Digital Filter FIR using VHDL

[ireon]

I should design a digital filter. I have coefficients values and I should write the VHDL code. I must implement the following circuit:

How can I describe it with the VHDL? I thought to use a for generate statement, but I don't manage to write the circuit. Someone could help me?

 

[axcdd]

You can create array that holds the coefficients values. Coefficient values needs to be integer type, so u have to scale them by multiplication with 2**X.
example: 0.074 * 2^8 = 18.94 ~ 19 = 00010011
then there is need to calculate number of protection bits in order not to overflow the output ( as i recall it could be log2( sum of known coefficients (real ones not scaled))
Z-1 equals to D-flip flop, so the design simply contains only DFF and mul commands. The only fact that i dont like here is that u have to implement not transponded FIR filter which is poor for FPGA for big filter depths ( to get high speed there is need to create tree of sum).

simple code (with prolly some error as i dont have time to test it)

---

Code VHDL - [expand]
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
--------------------------------------------------------------------------------
--! @file FIR_EDABOARD.vhd
--! @brief SIMPLE FIR
--! poor implementation version 
--! @author Axcdd
--! @date 2013.06.19
--------------------------------------------------------------------------------
 
 
--! COMPONENTS and ARRAY PACKAGE
 
library ieee;
use ieee.std_logic_1164.all;
use ieee.numeric_std.all;
 
entity DFF is
    generic (
      N : natural := 8 --! DATA_WIDTH;
    );
    port (
      CLK : in std_logic; --! clock signal
      RESET : in std_logic; --! asynchronize reset signal
      DIN : in signed(N-1 downto 0); --! input
      DOUT : out signed(N-1 downto 0) --! output
    );
end entity DFF;
    
architecture behave of DFF is
  
begin
  
  p0:process(CLK,RESET)
  begin
    --! asynchronize reset
    if RESET='1' then
      DOUT <= (others=>'0');
    elsif rising_edge(CLK) then
      DOUT <= DIN;
    end if;
  end process p0;
  
end architecture behave;
 
 
library ieee;
use ieee.std_logic_1164.all;
use ieee.numeric_std.all;
 
PACKAGE array_coefficient IS
  TYPE coefficient IS ARRAY(natural range <>) OF integer;
  
  component DFF is
    generic (
      N : natural := 8 --! DATA_WIDTH;
    );
    port (
      CLK : in std_logic; --! clock signal
      RESET : in std_logic; --! synchronize reset signal
      DIN : in signed(N-1 downto 0); --! input
      DOUT : out signed(N-1 downto 0) --! output
    );
    end component DFF;
    
END array_coefficient;
 
 
 
library ieee;
use ieee.std_logic_1164.all;
use ieee.numeric_std.all;
use work.array_coefficient.all;
 
 
entity FIR_EDABOARD is
  generic (
    N  : natural := 8;  --! DATA_WIDTH
    L  : natural := 26 ;  --! FIR DEEPTH
    W  : natural :=8;    --! coefficient  scalar (c * 2**W to make them integer)
    G  : natural :=2;  --! additional protection bits
    COE : coefficient  --! array of integer coefficients
  );
  port (
    CLK  : in std_logic;  --! FIR clk
    RESET  : in  std_logic;  --! asynchronize reset
    DIN  : in std_logic_vector(N-1 downto 0);  --! input
    DOUT  : out std_logic_vector(N-1 downto 0)  --! output
  );
end entity FIR_EDABOARD;
 
architecture ARCH_FIR of FIR_EDABOARD is
 
  type CON is array (0 to L-1) of signed(N-1 downto 0);
    signal QA : CON;  --! connections between DFF
 
  type CON1 is array (0 to L-2) of signed(N+W-1 downto 0);
    signal MUL : CON1; --! DFF outputs after MULTIPLY with coefficients
    
  begin
 
  -------------------------Z-1 part-------------------------  
   
  stage1: DFF generic map (N) port map (CLK, RESET, signed(DIN), QA(0)); --! input buffor not shown in picture
   
  g0:for i in 0 to L-2 generate
    Zminus1: DFF generic map (N) port map(CLK,RESET,QA(i),QA(i+1));
  end generate g0;
 
  ---------------MULTIPLICATION-----------------------------
  g1:for i in 0 to L-1 generate
    MULTIPLY: MUL(i) <= QA(i) * to_signed(COE(i),W); 
  end generate g1;
 
  ----------------OUTPUT-----------------------------------
  
  p0: process(MUL)
  variable temp: signed(N+W+G-1 downto 0);
  begin 
    --! sum of MUL
    temp := (others=>'0');
    for i in 0 to MUL'length-1 loop
      temp := temp + MUL(i);
    end loop;
    --! DOUT after dropping ',' bits.
    DOUT <= std_logic_vector(temp(N+W+G-1 downto W+G));
  end process p0;
  
end architecture ARCH_FIR;

---

 

[ireon]

OK, anyway I created a structural description of the circuit using the for generate statement:

library IEEE;
use IEEE.STD_LOGIC_1164.ALL;

entity filtro_fir is 
          port(clock,reset: in std_logic;
			      input: in std_logic_vector(15 downto 0);
					output: out std_logic_vector(25 downto 0)
					);
			      
end filtro_fir;

architecture Structural of filtro_fir is

component shift 
         port(clock, reset: in std_logic;
			     input: in std_logic_vector(15 downto 0);
				  output: out std_logic_vector(15 downto 0)
				  );
end component;

component multipleier
         port(input: in std_logic_vector(15 downto 0);
			     h: in std_logic_vector(9 downto 0);
			     output: out std_logic_vector(25 downto 0)
				  );
end component;

component adder
         port(input_1,input_2: in std_logic_vector(25 downto 0);
		        output: out std_logic_vector(25 downto 0)
				  );
end component;

type h_vector is array(0 to 8) of std_logic_vector(9 downto 0);

signal h: h_vector := (

8 => "0000000100",
0 => "0000001011",
1 => "0000011111",
2 => "0000110100",
3 => "0000111101",
4 => "0000110100",
5 => "0000011111",
6 => "0000001011",
7 => "0000000100");

type segnale_1 is array(0 to 8) of std_logic_vector(15 downto 0);
signal s1: segnale_1;

type segnale_2 is array(0 to 8) of std_logic_vector(25 downto 0);
signal s2: segnale_2;

type segnale_3 is array(0 to 8) of std_logic_vector(25 downto 0);
signal s3: segnale_3;


begin
s1(0)<=input;
U0: multipleier port map(input=>s1(0),h=>h(8),output=>s3(0));
filter_gen: for i in 0 to 7 generate
U1:shift port map(clock=>clock,reset=>reset,input=>s1(i),output=>s1(i+1));
U2:multipleier port map(input=>s1(i+1),h=>h(i),output=>s2(i));
U3:adder port map(input_1=>s3(i),input_2=>s2(i),output=>s3(i+1));
end generate;

output<=s3(8);

end Structural;

library IEEE;
use IEEE.STD_LOGIC_1164.ALL;

entity FF_D is
            port (D,CLK,RESET: in std_logic;
                  Q: out std_logic);
end FF_D;

architecture behavioral of FF_D is

begin
             process (CLK,RESET)
begin
                     if RESET='1' then
							
							Q <= '0';
                     elsif (CLK'event and CLK='1') then
                     Q <= D;
                     
							end if;

             end process;

end behavioral;

library IEEE;
use IEEE.STD_LOGIC_1164.ALL;

entity shift is
         port(clock, reset: in std_logic;
			         input: in std_logic_vector(15 downto 0);
						output: out std_logic_vector(15 downto 0)
						);
end shift;
						
architecture Structural of shift is

component FF_D
          port(D,CLK,RESET: in std_logic;
                  Q: out std_logic);
end component;
						
begin

shift_gen: for i in 0 to 15 generate
U_shift: FF_D port map(clk=>clock,reset=>reset,D=>input(i),Q=>output(i));
end generate;

end Structural;

library IEEE;
use IEEE.STD_LOGIC_1164.ALL;
use IEEE.STD_LOGIC_SIGNED.ALL;

entity multipleier is
         port(input: in std_logic_vector(15 downto 0);
			     h: in std_logic_vector(9 downto 0);
			     output: out std_logic_vector(25 downto 0)
				  );

end multipleier;

architecture Behavioral of multipleier is

begin

output<=input*h;

end Behavioral;

library IEEE;
use IEEE.STD_LOGIC_1164.ALL;
use IEEE.STD_LOGIC_SIGNED.ALL;

entity adder is
       port(input_1,input_2: in std_logic_vector(25 downto 0);
		      output: out std_logic_vector(25 downto 0)
				);
end adder;

architecture Behavioral of adder is

begin

output<=input_1+input_2;

end Behavioral;

The circuit has been generated but I have a warning message: "Using initial value ("0000001011","0000011111","0000110100","0000111101","0000110100","0000011111","0000001011","0000000100","0000000100") for h since it is never assigned". What is the problem? I created the new type assigning the h values with the for generate. The filter has 16 bit input and 9 coefficients.

 

[TrickyDicky]

Its not a problem. H is only given an initial value, so it is effectively a constant. If you dont assign a signal a value during runtime you will get this warning.

 

[ireon]

Ok. I executed a Test Bench and the circuit works perfectly.

 

[fasto2008]

I should design a digital filter of first order ADC Sigma-Delta converter with 8 bits, an over sampling ratio of 64 (OSR = 64),and sampling frequency of 10.24MHz.I make it with MAtLAB SIMULINK as shown in this figure:




Can somone help me by giving a vhdl Code of this filter because i must implement it.

Thanks for help

 

[TrickyDicky]

I should design
....

Can somone help me by giving a vhdl Code of this filter because i must implement it.

Thanks for help

I think the clue is that you should do the work, not someone else.