smallputin
Newbie level 5
anyone can tell me whether my algorithm is right using Gardner algorithm for timing recovery?
the algorithm is followed as:
suppose 4 samples per symbol. substract the register value and control word both in NCO at each sample point(n(i)-w(j)).we suppose the initial value of register and control word are 0 and 0.5 respectively.the result of the substraction should be moded 1(mod((n(i)-w(j)),1).And the result of the substraction is less or equal to 0. calculate the fraction interval(2*n(i)).we can donate the fraction interval as u.now use u to interpolate.if the sample point is first or second one,the interpolated value is y
.if the sample point is third or fourth ,the interpolated value is y(n-1/2).If the substracton value (n(i)-w(j)) is more than 0 and the sample point is second or fourth,so we should interpolate value and interpolated value donated as y,y(n-1/2).After that we calculate the next symbol.we compute timing error for each symbol.update the control word,put timing error though loop filter.
the algorithm is followed as:
suppose 4 samples per symbol. substract the register value and control word both in NCO at each sample point(n(i)-w(j)).we suppose the initial value of register and control word are 0 and 0.5 respectively.the result of the substraction should be moded 1(mod((n(i)-w(j)),1).And the result of the substraction is less or equal to 0. calculate the fraction interval(2*n(i)).we can donate the fraction interval as u.now use u to interpolate.if the sample point is first or second one,the interpolated value is y
.if the sample point is third or fourth ,the interpolated value is y(n-1/2).If the substracton value (n(i)-w(j)) is more than 0 and the sample point is second or fourth,so we should interpolate value and interpolated value donated as y,y(n-1/2).After that we calculate the next symbol.we compute timing error for each symbol.update the control word,put timing error though loop filter.