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.

AFE4403-HREVM - Weak signals and application errors

Hello,


I am having trouble with the AFE4403-HREVM ("wristwatch") evaluation kit. It has not been modified in any way nor reflashed. The sensor board used is the one that came preinstalled (OSRAM SFH7050). The PC used is an old machine running Windows XP SP3.

Measurements made with TX3 mode disabled are basically flat lines:

Zooming in only show noise:

With TX3 mode enabled, LED 1 signal is still flat. LED 2/LED 3 shows a very weak and noisy signal when zooming:

Running the calibration, when it works, gives a signal that looks more normal but still is pretty weak:

However, most of the time, the calibration fails after a few seconds with these error messages:


Things I've tried:

- enabling/disabling the notch filter: no improvement

- setting the wristwatch power switch to on: the calibration seems to crash less often if the power switch is on

Is there anything wrong with the hardware or the way I'm using it?

  • Boris,

    We have received your post and will be replying shortly.

    Thanks,

    Chuck Smyth

  • Boris,

    For the "reading stopped without...." error refer to this e2e post: e2e.ti.com/.../349138

    For the signal quality issues:

    Make sure the watch is very snug on the wrist to reduce ambient light.

    You can try manually adjusting the gain settings of the AFE.

    Another Note: The settings and firmware has worked with multiple subjects, but not all. There may need to be some fine tuning to work on every subject.
  • Thank you for your help.

    I've installed the GUI application on a much newer and faster PC running Windows 7 today. The application appears to work much more reliably and smoothly on this machine. Maybe the old PC was not fast enough?

    The results I get now, using calibration, are comparable to those one of my colleagues gets without using calibration. I've made sure the watch is attached snugly to my wrist, but since you've mentioned it doesn't work as well with every person, I assume this is not anormal.

    In any case, even if they're not strong, the signals look usable. In addition to the notch filter, may I suggest adding a high-pass filter in a future version of the application? There seems to be a low-frequency component that tends to be obscure the useful signals.