I am using a high speed clock (from an internal PLL) and attempting to divide it down to generate 2 clocks with varying duty and phase relationships. The code operates correctly if I single-step the input clock. As the frequency increases, the phase and duty of the secondary output (iDM_out in the code sample) gets corrupted. At some frequencies, the duty cycle of the secondary output will be correct. At other frequencies, the duty cycle might go to 90 or 199%. Ditto with the phase relationship (DMDelay). I've seem some articles about the need for clock buffers so I've trying a few types BUFF, CLKBUF and CLKINT on the output and it seems to make it worse. Does anybody have any ideas about what might be causing this condition?
use work.A208_pckgs.all;
library ieee;
use ieee.std_logic_1164.all;
use ieee.numeric_std.all;
entity DutyPhaseMod is
port (
rst : in std_logic;
GLA : in std_logic; -- GLA is 36x faster than target frequency
EXDuty : in std_logic_vector(4 downto 0); -- the number of GLA clock cycles for the EX duty cycle
DMDuty : in std_logic_vector(4 downto 0); -- the number of GLA clock cycles for the DM duty cycle
DMDelay : in std_logic_vector(4 downto 0); -- the number of GLA clock cycles for the DM phase delay
EX_out : out std_logic;
DM_out : out std_logic
);
end entity;
architecture behavioral of DutyPhaseMod is
signal iGLA : std_logic;
signal iFREQlen : integer range 0 to 35;
begin
process (rst, GLA)
variable iEX_out : std_logic;
variable iDM_out : std_logic;
variable iEXctr : natural range 0 to 35;
variable iDMctr : natural range 0 to 35;
begin
iFREQlen <= 35; -- number of clock cycles
if rst = '1' then -- reset the counters on reset signal
iEXctr := 0;
iDMctr := 0;
elsif rising_edge(GLA) then
if iEXctr <= unsigned(EXDuty) then -- first part of EX the duty cycle
iEX_out := '1';
iEXctr := iEXctr + 1;
elsif iEXctr < iFREQlen then -- second part of EX duty cycle
iEX_out := '0';
iEXctr := iEXctr + 1;
else -- set for the start of the next cycle
iEX_out := '1';
iEXctr := 0;
end if;
if iEXctr = unsigned(DMDelay) then -- reset for DM phase offset
iDMctr := 0;
elsif iDMctr <= unsigned(DMDuty) then -- first part of the DM duty cycle
iDM_out := '1';
iDMctr := iDMctr + 1;
elsif iDMctr <= iFREQlen - 1 then -- second part of the DM duty cycle
iDM_out := '0';
iDMctr := iDMctr + 1;
else -- set for the start of the next cycle
iDM_out := '1';
end if;
EX_out <= iEX_out;
DM_out <= iDM_out;
end if;
end process;
end behavioral;
I'm going to go out on a limb and answer my own question. Feel free to comment if you feel that my answer is not correct (or you can add/clarify further.) I decided to split my entity into two processes and it seems to work properly. The first process handles the base clock division on the rising_edge of the high speed clock. The second process handles the secondary (phase delayed) clock on the falling_edge of the high speed clock.
library ieee;
use ieee.std_logic_1164.all;
use ieee.numeric_std.all;
entity DutyPhaseMod is
port (
rst : in std_logic;
GLA : in std_logic;
EXDuty : in std_logic_vector(3 downto 0);
DMDuty : in std_logic_vector(3 downto 0);
DMDelay : in std_logic_vector(4 downto 0);
EX_out : out std_logic;
DM_out : out std_logic
);
end entity;
architecture behavioral of DutyPhaseMod is
signal iGLA : std_logic;
signal iFREQlen : integer range 0 to 35;
signal iEXctr : natural range 0 to 35;
begin
process (rst, GLA)
variable iEX_out : std_logic;
begin
iFREQlen <= 35;
if rst = '1' then
iEXctr <= 0;
elsif rising_edge(GLA) then
if iEXctr <= unsigned(EXDuty) then
iEX_out := '1'; --
iEXctr <= iEXctr + 1;
elsif iEXctr < iFREQlen then
iEX_out := '0';
iEXctr <= iEXctr + 1;
else
iEX_out := '1';
iEXctr <= 0;
end if;
EX_out <= iEX_out;
end if;
end process;
process (rst, GLA)
variable iDM_out : std_logic;
variable iDMctr : natural range 0 to 30;
begin
case FW_FREQ is
when F_46 =>
iFREQlen <= 35;
when F_82 =>
iFREQlen <= 23;
end case;
if rst = '1' then
iDMctr := 0;
elsif falling_edge(GLA) then
if iEXctr = unsigned(DMDelay) then
iDMctr := 0;
elsif iDMctr <= unsigned(DMDuty) then
iDM_out := '1';
iDMctr := iDMctr + 1;
elsif iDMctr <= iFREQlen - 1 then
iDM_out := '0';
iDMctr := iDMctr + 1;
else
iDM_out := '1';
end if;
DM_out <= iDM_out;
end if;
end process;
end behavioral;
Related
I want to make 4kb byte addressable memory. sorry I'm new in VHDL
I wanted my code works first write 4byte number in adress 8 (rdwr=1, addr=1000, size=10(2^2byte), idata=10001100)
then wait 8 cycles to implement writing time(ivalid=0)
Second read 4byte number from adress 8(rdwr=0, addr=1000, size=10(2^2byte))
In my purpose, the "ready" signal should be '0' while waiting for writing time
but the signal is always 'U' in simulation. I tried to figure out what is the problem but i couldn't
Can anyone help me where did a make mistake?
Here is my code
library IEEE;
use IEEE.STD_LOGIC_1164.ALL;
use IEEE.NUMERIC_STD.ALL;
entity Memory is
port (
clk: in std_logic;
ready: out std_logic; -- 0: busy, 1: ready
ivalid: in std_logic; -- 0: invalid, 1: valid
rdwr: in std_logic; -- 0: read, 1: write
addr: in unsigned(11 downto 0); -- byte address
size: in std_logic_vector(1 downto 0); -- 00/01/10/11: 1/2/4/8 bytes
idata: in std_logic_vector(63 downto 0);
ovalid: out std_logic; -- 0: invalid, 1: valid
odata: out std_logic_vector(63 downto 0)
);
end entity;
architecture Behavioral of Memory is
type ram_type is array (0 to (2**12)-1) of std_logic_vector(7 downto 0);
signal RAM : ram_type;
signal state : std_logic := '1'; --if ready '1'
signal queue : std_logic := '0'; --if something to do '1'
signal timer : integer := 0; --timer
signal curr_addr : unsigned(11 downto 0);
signal curr_size : std_logic_vector(1 downto 0);
signal curr_data : std_logic_vector(63 downto 0);
signal write : std_logic := '0';
signal read : std_logic := '0';
begin
process(clk)
variable vstate : std_logic := state;
variable vqueue : std_logic := queue; --if something to do '1'
variable vtimer : integer := timer; --timer
variable vcurr_addr : unsigned(11 downto 0) := curr_addr;
variable vcurr_size : std_logic_vector(1 downto 0) := curr_size;
variable vcurr_data : std_logic_vector(63 downto 0) := curr_data;
variable vwrite : std_logic := write;
variable vread : std_logic := read;
begin
--get input
if(rising_edge(clk)) then
ovalid <= '0';
if(vstate='1') then
if(ivalid='1') then
vcurr_addr := addr;
vcurr_size := size;
if(rdwr='0') then
--read
vread := '1';
else
vwrite := '1';
vcurr_data := idata;
end if;
vqueue := '1';
vtimer := 2**(to_integer(unsigned(vcurr_size)))-1;
end if;
end if;
--process
if(vqueue = '1') then
if(vtimer > 0) then
--wait for next cycle
ready <= '0';
vstate := '0';
vtimer := vtimer - 1;
else
--ok to go
if(vwrite = '1') then
--write
for x in 0 to 2**(to_integer(unsigned(vcurr_size)))-1 loop
for y in 0 to 7 loop
RAM(to_integer(vcurr_addr) + x)(y) <= vcurr_data(y + 8*x);
end loop;
end loop;
elsif(vread = '1') then
--read
for x in 0 to 7 loop
for y in 0 to 7 loop
if(x < 2**(to_integer(unsigned(vcurr_size)))) then
odata(y + 8*x) <= RAM(to_integer(vcurr_addr) + x)(y);
else
odata(y + 8*x) <= '0';
end if;
end loop;
end loop;
ovalid <= '1';
end if;
vqueue := '0';
vstate := '1';
ready <= '1';
end if;
end if;
--save variable to signals
state <= vstate;
queue <= vqueue;
timer <= vtimer;
curr_addr <=vcurr_addr;
curr_size <=vcurr_size;
curr_data<= vcurr_data;
write <= vwrite;
read <= vread;
end if;
end process;
end architecture;
I've written the following VHDL code and test-bench to control the brightness of LEDs. Then tried with Altera ModelSim to simulate the code. However, I'm facing some technical difficulties. It would be really nice if someone will compile the code and share me the simulated result with me. Thanks in advance.
library IEEE;
use IEEE.STD_LOGIC_1164.ALL;
use IEEE.NUMERIC_STD.all;
use IEEE.STD_LOGIC_UNSIGNED.ALL;
entity main_testbench is
end main_testbench;
architecture behavior of main_testbench is
component led_controller is
Port ( clk, reset: in std_logic; -- system clock is assumed 10KHz
in_word: in std_logic_vector(7 downto 0); -- LS 4 bits - frequency & MS 4 bits - duty-cycle
LEDs: out std_logic_vector(3 downto 0));
end component;
signal clk, reset: std_logic := '0';
signal in_word: std_logic_vector(7 downto 0) := "00010001"; -- 0.2 Hz, 10% duty cycle
signal LEDs: std_logic_vector(3 downto 0) := "0000";
type in_word_commands is array (0 to 15) of std_logic_vector(7 downto 0);
signal in_words: in_word_commands := ("00010001", "00010010", "00010100", "00011000", -- 10% duty cycle with 0.2Hz, 0.5Hz, 1Hz, 2Hz
"00100001", "00100010", "00100100", "00101000", -- 30% duty cycle with 0.2Hz, 0.5Hz, 1Hz, 2Hz
"01000001", "01000010", "01000100", "01001000", -- 60% duty cycle with 0.2Hz, 0.5Hz, 1Hz, 2Hz
"10000001", "10000010", "10000100", "10001000"); -- 85% duty cycle with 0.2Hz, 0.5Hz, 1Hz, 2Hz
signal command_num : integer := 0;
begin
dut: led_controller port map (clk, reset, in_word, LEDs);
clk <= not clk after 50 us; -- 0.1ms/2 = 50us
command_num <= command_num + 1 after 5000 ms; -- 5000ms = 5s
in_word <= in_words(command_num);
end behavior;
controller:
library IEEE;
use IEEE.STD_LOGIC_1164.ALL;
use IEEE.NUMERIC_STD.all;
entity led_controller is
Port ( clk, reset: in std_logic;
in_word: in std_logic_vector(7 downto 0);
LEDs: out std_logic_vector(3 downto 0));
end led_controller;
architecture behavior of led_controller is
--------------------- signals ---------------------
type freq is array (0 to 3) of integer range 0 to 50000;
signal frq: freq := (25000, 10000, 5000, 2500);
signal led_freq_count: integer range 0 to 50000 := frq(0);
type d is array (0 to 3) of integer range 0 to 100;
signal duty: d := (10, 30, 60, 85);
signal duty_cycle: integer range 0 to 100 := duty(0);
signal LED_switch, new_command: std_logic := '0';
begin
--------- clock process / sync reset configuration ---------------
process (clk)
variable duty_counter: integer range 0 to 100 := 100;
variable freq_counter: integer range 0 to 50000 := led_freq_count;
begin
if rising_edge(clk) then
------- if reset was high or new in_word were arrived --------
if reset = '1' or new_command = '1' then
LEDs <= "0000";
duty_counter := 100;
freq_counter := led_freq_count;
new_command <= '0';
else
------- blinking process --------
if freq_counter = 0 then
freq_counter := led_freq_count;
LED_switch <= not LED_switch;
end if;
freq_counter := freq_counter - 1;
if duty_counter = 0 then
duty_counter := 100;
end if;
duty_counter := duty_counter - 1;
------ output assignment -------
if LED_switch = '1' and duty_counter < duty_cycle then
LEDs <= "1111";
else
LEDs <= "0000";
end if;
end if;
end if;
end process;
--------- input process---------------
process (in_word)
begin
case in_word(3 downto 0) is
when "0001" => led_freq_count <= frq(0);
when "0010" => led_freq_count <= frq(1);
when "0100" => led_freq_count <= frq(2);
when "1000" => led_freq_count <= frq(3);
when others => led_freq_count <= frq(0);
end case;
case in_word(7 downto 4) is
when "0001" => duty_cycle <= duty(0);
when "0010" => duty_cycle <= duty(1);
when "0100" => duty_cycle <= duty(2);
when "1000" => duty_cycle <= duty(3);
when others => duty_cycle <= duty(0);
end case;
new_command <= '1';
end process;
end behavior;
if freq_counter = 0 then
freq_counter := led_freq_count;
LED_switch <= not LED_switch;
else
freq_counter := freq_counter - 1;
end if;
if duty_counter = 0 then
duty_counter := 100;
else
duty_counter := duty_counter - 1;
end if;
I try to implement a divisor for std_logic_vector with the language VHDL.
After I have implemented it, I have to use GHDL but encountered this error :
vhdl:error: bound check failure at divisor.vhdl
It's correspond to this :
else ....
nb_reg <= (2*nb);
My code is :
library ieee;
use ieee.std_logic_1164.all;
use ieee.numeric_std.all;
entity div is
generic
(
DATA_WIDTH : natural := 8;
ADDR_WIDTH : natural := 3
);
port (
clk : in std_logic;
reset : in std_logic;
raddr : in natural range 0 to 2**ADDR_WIDTH - 1;
waddr : in natural range 0 to 2**ADDR_WIDTH - 1;
max : in std_logic_vector((DATA_WIDTH -1) downto 0);
data : in std_logic_vector((DATA_WIDTH -1) downto 0);
q : out std_logic_vector((DATA_WIDTH -1) downto 0)
);
end div;
architecture rtl of div is
-- Build a 2-D array type for the RAM
subtype word_t is unsigned((DATA_WIDTH-1) downto 0);
type memory_t is array(natural range <>) of word_t;
-- Declare the RAM signal.
signal ram, div : memory_t((2**ADDR_WIDTH)-1 downto 0);
-- Declare the n and max_reg signal.
signal n,m : unsigned(((2*DATA_WIDTH)-1) downto 0);
signal max_reg : unsigned(((2*DATA_WIDTH)-1) downto 0);
signal nb : unsigned(((2*DATA_WIDTH)-1) downto 0);
signal nb_reg : unsigned(((4*DATA_WIDTH)-3) downto 0);
-- Build div_reg, ram_reg
type memory_reg is array(natural range <>) of unsigned(((2*DATA_WIDTH)-1) downto 0);
signal ram_reg: memory_reg((2**ADDR_WIDTH)-1 downto 0);
begin
process(clk,reset)
begin
if (reset = '1') then
ram(waddr) <= X"00";
div(waddr) <= X"00";
max_reg <= X"0000";
ram_reg(waddr) <= X"0000";
n <= X"0000";
nb <= X"0000";
--nb_reg(((4*DATA_WIDTH)-3) downto 0) <= "0";
m <= X"0000";
elsif(rising_edge(clk)) then
ram(waddr) <= unsigned(data);
max_reg((DATA_WIDTH -1) downto 0) <= unsigned(max);
ram_reg(waddr)((DATA_WIDTH-1) downto 0) <= ram(waddr)((DATA_WIDTH-1) downto 0);
nb <= (nb+1);
if (ram(waddr) = max_reg)then
div(waddr) <= div(waddr)+1;
elsif (ram(waddr) > max_reg)then
while ((div(waddr)*(ram(waddr)) > max_reg) or (m <(DATA_WIDTH -1))) loop
div(waddr) <= (div(waddr))+1;
max_reg <= (max_reg) - (ram_reg(waddr));
m <= m+1;
end loop;
m <= (m-1);
while (((div(waddr)*ram_reg(waddr)) < max_reg-1) or (n <(DATA_WIDTH)-(m))) loop
ram_reg(waddr)(((2*DATA_WIDTH)-1) downto 1) <= ram_reg(waddr)(((2*DATA_WIDTH)-2) downto 0);
ram_reg(waddr)(0) <= '0';
n <= n+1;
nb_reg <= (nb*2);
nb(((2*DATA_WIDTH)-1) downto 0) <= nb_reg(((2*DATA_WIDTH)-1) downto 0);
end loop;
ram_reg(waddr) <= ram_reg(waddr) - (max_reg);
div(waddr) <= (div(waddr))+(1/(nb));
else
while (((div(waddr)*ram_reg(waddr)) < max_reg-1) or (n <(DATA_WIDTH)-(m))) loop
ram_reg(waddr)(((2*DATA_WIDTH)-1) downto 1) <= ram_reg(waddr)(((2*DATA_WIDTH)-2) downto 0);
ram_reg(waddr)(0) <= '0';
n <= n+1;
nb_reg <= (2*nb);
nb(((2*DATA_WIDTH)-1) downto 0) <= nb_reg(((2*DATA_WIDTH)-1) downto 0);
end loop;
ram_reg(waddr) <= ram_reg(waddr) - (max_reg);
div(waddr) <= (div(waddr))+(1/(nb));
end if;
else null;
end if;
end process;
q <= std_logic_vector(div(waddr));
end rtl;
The test bench :
-- descrizione del Test_bench
library IEEE;
use IEEE.std_logic_1164.all;
use IEEE.numeric_std.all;
entity div_tb is
end div_tb;
architecture behaviour of div_tb is
--dichiarazione dei COMPONENT ovvero la Unit Under Test
component div is
generic
(
DATA_WIDTH : natural := 8;
ADDR_WIDTH : natural := 3
);
port
(
clk : in std_logic;
reset : in std_logic;
raddr : in natural range 0 to 2**ADDR_WIDTH - 1;
waddr : in natural range 0 to 2**ADDR_WIDTH - 1;
max : in std_logic_vector((DATA_WIDTH -1) downto 0);
data : in std_logic_vector((DATA_WIDTH -1) downto 0);
q : out std_logic_vector((DATA_WIDTH -1) downto 0)
);
end component;
-- Clock period definitions
constant clk_period : time := 1 us;
constant DATA_WIDTH : natural := 8;
constant ADDR_WIDTH : natural := 3;
signal CLK_tb: std_logic := '0';
signal RESET_tb: std_logic := '1';
signal raddr_tb, waddr_tb : natural range 0 to 2**ADDR_WIDTH - 1;
signal data_tb, q_tb, max_tb : std_logic_vector((DATA_WIDTH -1) downto 0);
signal I : integer := 0; -- variabile per il conteggio dei clock
begin
clk_process: process --processo di generazione del CLK
begin
CLK_tb <= '0';
wait for clk_period/2;
CLK_tb <= '1';
wait for clk_period/2;
I<=I+1;
if I=200 then wait; -- durata della simulazione: 30 colpi di CLK
else null;
end if;
end process;
-- istanziazione della Unit Under Test
UUT: div generic map (ADDR_WIDTH => 3, DATA_WIDTH => 8)
port map (clk=>clk_tb, reset=>RESET_tb, raddr => raddr_tb, waddr => waddr_tb , data => data_tb, q => q_tb, max => max_tb);
stimoli: process
begin
RESET_tb <= '1';
wait for clk_period*3;
RESET_tb <= '0';
wait;
end process;
we: process
begin
max_tb <= "11100110";
wait;
end process;
Data : process
begin
data_tb <= "00000000"; raddr_tb <= 0; waddr_tb <= 0; wait for clk_period*3;
data_tb <= "01010110"; raddr_tb <= 1; waddr_tb <= 1; wait for clk_period*8;
data_tb <= "01000110"; raddr_tb <= 2; waddr_tb <= 2; wait for clk_period*8;
data_tb <= "11001110"; raddr_tb <= 3; waddr_tb <= 3; wait for clk_period*8;
data_tb <= "01000111"; raddr_tb <= 4; waddr_tb <= 4; wait for clk_period*8;
data_tb <= "11100110"; raddr_tb <= 5; waddr_tb <= 5; wait for clk_period*8;
data_tb <= "01000110"; raddr_tb <= 6; waddr_tb <= 6; wait for clk_period*8;
data_tb <= "01010110"; raddr_tb <= 7; waddr_tb <= 7; wait for clk_period*8;
wait;
end process;
end behaviour;
I use the 2007 version of GHDL and it is not possible for me to update, my teacher wants me to use this one.
Could somebody help me with this code?
The multiplication operator in the numeric_std package is defined thus for multiplying a natural by an unsigned:
-- Id: A.18
function "*" (L: NATURAL; R: UNSIGNED) return UNSIGNED;
-- Result subtype: UNSIGNED((R'LENGTH+R'LENGTH-1) downto 0)
-- Result: Multiplies an UNSIGNED vector, R, with a nonnegative
-- INTEGER, L. L is converted to an UNSIGNED vector of
-- SIZE R'LENGTH before multiplication.
As you can see, the output is twice the width of the unsigned argument minus 1. In this line
nb <= std_logic_vector(2*unsigned(nb));
you are assigning the result back to the unsigned operand, which obviously is not a different width to itself.
So, you either need an intermediate variable of the right width or perhaps a shift left?
I am working on trying to interface a cheap FPGA (the ep2c5t144 Altera Cyclone II mini board) with a SNES in order to ACT as a SNES controller. So far, it seems to work on and off... with the current problem being, it works for about 1 second after switched on... but then seems to get stuck in a state until it is reset.
Since I have spent a long time looking at the code for a logic issue, I'm starting to wonder whether it's some strange quirk of using FPGAs, but I've already tried testing for any states which aren't defined, and that hasn't fixed the problem. I will post the SNES code below, and the output from my cheap logic analyser which shows the problem. Warning, the code is quite messy... especially with me changing thing around to try to fix it. Any ideas at all with be much appreciated!
Many thanks in advance for any help!
Problem from logic analyser:
When a request works - State transitions occur as expected
When a request fails - SEEMS to incorrectly transition directly to "working" state and get stuck for some reason
library ieee;
use ieee.std_logic_1164.all;
use ieee.numeric_std.all;
entity snes_controller is
generic (
hp : integer := 300
);
port (
clk : in std_logic;
latch : in std_logic;
data : out std_logic := '0';
clock : in std_logic;
enable : in std_logic;
btn_B : in std_logic;
btn_Y : in std_logic;
btn_select : in std_logic;
btn_start : in std_logic;
btn_up : in std_logic;
btn_down : in std_logic;
btn_left : in std_logic;
btn_right : in std_logic;
btn_A : in std_logic;
btn_X : in std_logic;
btn_L : in std_logic;
btn_R : in std_logic;
helpA : out std_logic := '0';
helpB : out std_logic := '0';
helpC : out std_logic := '0';
helpD : out std_logic := '0';
helpE : out std_logic := '0'
);
end entity;
architecture Behav of snes_controller is
signal buttons : unsigned(16 downto 0) := "10000000000000000";
type state_type is (s_idle, s_latching_1, s_latching_2, s_working);
signal state : state_type := s_idle;
type cycle_type is (c_high, c_low);
signal cycle : cycle_type := c_high;
begin
process (clk)
variable i : integer range 0 to 16;
variable count : integer range 0 to hp;
begin
if(rising_edge(clk)) then
data <= not buttons(i);
if(state = s_latching_1 or state = s_latching_2 or state = s_working) then
if(count < hp) then
count := count+1;
else
count := 0;
if(state = s_latching_1) then
if(latch = '1') then
state <= s_latching_2;
buttons(0) <= btn_B;
buttons(1) <= btn_Y;
buttons(2) <= btn_select;
buttons(3) <= btn_start;
buttons(4) <= btn_up;
buttons(5) <= btn_down;
buttons(6) <= btn_left;
buttons(7) <= btn_right;
buttons(8) <= btn_A;
buttons(9) <= btn_X;
buttons(10) <= btn_L;
buttons(11) <= btn_R;
else
state <= s_idle;
end if;
elsif(state = s_latching_2) then
state <= s_working;
i := 0;
cycle <= c_high;
elsif(state = s_working) then
if(latch = '1') then
state <= s_idle;
helpD <= '1';
elsif(cycle = c_low) then
cycle <= c_high;
if(i < 16) then
i := i+1;
else
state <= s_idle;
helpD <= '0';
helpE <= '0';
end if;
else
cycle <= c_low;
end if;
end if;
end if;
elsif(state = s_idle) then
if(latch = '1') then
state <= s_latching_1;
count := 0;
i := 0;
end if;
else
helpE <= '1';
state <= s_idle;
count := 0;
i := 0;
end if;
end if;
end process;
process(state)
begin
if(state = s_idle) then
helpA <= '0';
helpB <= '0';
elsif(state = s_latching_1) then
helpA <= '1';
helpB <= '0';
elsif(state = s_latching_2) then
helpA <= '0';
helpB <= '1';
elsif(state = s_working) then
helpA <= '1';
helpB <= '1';
else
helpA <= clk;
helpB <= not clk;
end if;
if(cycle = c_low) then
helpC <= '0';
elsif(cycle = c_high) then
helpC <= '1';
end if;
end process;
end Behav;
You are using asynchronous external inputs and feed them into a synchronous, clock-based, state machine. Metastability in sampling may lead to your problem. Make sure you implement at least a two-flop synchronizer for every input signal.
Read more about it here: http://webee.technion.ac.il/~ran/papers/Metastability%20and%20Synchronizers.posted.pdf
library ieee;
use ieee.std_logic_1164.all;
use ieee.numeric_std.all;
entity snes_controller is
generic (
hp : integer := 300
);
port (
clk : in std_logic;
latch : in std_logic;
data : out std_logic := '0';
clock : in std_logic;
enable : in std_logic;
btn_B : in std_logic;
btn_Y : in std_logic;
btn_select : in std_logic;
btn_start : in std_logic;
btn_up : in std_logic;
btn_down : in std_logic;
btn_left : in std_logic;
btn_right : in std_logic;
btn_A : in std_logic;
btn_X : in std_logic;
btn_L : in std_logic;
btn_R : in std_logic;
helpA : out std_logic := '0';
helpB : out std_logic := '0';
helpC : out std_logic := '0';
helpD : out std_logic := '0';
helpE : out std_logic := '0'
);
end entity;
architecture Behav of snes_controller is
signal synch0 : unsigned(11 downto 0) := (others => '0');
signal synch1 : unsigned(11 downto 0) := (others => '0');
signal synch2 : unsigned(11 downto 0) := (others => '0');
signal buttons : unsigned(16 downto 0) := "10000000000000000";
type state_type is (s_idle, s_latching_1, s_latching_2, s_working);
signal state : state_type := s_idle;
type cycle_type is (c_high, c_low);
signal cycle : cycle_type := c_high;
begin
process (clk)
variable i : integer range 0 to 16;
variable count : integer range 0 to hp;
begin
if(rising_edge(clk)) then
synch0(0) <= btn_B;
synch0(1) <= btn_Y;
synch0(2) <= btn_select;
synch0(3) <= btn_start;
synch0(4) <= btn_up;
synch0(5) <= btn_down;
synch0(6) <= btn_left;
synch0(7) <= btn_right;
synch0(8) <= btn_A;
synch0(9) <= btn_X;
synch0(10) <= btn_L;
synch0(11) <= btn_R;
synch1 <= synch0;
synch2 <= synch1;
data <= not buttons(i);
if(state = s_latching_1 or state = s_latching_2 or state = s_working) then
if(count < hp) then
count := count+1;
else
count := 0;
if(state = s_latching_1) then
if(latch = '1') then
state <= s_latching_2;
buttons(11 downto 0) <= synch2(11 downto 0);
else
state <= s_idle;
end if;
elsif(state = s_latching_2) then
state <= s_working;
i := 0;
cycle <= c_high;
elsif(state = s_working) then
if(latch = '1') then
state <= s_idle;
helpD <= '1';
elsif(cycle = c_low) then
cycle <= c_high;
if(i < 16) then
i := i+1;
else
state <= s_idle;
helpD <= '0';
helpE <= '0';
end if;
else
cycle <= c_low;
end if;
end if;
end if;
elsif(state = s_idle) then
if(latch = '1') then
state <= s_latching_1;
count := 0;
i := 0;
end if;
else
helpE <= '1';
state <= s_idle;
count := 0;
i := 0;
end if;
end if;
end process;
process(state)
begin
if(state = s_idle) then
helpA <= '0';
helpB <= '0';
elsif(state = s_latching_1) then
helpA <= '1';
helpB <= '0';
elsif(state = s_latching_2) then
helpA <= '0';
helpB <= '1';
elsif(state = s_working) then
helpA <= '1';
helpB <= '1';
else
helpA <= clk;
helpB <= not clk;
end if;
if(cycle = c_low) then
helpC <= '0';
elsif(cycle = c_high) then
helpC <= '1';
end if;
end process;
end Behav;
Additionally, I suggest to create some sort of filter to handle debouncing of button clicks. http://www.eng.utah.edu/~cs5780/debouncing.pdf
I have an issue with self implemented UART in VHDL.
I wrote VHDL code which generates proper waveform when running on Altera ModelSim:
UART.vhd:
LIBRARY ieee;
USE ieee.std_logic_1164.all;
use ieee.numeric_std.ALL;
entity UART is
port (
clk_10mhz: in STD_LOGIC;
uart_clk: out STD_LOGIC;
txPin: out STD_LOGIC
);
end entity;
architecture Test of UART is
signal txStart: STD_LOGIC := '0';
signal txIdle: STD_LOGIC;
signal txData: STD_LOGIC_VECTOR(7 downto 0);
component TX is
port (
clk_in: in STD_LOGIC;
start: in STD_LOGIC;
data: in STD_LOGIC_VECTOR(7 downto 0);
tx: out STD_LOGIC;
txIdle: out STD_LOGIC;
debug_clk: out STD_LOGIC
);
end component TX;
begin
process (clk_10mhz)
variable clkDividerCounter : integer range 0 to 10000000;
variable textToSend : string(1 to 31) := "Hello darkness my old friend!" & LF & CR;
variable currentCharacterIndex : integer range 1 to 31 := 1;
variable startSending : std_logic := '0';
variable characterReceivedByTX : std_logic := '1';
begin
if (rising_edge(clk_10mhz)) then
if (startSending = '1') then
if (txIdle = '0') then
characterReceivedByTX := '1';
end if;
if (txIdle = '1' and characterReceivedByTX = '1') then
txData <= std_logic_vector(to_unsigned(character'pos(textToSend(currentCharacterIndex)), 8));
txStart <= '1';
if (currentCharacterIndex < 31) then
currentCharacterIndex := currentCharacterIndex + 1;
characterReceivedByTX := '0';
else
txStart <= '0';
currentCharacterIndex := 1;
startSending := '0';
end if;
end if;
else
if (clkDividerCounter < 10000000) then
clkDividerCounter := clkDividerCounter + 1;
startSending := '0';
else
clkDividerCounter := 0;
startSending := '1';
end if;
end if;
end if;
end process;
u1: TX port map (clk_10mhz, txStart, txData, txPin, txIdle, uart_clk);
end Test;
TX.vhd:
LIBRARY ieee;
USE ieee.std_logic_1164.all;
use ieee.numeric_std.ALL;
entity TX is
port (
clk_in: in STD_LOGIC;
start: in STD_LOGIC;
data: in STD_LOGIC_VECTOR(7 downto 0);
tx: out STD_LOGIC := '1';
txIdle: out STD_LOGIC := '1';
debug_clk: out STD_LOGIC := '0'
);
end entity;
architecture Test of TX is
signal idle: STD_LOGIC := '1';
begin
process (clk_in)
variable bitIndex : integer range 0 to 9;
variable clkDividerCounter : integer range 0 to 1042;
variable dataFrame : STD_LOGIC_VECTOR(9 downto 0);
variable dataFrameCurrentIndex : integer range 0 to 9;
begin
if (rising_edge(clk_in)) then
if (start = '1' and idle = '1') then
dataFrame(0) := '0';
dataFrame(8 downto 1) := data;
dataFrame(9) := '1';
dataFrameCurrentIndex := 0;
idle <= '0';
end if;
if (idle = '0') then
if (clkDividerCounter < 521) then
debug_clk <= '0';
else
debug_clk <= '1';
end if;
if (clkDividerCounter < 1041) then
clkDividerCounter := clkDividerCounter + 1;
else
if (dataFrameCurrentIndex < 9) then
tx <= dataFrame(dataFrameCurrentIndex);
dataFrameCurrentIndex := dataFrameCurrentIndex + 1;
else
tx <= dataFrame(dataFrameCurrentIndex);
idle <= '1';
end if;
clkDividerCounter := 0;
end if;
else
debug_clk <= '0';
end if;
end if;
end process;
txIdle <= idle;
end Test;
Unfortunately, on hardware, instead of "Hello darkness my old friend!" sent, it sends "HHello darkness my old friend!" with double H at the beginning.
I checked it on SignalTap II and waveform confirms the problem:
What can cause this problem? How may I debug such an issue?