Different concepts of integral experiments for fusion blanket neutronics are investigated. The first is with the neutron source (tritium target of a neutron generator) located inside of or in immediate proximity to the stack of blanket materials under consideration. The second is based on irradiation of the stack by means of a collimated and, therefore, monoenergetic T(d, n)4He neutron beam with a tritium target placed outside the stack. The comparison between the different concepts is carried out by means of the Monte Carlo transport code MCNP with continuous energy treatment. The comparison between the two approaches reveals that the integral experiments with a collimated monoenergetic T(d,n)4He neutron beam result in a neutron spectrum that is better correlated with the details of elastic and inelastic scattering to the first level of the material's nuclei than the one with a neutron source inside a stack. In the case of a collimated neutron beam, there is a clearer separation between energy regions of different neutron interactions and, therefore, the source of discrepancies between measurement and calculation can be identified more easily and corrected by a proper treatment of the cross sections of the specified nuclear reactions.