for i=1: nsamples
ydata(i,:)=[5+10*cos(2*pi*time+pi/6)];
end
That calculates 5+10*cos(2*pi*time+pi/6) and assigns it to the first row of ydata, then calculates the same thing again and assigns it to the second row of ydata, etc., etc., for nsamples rows. At the end ydata is a matrix where each row is the same and each row contains the calculated signal for all times.
What you mean to do is simpler than that; just calculate ydata once:
deltat = 0.1;
nsamples = 2^10;
time = (0 : nsamples-1) * deltat;
ydata = 5+10*cos(2*pi*time+pi/6);
figure(1);
plot(time,ydata,'b')
% zooming in
figure
plot(time,ydata,'b')
xlim([0 10])
