Frequency respone trace

1

Hi

I have done some response of amps

I got the picture in attach

I can’t understand why from 1kHz to the high end the trace became thick

With both sw

Thanks

KIT3_DX_Risp_2,5watt
KIT3_DX_Risp_2,5watt1550×878 109 KB

2

Looks like some kind of instability.

It’s quite visible between 200Hz and 1kHz

3

The test was done on AudioNote Kit III, SEPP of 300B, and the doubt is about the secondary of OT where the 0 terminal is not to ground.

I will check, maybe this is the reason

Walter

4

More averages and higher fft size?

5

Try again without using the chirp option

6

Ok thanks

Walter Gentilucci
+39 335343255

7

First, what you’re seeing is typical chirp output.

I think you’ll find the latest version (1.4.1) of QA40xPlot does a better job with Chirps due to forcing rectangular fft windows. With that said, Chirps are a low-energy test so there are things you can do to improve the Chirp result.

  1. Use a longer FFT. This will improve low-frequency results and also improve smoothing results. I’d probably use 256K at your sample rate.

  2. Increase the voltage. Because Chirp is trying to extract information from a little energy increasing the energy will improve the accuracy some.

  3. Use smoothing. Even .01 octave will do great things at the higher frequencies without costing any real accuracy.

  4. Reduce the frequency range. Set the range to cover the area you are really concerned about but every octave of additional frequency range reduces the energy in other octaves.

Finally, for the utmost accuracy, @handsome is correct that turning off chirp and using a standard single-frequency sweep is rock-solid accurate (especially with averaging). It will produce many fewer frequency points so it isn’t as cool and smooth looking (and it is much much slower) but the values may be more accurate at the frequency limits.

Mark