FFT size on Low Frequencies

I’ve been trying to measure a transformer-balanced unit. The output transformer (in combination with a DC blocking cap) creates a peak at the very low frequencies.

However, it is rather difficult to measure the exact frequency and amplitude since it varies depending on the FFT size.

Please see this plots below. Which FFT size should I trust? FWIW, the applied signal is +4dbu.

I thought that the higher the FFt size, the smoother the curve would be (this is the case also when using REW) but with the QA403 it doesn’t seem to be the case.

In fact, I’m not even sure if this LF peak is real or some digital measurement artifact.

Nobody has faced a similar situation?

Perhaps Matt can give some advice?

The measurement resolution is given by: SR / FFT size.
In your case, you want to measure a 2.5 Hz resonance, thus the higher FFT size should be used. Notice that as you increase the FFT size, the peaks gets closer to each other, with 512k and 1024k quite ovelapping: you reached the right resolution for that measurement.

1 Like

Remember your QA403 is AC coupled (~1Hz) on the inputs.