It is very sensitive to what you input. So here is what you should do until you master the concept
1-Enter your mass in consistent units. It should be that when multiplied by acceleration (m,s units or ft,lbs or whatever you used), you should get the correct force unit.
2- Enter your spectral acceleration function as it is. REMEMBER THIS IS NOT ACCELERATION. IT IS A RATIO OF HORIZONTAL TO GRAVITATIONAL. this curve accounts only seimic zone factor and site information, but no other factors and no units yet.
2-Now you must convert it to accelaration by factoring it by g, in your own units. And then further by importance factor I and may be a system factor R.
3-Thus your overall scale will be for example gxI/R, and it usually is around 2-3. Anything less is wrong.
Alternatively you can factor the spectrum function by the same value right at the beginning and then scale it to 1.00
The best way to test this one is to model a pole, put a mass on top, define a function that is 0.4 everywhere, meaning that you are accelerating the mass at 0.4g. Then run and check to see that the shear in the pole is 0.4g x mass(ie F= m. a)