mirror of
https://github.com/fairwaves/UHD-Fairwaves.git
synced 2025-10-30 19:43:32 +00:00
113 lines
3.3 KiB
Verilog
113 lines
3.3 KiB
Verilog
//
|
|
// Copyright 2011 Ettus Research LLC
|
|
//
|
|
// This program is free software: you can redistribute it and/or modify
|
|
// it under the terms of the GNU General Public License as published by
|
|
// the Free Software Foundation, either version 3 of the License, or
|
|
// (at your option) any later version.
|
|
//
|
|
// This program is distributed in the hope that it will be useful,
|
|
// but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
// GNU General Public License for more details.
|
|
//
|
|
// You should have received a copy of the GNU General Public License
|
|
// along with this program. If not, see <http://www.gnu.org/licenses/>.
|
|
//
|
|
|
|
// Medium halfband decimator (intended to be followed by another stage)
|
|
// Implements impulse responses of the form [A 0 B 0 C 0 D 0.5 D 0 C 0 B 0 A]
|
|
//
|
|
// These taps designed by halfgen_test:
|
|
// 2 * 131072 * halfgen_test(.8/8,4,1)
|
|
// -597, 0, 4283, 0, -17516, 0, 79365, 131072, 79365, 0, -17516, 0, 4283, 0, -597
|
|
|
|
|
|
module med_hb_int
|
|
#(parameter WIDTH=18)
|
|
(input clk,
|
|
input rst,
|
|
input bypass,
|
|
input stb_in,
|
|
input [WIDTH-1:0] data_in,
|
|
input [7:0] output_rate,
|
|
input stb_out,
|
|
output reg [WIDTH-1:0] data_out);
|
|
|
|
localparam coeff_a = -597;
|
|
localparam coeff_b = 4283;
|
|
localparam coeff_c = -17516;
|
|
localparam coeff_d = 79365;
|
|
|
|
reg phase;
|
|
reg [WIDTH-1:0] d1, d2, d3, d4, d5, d6, d7, d8;
|
|
|
|
localparam MWIDTH = 36;
|
|
wire [MWIDTH-1:0] prod;
|
|
|
|
reg [6:0] stbin_d;
|
|
|
|
always @(posedge clk)
|
|
stbin_d <= {stbin_d[5:0],stb_in};
|
|
|
|
always @(posedge clk)
|
|
if(stb_in)
|
|
begin
|
|
d1 <= data_in;
|
|
d2 <= d1;
|
|
d3 <= d2;
|
|
d4 <= d3;
|
|
d5 <= d4;
|
|
d6 <= d5;
|
|
d7 <= d6;
|
|
d8 <= d7;
|
|
end
|
|
|
|
wire [WIDTH-1:0] sum_a, sum_b, sum_c, sum_d;
|
|
add2_and_round_reg #(.WIDTH(WIDTH)) add_a (.clk(clk),.in1(d1),.in2(d8),.sum(sum_a));
|
|
add2_and_round_reg #(.WIDTH(WIDTH)) add_b (.clk(clk),.in1(d2),.in2(d7),.sum(sum_b));
|
|
add2_and_round_reg #(.WIDTH(WIDTH)) add_c (.clk(clk),.in1(d3),.in2(d6),.sum(sum_c));
|
|
add2_and_round_reg #(.WIDTH(WIDTH)) add_d (.clk(clk),.in1(d4),.in2(d5),.sum(sum_d));
|
|
|
|
MULT18X18S mult1(.C(clk), .CE(1), .R(rst), .P(prod1), .A(stbin_d[1] ? coeff_a : coeff_b),
|
|
.B(stbin_d[1] ? sum_a : sum_b) );
|
|
MULT18X18S mult2(.C(clk), .CE(1), .R(rst), .P(prod2), .A(stbin_d[1] ? coeff_c : coeff_d),
|
|
.B(stbin_d[1] ? sum_c : sum_d) );
|
|
|
|
wire [MWIDTH:0] accum;
|
|
acc #(.IWIDTH(MWIDTH),.OWIDTH(MWIDTH+1))
|
|
acc (.clk(clk),.clear(stbin_d[2]),.acc(|stbin_d[3:2]),.in(prod),.out(accum));
|
|
|
|
wire [WIDTH+2:0] accum_rnd;
|
|
round_reg #(.bits_in(MWIDTH+1),.bits_out(WIDTH+3))
|
|
final_round (.clk(clk),.in(accum),.out(accum_rnd));
|
|
|
|
wire [WIDTH-1:0] clipped;
|
|
clip_reg #(.bits_in(WIDTH+3),.bits_out(WIDTH))
|
|
final_clip (.clk(clk),.in(accum_rnd),.out(clipped));
|
|
|
|
reg [WIDTH-1:0] saved, saved_d3;
|
|
always @(posedge clk)
|
|
if(stbin_d[6])
|
|
saved <= clipped;
|
|
|
|
always @(posedge clk)
|
|
if(stbin_d[3])
|
|
saved_d3 <= d3;
|
|
|
|
always @(posedge clk)
|
|
if(bypass)
|
|
data_out <= data_in;
|
|
else if(stb_in & stb_out)
|
|
case(output_rate)
|
|
1 : data_out <= d6;
|
|
2 : data_out <= d4;
|
|
3, 4, 5, 6, 7 : data_out <= d3;
|
|
default : data_out <= d2;
|
|
endcase // case(output_rate)
|
|
else if(stb_out)
|
|
data_out <= saved;
|
|
|
|
endmodule // small_hb_int
|
|
|