Using AFE4490EVM with the supplied sensor on my own finger

Devidas,

The screenshot shows you have large amount of power line noise(50Hz/60Hz) riding on the signal of interest.  Try to switch from time domain to FFT plot.  You should see large spike at your power line frequency.  The signals you see in text boards are filtered and/or ideal pleth signals.   

You can try the following to improve things. 

1.  Run your EVM on a laptop powered by a batteries only.  Unplug the and remove the AC adapter.  Place it at least an arm's length away.  Now capture data.

2.  Enable the 50/60Hz notch filter on the GUI to help attenuate the power line noise.  Now capture data. 

3.  Finally, save the raw waveforms to a file and post process the signals in Matlab(for example) using a sharp LPF with cutoff at say 5Hz.  The baseline drift could be removed with a 0.5 Hz High pass filter.  Once you filter the signal, you'll see it looks more like a textbook signal.

TI provides reference example for PulseOx.  See link below. 

http://www.ti.com/tool/tida-00010
In the test results document, you'll see the textbook pleth waveform from actual hardware.  It uses the AFE4490.

 

Thank you for the reply.

I too suspected the mains interference. Mostly it may be  picked up in the sensor or its cable.

Can changing the finger sensor module improve data ?

I am not doing a standard pulse-oximeter but trying for a high resolution system.

I have done a discreet system and now wish to switch to AFE4490 to save cost and space.

Regards,

Devidas

Davidas,

Yes, try and switch the sensor/cable. The shorter the better. 

 

i am using pulse oximter with finger clip reference design.

i need to know which finger clip sensor is compatible? Any part number available ? i am need of specification details

Duplicate post.

Please see response in the post below:

https://e2e.ti.com/support/applications/high_reliability/f/30/p/328919/1515468#1515468