reduced their MCLK to operate at low frequencies of interest, please look, for example, here: Three electrode measurement for AD5933? , posts by rvalencia .

  • Running sweep without end could indicate some I2C issue. Your movie also shows unchanging value in the right column, which is suspicious - the status register and Re and Im registers do not contain correct values or red incorrectly. People reported I2C errors at low MCLK, but below 100KHz - much lower than yours. As to number of selling cycles - it should not be zero, but that would be the only restriction.
    Again, commenting without seeing the circuit is always a crapshoot.

  • I have uploaded my circuit and external oscillator output. It's working for  16Mhz or 4Mhz, but for lower frequencies, it's not working. Now I am giving 2Mhz frequency, I want to take measurement from 300 Hz to 5000 Hz.

Reply Children
  • This circuit should work fine if Vexc is set at maximum and gain is set at 1x.
    Do you have control over I2C speed of your microcontroller? Something similar to, say, Arduino's Wire.setClock() function that could reduce the I2C clock below 100KHz?

  • Thanks for your reply, I am using Teensy3.2 and Arduino IDE for programming. yes, I can control the i2c speed, I have mentioned the function below. now I am using 400Khz. but minimum speed will be 100kHz,

    Vexc = 1.98Vp-p;

    gain= X1;

    Wire.begin(I2C_MASTER, 0x00, I2C_PINS_18_19, I2C_PULLUP_EXT, 400000);

  • I have tried with Arduino UNO with 50kHz i2c frequency, but I am facing the same problem.

  • AFAIK with Wire library you can go as low as 31KHz if you call Wire.setClock(31000L) after Wire.begin(). Maybe that would resolve your operation issues at low MCLK.

  • Snorlax, I think I am doing some mistake in my code, I saw the post as you mentioned above, can you please tell what is L in the picture below that must be an integer as gennyelectro said, I am using the function for frequency code I mentioned below  

    byte getFrequency(float freq, int n){
    	long val = long((freq/(MCLK/4)) * pow(2,27));
    	byte code;
    	  switch (n) {
    	    case 1:
    	      code = (val & 0xFF0000) >> 0x10; 
    	    case 2:
    	      code = (val & 0x00FF00) >> 0x08;
    	    case 3:
    	      code = (val & 0x0000FF);
    	      code = 0;
    	return code;  

  • Your function should work just fine as long as you change your global MCLK value in accordance with the clock frequency you feed the AD5933 chip.
    As to the "L" gennyelectro discussed - it is the integer number of full cycles encompassed by the 1024-long sampled sequence to eliminate the effects of the DFT artifacts. Not sure if it is directly related to your problem.

  • sorry for the late reply. I have solved my problem, Actually i was not waiting between reads.

  • Great, best of luck with the rest of your endeavor!

  • Hi   and  , may I know if you have added any additional circuitry between Out1 (DS1077L-66) and MCLK (AD5933). I have tried to generate 2MHz using DS1077L-40. I notice that the overshoots on the rising and falling edges of the clock waveform are very high. I have checked the generated waveform using an oscilloscope. The max and min voltage of the clock waveform are 6.72V and -1.68V respectively as shown in the picture below, which is way beyond the absolute maximum rating indicated in the datasheet (-0.3V to Vdd+0.3V).

    I have added a 500pF shunt capacitor to filter out the spikes to ensure the waveform is within the absolute maximum rating as shown in the attached picture below. However, it seems like my AD5933 IC cannot detect the clock signal as my calibration keeps failing (calibration works on evaluation board with signal generator using the same source code, so it is definitely not the coding problem).

    Would really appreciate if you could share me any of your experience please. Thanks in advance.

  • Not sure I can be of help without seeing the setup. The AD5933 in my experience worked fine with just about any source of MCLK from general-purpose lab generators to chip oscillators of various kinds including programmable ones, just not the the DS1077L-40 specifically. Whenever chip oscillators were used they shared power supply with the AD5933, so the levels were automatically within working range. Also the wires/traces between the oscillator output and the MCLK input were typically under 10cm in length, so excessive over/undershoots were never observed.
    How do you conclude that the calibration is failing? Does the problem occur when the output signal from the DS1077L-40 is fed to the evaluation board? If so, would it be possible to supply power to the DS1077L-40 from the evaluation board so that it would be the same as the AD5933 receives? 
    500 pF is probably too much, the DS1077L-40 datasheet seems to suggest that 50pF is the max load. To reduce the spikes while preserving reasonably sharp transitions people tend to use 10-30Ω resistors between the oscillator output and MCLK input.