Honored Contributorsir i use fixed-pkg to represent fraction bits,but when the output is 0.05118,whose binary representation is something like 0.00001101,in waveform i get only 0.should i use float_pkg to get the correct output?:cry:
Honored Contributordo you have enough fractional bits in your signal?
and no, don't use float. without seeing the rest of your code we cant say what's wrong.Honored Contributorsir here is my code,plz suggest me.
Honored ContributorHow do you know in which step of the calculation the result is set to zero? I can't guess without knowing the component code.
Honored Contributor@FvM sir, i am attaching the component.plz help.
Honored ContributorThe code has several length matching errors and doesn't compile. Also, what's the fpdiv component?
Assignments with length msimatch: p<= to_sfixed(i)*data_in; b<= (i*Pi); cv_q<=to_slv(b);Honored Contributor@ fvm sir,
ihave taken i (2 downto 0,data_in (14 bits),so in multiplication result is 16 bits. as i take i an ufixednumber,so i convert in in sfixed. now i take pi (ufixed(2 downto -7) ) so after multiplication it should be 6 downto -7,but i take b(b:ufixed(8 downto -8);) to match with the fixed point division 16 bit inputs. yep there is error,u r right,i will use correct range, other else.............. sir.
