AFE4400 signal denoising

Arnab Chakraborty

Other Parts Discussed in Thread: AFE4400

Hi support group

I am capturing AFE4400 data from MSP430 and fluke simulator. There I have set the signal to be drifted due to respiratory artifacts. I have designed a second order butterworth filter in a band of (0.5-3) Hz to negate the baseline drift, but after filtering the signal shape is totally changing and I am not able to get the heart rate. I used the command [b,a]=butter(2,Wn) in MATLAB to design the filter and used the coefficients in b and a to filter the data. Am I doing anything wrong? Please let me know.

Thanks

Arnab

over 9 years ago

0 Amy Schnoor over 9 years ago

TI__Mastermind 18285 points

Arnab,

I am assigning your post to the correct applications engineer. He should respond soon.

Question: Does everything look fine without the filtering?

0 Arnab Chakraborty over 9 years ago in reply to Amy Schnoor

Intellectual 635 points

Hello Amy

Yes everything look fine without filtering. I mean the signal coming out of AFE. I will post the graphs of the original signal as well as the filtered one tomorrow.

Thanks
Arnab

0 Arnab Chakraborty over 9 years ago in reply to Amy Schnoor

Intellectual 635 points

Hello Amy

Did you assign the post to anyone?I am waiting for your reply..

Thanks

Arnab

0 Praveen Aroul over 9 years ago in reply to Arnab Chakraborty

TI__Mastermind 49916 points

Arnab,

The Matlab command is valid. The cutoffs are fine as well. What sampling Frequency did you use?
Have you tried plotting the frequency response using the command freqz and confirm that it has a band pass response?

Our suggestions would be:
• Plot the filter response with freqz command in Matlab
• Collect data for when the user is rest and see if you could count the peaks on both the filtered and unfiltered responses.

0 Arnab Chakraborty over 9 years ago in reply to Praveen Aroul

Intellectual 635 points

Hello Praveen

In rest position there is no problem. Do you mean to say without adding any noise? There are two types of noise that I can add. One is 50Hz line frequency noise and the other one is respiratory noise which will introduce a low pass noise component. I am having problem in removing that component only (peaks are getting distorted).I will send you the plots in subsequent posts. The sampling frequency I have used is 500Hz (equal to the PRF that I have set). The freqz plot is also proper.

Thanks

Arnab

Data converters

Data converters forum

AFE4400 signal denoising