Clarified WG valid amplitude clipping range in relation to c_diag_wg_gain_w.

libraries/base/diag/src/vhdl/diag_pkg.vhd

@@ -81,6 +81,9 @@ PACKAGE diag_pkg IS
   CONSTANT c_diag_wg_gain_w             : NATURAL := 1;  -- Normalized range [0 1>  maps to fixed point range [0:2**c_diag_wg_ampl_w>
                                                          -- . use gain 2**0             = 1 to have fulle scale without clipping
                                                          -- . use gain 2**g_calc_gain_w > 1 to cause clipping
+                                                         -- For c_diag_wg_gain_w = 1 there is clipping from [1 2> For normalized values >= 2**c_diag_wg_gain_w = 2
+                                                         -- the behaviour becomes more or less undefined. Due to wrapping it appears that normalized values [2 3>
+                                                         -- result in a sinus again. Therefore use normalized range [0 2**c_diag_wg_gain_w>.
   CONSTANT c_diag_wg_ampl_unit          : REAL := 2**REAL(c_diag_wg_ampl_w-c_diag_wg_gain_w)*c_diag_wg_ampl_norm;  -- ^= Full Scale range [-c_wg_full_scale +c_wg_full_scale] without clipping
   CONSTANT c_diag_wg_freq_unit          : REAL := 2**REAL(c_diag_wg_freq_w);                                       -- ^= c_clk_freq = Fs (sample frequency), assuming one sinus waveform in the buffer
   CONSTANT c_diag_wg_phase_unit         : REAL := 2**REAL(c_diag_wg_phase_w)/ 360.0;                               -- ^= 1 degree