Improved description of sdp_bf_weights. Defined unused field for second word of sdp_bf_scale.

applications/lofar2/libraries/sdp/sdp.peripheral.yaml

@@ -72,10 +72,17 @@ peripherals:
           "The beamlet weigths are stored in g_nof_instances = N_pol_bf * P_pfb = 2 * 6 = 12, where
            P_pfb = S_pn / Q_fft = 6. Per instance there is a block of Q_fft * S_sub_bf =
            2 * 488 = 976 complex BF weights. The N_pol_bf = 2 represents the two beamformer
-           polarizations, to distinguish these from the N_pol = 2 antenna polarizations. The pairs
-           of Q_fft signal inputs that are multiplexed per block are mapped to the N_pol = 2
-           polarizations of an antenna.
-           Therefore A_pn = P_pfb = 6 is the number dual polarization antennas per peripheral.
+           polarizations, to distinguish these from the N_pol = 2 antenna polarizations. The
+           beamlet weigths for S_pn = P_pfb * Q_fft = 12 signal inputs are therefore defined by:
+
+           (cint16)bf_weights[N_pol_bf][P_pfb]_[Q_fft][S_sub_bf]
+
+           where N_pol_bf = 2, P_pfb = 6 and Q_fft = 2 and S_sub_bf = 488, defined in sdp_pkg.vhd.
+
+           The pairs of Q_fft signal inputs that are multiplexed per block are mapped to the N_pol = 2
+           polarizations of an antenna. Therefore A_pn = P_pfb = 6 is the number dual polarization
+           antennas per peripheral. The  beamlet weigths for S_pn = A_pn * N_pol = 12 signal inputs
+           are therefore defined by:
 
            (cint16)bf_weights[N_pol_bf][A_pn]_[N_pol][S_sub_bf]
 
@@ -113,11 +120,11 @@ peripherals:
            requantizes the result to beamlet data output with less bits.
            The beamlet scale factor has g_gain_w bits and the value 2**g_lsb_w represents a gain of 1.
            For example for g_gain_w = 16, g_lsb_w = 15, a beamlet sum of 18 bits and beamlet data
-           output of 8 bits a scale value of:
+           output of 8 bits, a scale value of:
            . 2**g_lsb_w = 2**15 selects the lowest 8 bits of the beamlet sum and clips the highest
-             10 bits
+             10 bits,
            . 2**11 rounds the lowest 4 bits, selects the next 8 bits of the beamlet sum and clips
-             the highest 6 bits.
+             the highest 6 bits,
            . 2**5 rounds the lowest 10 bits and selects the highest 8 bits of the beamlet sum."
         slave_type: REG
         fields:
@@ -125,8 +132,10 @@ peripherals:
               field_description: ""
               width: g_gain_w
               address_offset: 0x0
-              bit_offset: g_lsb_w
               number_of_fields: 1
               radix: unsigned
               #radix_width: g_gain_w
-              radix_resolution: -g_lsb_w
+              radix_resolution: 0 - g_lsb_w
+          - - field_name: unused
+              field_description: "Not used."
+              address_offset: 0x4