Phase not unwrapped before group delay is calculated

[roseengineering]

NanoVNA-D/plot.c

Line 331 in 00e21dc

return vna_atan2f(i, r) / (2 * VNA_PI * deltaf);

To properly measure group delay, ie. -dph / (2 * pi * df), the phase must be "unwrapped" before this calculation is performed, otherwise for large phase shifts (but <180 deg) between points, the group delay calculation might be in error. For a non-dispersive DUT like a coaxial cable, highly accurate group delay measurements require a large frequency "aperture". For example to measure coax length, a df of .3 / delay is recommended. Unfortunately because of this non-unwrapping issue measurements could be wrong. See my measurement at https://twitter.com/gmagiros/status/1396204275089477636?s=20

The python code below show how to unwrap phase and then calculate the group delay between two points.

for n in range(1, len(ph)):  # where ph is the phases in radians
    dph = ph[n] - ph[n-1]
    if dph <= -np.pi: dph += 2 * np.pi
    if dph >= np.pi: dph -= 2 * np.pi
    df = f[n] - f[n-1]
    g[n-1] = -dph / (2 * np.pi * df)