FRF by abaqus vs FRF by experiment

Dear All,
 I have been trying to obtain the FRF of a cantilever beam using abaqus. My intention is to compare the FRF pIots that are obtained experimentally and the ones obtained in abaqus. I  have plotted displacement vs frequency and acceleration vs frequency plots. The graphs look fine and I am able get relatively good estimates for the natural frequencies but the problem is the units of the y axis of the graphs when compared to my experimental data.  
 

 The FRF is supposed to be the ratio of the output to input. I used a hammer for the impact and an accelerometer for measuring the response in my experiment. Therefore, the unit of the y axis of the FRF plot should be  m/s^2/N when the FRF in the experiment is plotted. However, I am not entirely sure what unit abaqus provides the FRF in. I am assuming I get m/s^2 for acceleration and as the force is 1N I am supposed to be having the same unit as m/s^2/N. Am I right here?
 

 How can one explain this difference of FRF in magnitude ? 
 

 Kind regards,
 Chan,