This is the spectrum coming off of the sample. there should be a peak at 1038Hz... but there isn't. And what is even weirder is that the spectrum analyzer that is built into Simulink shows a peak where I expect but when I do it here it doesn't show up. 

Update: I think I have found why there is a discrepancy between the two versions of the power spectrum. The spectrum analyzer in the Simulink model requires non-continuous data so you have to use a block to make your data discrete. The sampling rate of that block affects where the peak of the mode is seen. So it seems that that the mode seen in the previous post was just caused by making the data non-continuous.