In theory, the proton shift depends on the overall deformation of the magnetic field that is caused by the molecule structure and the solvent interactions, as well as interaction with impurities and which are temperature dependent. Thus, the proton shift is the same, if being pure and taken in the same but pure solvent and the same temeperature; regardless the functional frequency of the instrument because the frequency affects the peak resolution and not the proton shift. Consequently, product/solvent impurities may be responsible for slight variations of the shift from sample to sample.
However, modern instruments lock the zero shift by software driven electronic devices, which are sensible to errors due to electric current fluctuations or magnetic interactions from e.g. a moving metallic carriage that is passing nearby. Thus, very slight variations of proton shift from instrument to instrument must also be expected.
PS: It is obvious that slight variations of the ternary protein structure, may cause significant variations of the peak shift and appearance due to long distance, intramolecular interactions.