The feasibility of low-energy fragmentation experiments using a magnetic spectrometer is discussed. The main challenge is the multiplicity of the ionic charge states, which can hamper the identification in both Z and A of the fragments. Three topics are covered. First, a specific set-up for ionisation chambers, based on a very large gas thickness, is presented. Its satisfactory performances are discussed in light of the observations during a 500A MeV Pb+p experiment performed at the FRS (GSI). As a second topic, the possibility to use a thick layer of matter (a degrader) as a passive measurement device to identify the nuclear charge and the ionic charge state of fragments is discussed. This method, successfully used for Z identification in experiments such as Pb+p at 1A GeV, fails to measure the charge states at 500A MeV for the same system. It is shown that surface defects of the degrader are probably responsible for this failure. The third topic is the description of new analysis techniques developed in order to account for and subtract the contribution of polluting charge states in the spectrometer, thus making possible a clean estimation of the production cross-sections of all fragments. The combination of those new experimental and analysis techniques made the 500A MeV spallation experiment a success.