Hi Cherie,
If you want to produce a .1D file with twice as many data points as your time series, you simply tell waver to use a TR (-dt I think is the flag) that is half the size of your actual TR, in your case TR=1.25 instead of TR=2.5sec.
The only AFNI program that I know of that can use .1D files that are twice as long as FMRI time series is 3dDeconvolve using the -stim_nptr flag. This option allows one to model a hemodynamic response at a higher frequency than the rate obtained in the data assuming you've designed your experiment to meet the criteria for using the -stim_nptr flag. You'd have to read up on the implementation of this to know if it will work for you.
-jim