Strange FFT behaviour in WaveForms

[Andras]

I use WaveForms 3.17.18 on Windows 11 with Analog Discovery 2.

The wave generator output (W2) is directly connected to scope input CH1.

W2 is set to 123.4567 Hz Sine wave with 500 mV amplitude.

I set the Scope to 1kHz and its rage to 200mV/div. I use the built-in FFT in the range of 120 - 130 Hz with the Peak(V) setting.

It normally looks good but periodically I get a strange reading like this:

I'm not sure what is happening here, could you help me find out?

I also made a screencap of this and uploaded here:
https://drive.google.com/file/d/1Bxn0S0H6SM2J9r3cXOgJ3lleH5cPN6v7/view?usp=sharing

 

[attila]

Szia @Andras

Yes, it is expected.

The FFT is performed on the whole captured samples. Which in case for scan shift/screen and record equals to the visible samples. For repeated capture the visible may be less since it is rounded off (like 8000 for 8192) trying to get nicer sample rate for the given time base and at low time base only a portion is shown (like for 8192 buffer and 100MHz max the minimum capture is 81.92us)

With scan screen the hump is significant during middle crossing because before FFT the samples are windowed (default FlatTop) having a peak in the middle.

The FlatTop looks like M1 and the scan is stopped close to the middle:

[attila]

Szia @Andras

This is due to discontinuity in the scan screen data. Use shift or repeated mode.

[Andras]

Alright, thank you!

I'm just curious: is this an expected behavior?

Do I understand it correctly that the screen mode would mean that the source of the FFT are the samples that are currently visible on the screen? If so, and you look at the video I attached, why is this hump in FFT only there at a certain point? The array of samples look continuous on the screen after the first round-trip.

[Andras]

Szia @attila,

Thank you for your explanation, I definitely learned something again! :) I didn't know about the window function, but it makes perfect sense now!