Although the direct influences of sunspots upon the solar constant leads to short-term decreases, an opposite nearly in phase 11 year variation in the solar constant can be modelled. This opposite behavior results primarily from global faculae (polar, network, and active region). The model attributes the observed variations in the solar constant entirely to magnetic features in the solar atmosphere. Hence, the model does not require any indirect effect of the magnetic cycle on the brightness of non-magnetic areas of the Sun. The model uses the well studied, active region influences upon the solar constant, together with the influences of global faculae, associated with polar and general solar magnetism, studied by Sheeley. Using the present model, together with the sunspot prediction of Schatten and Sofia (1987), we predict the average solar constant to the year 1997, normalized to the mean Active Cavity Radiometer (ACRIM) and Earth Radiation Budget (ERB) observations. The model also provides a relation ΔS0=0.006 ΔRm max, for the very long term solar constant variations due to the secular changes in solar activity. We obtain a change of ~1 W/m2 for the difference between the late twentieth century solar constant and the 17th century solar constant. This supports Eddy's view that this difference could give rise to the glacial increase during the little ice age of the 17th century.