sweep question

[luoGanda]

What would you like to know?

There is a test code - a sweep between 0 and 120 Mhz.
Code is mostly initialized and doing processing rx_callback as hackrf_sweep.c.
But from 0...20MHz i/q samples are ok, above that till 120 are almost 0 in amplitude, why?
(in other software that seems to work - not testes yet totally - eg hackrf_sweep)

This code does not print at all, while for <20MHz if prints as it should:

for (i = 0; i < num_fft_bins; i++) {
    if(frequency>=20000000 && (buf[i*2]>100 || buf[i*2+1]>100))
        printf("AmplOK\n");
}

In fact - most buf >20MHz is within -3 to +5.
Of course - it could be something at initialization step - but i checked more
times and everything looks ok. Eg using INTERLEAVED or LINEAR does not change a thing.

Also,
there is "buf += BYTES_PER_BLOCK - (data.num_fft_bins * 2);",
why not just use buf+=10? because of possible garbage bytes at start?

[luoGanda]

tried from scratch and seems samples amplitudes are somehow ok,
not sure why - but it could be setting bw at start.
It seems that also - at least for this - lna/vga must be quite high, eg cca >= amp/2(maybe it also a little depends on span).

Ok, this would be nice to know:

there is "buf += BYTES_PER_BLOCK - (data.num_fft_bins * 2);",
why not just use buf+=10? because of possible garbage bytes at start?