The AMS experiment has recently measured the proton and helium spectra in cosmic rays (CRs) in the GeV-TeV energy region. The two spectra are found to progressively harden at rigidity R = pc/Z >\,200 GV, while the p/He ratio is found to fall off steadily as p/He\sim\,R^{-0.08}. The p/He decrease is often interpreted in terms of particle-dependent acceleration, which is in contrast with the universal nature of DSA mechanisms. A different explanation is that the p-He anomaly reflects a flux transition between two components: a sub-TeV flux component (L) provided by hydrogen-rich supernova remnants with soft acceleration spectra, and a multi-TeV component (G) injected by younger sources with amplified magnetic fields and hard spectra. In this scenario the universality of particle acceleration is not violated because both sources provide composition-blind injection spectra. The present work is aimed at testing this model using the low-energy CR flux which is expected to be L-dominated. However, at E\sim\,0.5-10 GeV, the fluxes are affected by energy losses and solar modulation for which a proper modeling is required. To set the properties of the L-source, I have used the Voyager-1 data collected in the interstellar space. To compare my calculations with the AMS data, I have performed a determination of the force-field modulation parameter using neutron monitor measurements. I will show that the recent p-He data reported by AMS and Voyager-1 are in good agreement with the predictions of such a scenario, supporting the hypothesis that CRs are injected in the Galaxy by universal, composition-blind accelerators. At energies below \sim\,0.5 GeV/n, however, the model is found to underpredict the data collected by PAMELA from 2006 to 2010. This discrepancy is found to increase with increasing solar activity, reflecting an expected breakdown of the force-field approximation.