This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

AFE4300EVM-PDK: Calibration procedure

Part Number: AFE4300EVM-PDK
Other Parts Discussed in Thread: AFE4300

Good evening,

i'm working for the first time on the module in object and i'm encountering difficulties with the calibration procedure for the I/Q mode given in http://www.ti.com/lit/an/sbaa202/sbaa202.pdf. In details, i have performed the following steps:

1) I've short the jumpers J11, J12, J22, J23, J24, J25.

2) From the AFE4300 Device GUI, i have set the board as follow:

3) I've connected the IOUT0 and IOUT1 to the on-board impedance simulator as follows

4) I've set the switch SW3 in order to get 750 Ohm. This is my calibration resistor Rx.

5) From the ADC Capture & Analysis, i've acquired the following I/Q values: I = 0.508 V and Q = -0.110 V

6) I've set the switch SW3 in order to get 1 kOhm. This is my calibration resistor Ry.

7) From the ADC Capture & Analysis, i've acquired the following I/Q values: I = 0.676 V and Q = -0.143 V

8) By adopting the formulas for the magnitude and phase, i've computed the interpolating line passing through the magnitudes of the reference resistors. In particular, i've got a slope of 1460 and an offset of -9.1. The phase of Rx is -12°.

9) As sample network, i've set the on board impedance simulator in order to get a series RC circuit with R = 604 Ohm and C = 400 pF.

7) From the ADC Capture & Analysis, i've acquired the following I/Q values: I = 0.412 V and Q = -0.051 V

8) By inserting these values in the interpolated line equation i get a magnitude of 597 Ohm. The phase computed from the values given in 7) is -7°, and so the corrected phase should be -7°-(-12°) = 5° that is not correct.

In order to check if there was a bug in my procedure, i've used the functionalities of calibration and measurement in the BCM Measurement tab and it gives results quite similar to what i've found independently

Does someone see a bug in my procedure? Thanks

  • Hi Emanuele,

    Welcome to TI E2E forum!

    Can you send your questions to biosensing_afe-support@list.ti.com along with a reference to this E2E post?

  • Hi Emanuele,

      What is the excitation frequency that you are using? Looks like you are getting correct phase.

    Regards,

    Midhun

  • Hi,

    the frequency is 64 kHz. For this frequency, the impedance's phase should be arctan(-1/(2*pi*freq*R*C))=-84.45°, so i cannot understand the returned calibrated phase of 5°.

    Regards

    Emanuele

  • Hi Emanuele,

    Looks like your I and Q are swapped. Can you follow the guideline gives in 8.1.3 Achieving Deterministic Phase in the IQ Mode in the AFE4300 datasheet and check the results.

    Regards,

    Midhun Raveendran

  • Hi,

    thank you for your reply. Supposing that the I and Q channels were swapped, by using the data that i reported in my first post i would get arctan(0.508/(-0.110))=-78° for the calibration resistor Rx and arctan(0.412/(-0.051))=-83°, and so the calibrated phase would be -83°-(-78°)=-5°, but i expect -84° about for my RC circuit. 

    I saw the chapter that you indicated in the datasheet. About the first solution proposed (i.e., do a fresh calibration each time the registers are reprogrammed for a new excitation frequency), the measured data that i reported in the first post are given by the GUI of the evaluation board after that i have set the excitation frequency, so i think to have accomplished this suggestion. About the second solution proposed in the same chapter, i don't know how to set a synchronous clock for the SPI from the GUI of the evaluation board. Is there a way?

    Best Regards

    Emanuele

  • Hi Emanuele,

    I went through your original post where you mention you are measuring the impedance of the series R-C network. However the on-board impedance realizable is a parallel R-C. For the R,C and f values, there is a roughly 90 deg difference in phase between the series R-C and parallel R-C configurations. Can you please check if the impedance you are measuring is indeed a parallel R-C and not a series R-C and change your calculations appropriately?


    Also regarding setting a synchronous clock for the SPI, the EVM does not support this option. If you are designing your own board, you'd need to have a way to derive the SPI clock and the AFE input clock from a common master clock to be able to achieve this.

    Also it is suggested to do calibration as well as taking actual measurement within a single power cycle and device should not be reset in between.

    Regards,

    Midhun Raveendran

  • Hi Midhun,

    thanks for your reply. Effectively, the on-board impedance simulator lets me to simulate a parallel RC. I was confused by another test circuit that I realized on a breadboard. Therefore i get the correct phase by swapping the I and Q signal. I'm not designing a custom board, I'm only working on the EVM. Now that I am aware of this swap, I can compute the impedances in the right way. But why the EVM (or the related GUI) is returning swapped signals?

    Thanks

    Emanuele

  • Hi Emanuele,

    In the 1st post you have said the measurement results are giving 5 deg phase. If it is a parallel R-C circuit, the expected phase is -5.5 deg. So this is matching with the expected phase. There is no I,Q swap. 

    The phase sign is reversed in AFE4300. So measured phase is showing positive phase instead of negative phase.

    Regards,

    Midhun Raveendran

  • Hi Midhun,

    thanks for your replies. I messed up with my different test circuits and measurements. Moreover, I was not aware of the sign change performed by the AFE4300.

    Best Regards

    Emanuele