Applied Fft to my temperature sensor readings and don't understand the result

Hi everybody,

I'm reading data from my temperature sensor with arduino at a sampling frequency of 100Hz. Arduino writes data to the serial port and Matlab gets it perfectly (no data is missing). When I do the fft to see at which frequency I'm getting this result.

Second image is the zoomed graph of the 1st.

I'm new to digital signal processing, so I'm a bit confused: shouldn't DC gain be negligible? Is there an explanation for this result, or is it wrong.

Time series is plotted here also

Thanks in advance

Answer 1

[Christoph F.]

> Is there an explanation for this result, or is it wrong.

The result looks correct. The Fourier transform shows a large DC component, because the signal has a large DC component.

> I'm new to digital signal processing, so I'm a bit confused: shouldn't DC gain be negligible?

What do you mean by "DC gain"? From the graph, it looks like the average value (-> DC component) is about 32.75°C, and all other variations are in the +/- 2.5°C range. A large DC component in the frequency domain correctly represents the temperature data.

[Vasco Sampaio]

Thank you very much for your explanation Christoph. I understand it now, makes sense

Answer 2

[Rijal Al Kautsar]

can i know coding from arduino and matlab ?