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.

Calibration procedure for AFE4400?

Other Parts Discussed in Thread: AFE4400

Is there any particular calibration sequence that we need to use for the AFE4400?  We are still learning how to use this device, so I do realize that this might be a naive question.

  • Hello Dean,

    The TIA gain (Rf), TIA bandwidth (Cf), LED current (ILED), second stage gain etc has to be programmed such that the signal quality at the ADC output is reasonably good for extracting the heart rate value out of it. Ideally it is good to have an automatic gain and offset calibration algorithm.

    The optimum gain and LED current depends on the following conditions

    1) The current transfer ratio from the LED to the photodiode (CTR)
    2) The perfusion index at the ADC output (This is the AC to DC ratio of the signal)

    I have described a brief procedure to do the gain and LED calibration manually.

    1) At first, run the system with lowest gain (Rf = 10K) and low LED current (For example, ILED = 2mA). Disable the stage 2. Measure the DC and AC in the ADC output.
    2) If the AC value is not reasonable good, Increase the LED current or Rf setting such that DC becomes half of the ADC full scale. This will increase the AC signal also
    3) If the AC value is still not sufficient, then enable stage and cancel the extra DC current (but ensure that the data in the ambient phase is not saturated)
    4) The Cf value can be set such that the TIA time constant is much lower than the LED pulse width
  • Thank you for your response. I have a quick and naive question here. Should this calibration be performed without the finger inside the sensor clip or without the finger inserted?
  • With the finger inserted.
  • Thanks again for the information. I've adjusted the LED current and have been seeing more sensible SpO2 readings. I haven't tried enabling stage 2 yet, but I will.

    Now, when you say that the DC readings have to reach half of the full scale value, do you mean that we should be getting readings in the range of 0x0FFFFF? I just want to make sure that we're not misinterpreting anything.

    Also, before we start tinkering with stage 2, what do you mean when you say that the AC readings have to be reasonably good? What criteria should we look for? I ask because even with our increased DC and AC reading values, we're still getting SpO2 values that are mostly under 95%.
  • It's been a while, so I wanted to revisit these last few questions... again, just to make sure that we understand things correctly.  Do we have any criteria for what's "reasonably good"?  Also, should we expect the DC readings to be in the range of 0xFFFFF?

  • I've been able to greatly improve the results by adjusting the gain and LED current, and by enabling stage 2, as suggested.  

    Do you have any recommendations for how to cancel the extra DC current, as stated in the calibration procedure that you provided?  I understand that this can be done using the TIA_AMB_GAIN register (0x21), Do you have any recommendations for how the precise amount of cancellation current should be determined?

  • Praveen, the procedure that you provided is much appreciated.  I've been reviewing, trying different settings, and I think we have a better understanding of how this works.

    I have a question, though.  In step 3, you suggested turning on the second stage amplifier and using it to "cancel the extra DC current."  Are you referring to the ambient light current?  If so, it seems to me that the proper approach is to get the ambient light readings from registers 0x2B and 0x2D, then adjust the AMBDAC bits in register 0x21 such that these readings are close to zero.

    Is that correct?  If so, then how should we interpret the statement "ensure that the data in the ambient phase is not saturated"?  Would this pretty much guarantee that the ambient readings would not be saturated, since they'd be close to zero?

  • Hello Dean,

    Yes you are correct. I was referring to the ambient light current to cancel the extra DC current.
    The procedure to read the registers is also correct.

    What I meant by "ensure that the data in the ambient phase is not saturated", is that when you choose the TIA gain after reducing the effect of the ambient signal at the ADC input, ensure that the TIA gain will not saturate at the TIA output for the assumed variation around the measured ambient signal. Choose the second stage gain based on what it takes to get back to close to 50% full scale after subtracting the ambient signal.