problem: to_ufixed(0.483,n1)

[fanwel]

Hi all;

I want to convert 0.483 to ufixed type. Manually, I get 0000.0111 (3 downto -4) when I calculate by hand.
But, this value is not same with the value in modelsim which is 0000.1000. I write like this;

ip1<=to_ufixed(0.483,n1);

What the correct value of 0.483 in ufixed? Do I calculate in wrong way?..needs help, thanks

 

[TrickyDicky]

0.483 x 2^4 = 7.728 which is 8 (1000) when you round to nearest.

to_ufixed has extra options for saturation and rounding you can add to the function.

 

[fanwel]

Sorry, what actually the formula to convert to ufixed? thanks for reply

 

[TrickyDicky]

Y x 2^f, where f is the number of fractional bits.

This will give you the number you can easily convert to binary.

 

[fanwel]

I try write a code for convert integer to ufixed:

package my_data_types is
type vector is array (natural range <>) of integer;
type ufixed is array (natural range <>) of std_logic;
end my_data_types;

library ieee;
library ieee_proposed;
use ieee_proposed.fixed_pkg.all;
use work.my_data_types.all;

entity fix is
port (clk: in bit;
nprev: in vector (0 to 7);
ip1: out ufixed (3 downto -4));
end fix;

architecture fix of fix is
signal n1: ufixed (3 downto -4);
begin
process(clk)
begin
if (clk'event and clk='1') then
for i in 0 to 7 loop
ip1(i) <= to_ufixed (nprev(i),n1);
end loop;
end if;
end process;
end fix;

I get this error:
Error (10482): VHDL error at fix.vhd(3): object "std_logic" is used but not declared.

How to declare the "std_logic" in package?..

 

[TrickyDicky]

you forgot to include the std_logic library

library ieee;
use ieee.std_logic_1164.all;

---------- Post added at 11:30 ---------- Previous post was at 11:29 ----------

Also, a big big big note:

DO NOT DECLARE UFIXED. It is in the fixed_pkg.

 

[fanwel]

What the right way to declare ufixed as an array?..thanks for reply

 

[TrickyDicky]

you dont need to define ufixed, it is already in the fixed_pkg.

so you you need to do is:

library floatfixlib; --(or IEEE_PROPOSED if you're using an older version)
use floatfixlib.fixed_pkg.all;

.....
signal a : ufixed(10 downto -27);

...etc

 

[fanwel]

Now I get this error:
Error (10511): VHDL Qualified Expression error at fix.vhd(23): to_ufixed type specified in Qualified Expression must match std_ulogic type that is implied for expression by context.
At this line: ip1(i) <= to_ufixed (nprev(i),n1);

What my mistake?..thanks for reply

 

[TrickyDicky]

ip1(i) is a single bit. You cannot assign an entire array (the ufixed) to a single bit.

You need to decalre an array of ufixed values.