The outgoing neutron energy spectra from neutron-induced fission of various actinides are important for basic understanding of the fission process near the scission point as well as playing a large role in neutron transport codes, which are heavily relied upon in the design of advanced nuclear reactors and simulations of critical assemblies. The reliability of the results of neutron transport models is a strong function of the quality of the nuclear data used as input. Currently, the world's experimental database of fission neutron spectra is severely incomplete (especially for higher incident neutron energies) with large uncertainties in key portions of the outgoing energy spectra. Many transport codes use evaluated data libraries, which are based on the approach of the Los Alamos model. Other theoretical models have been developed, but the available data cannot distinguish the results of different models (as is the case for 239Pu). Better measurements are needed for all incident and outgoing neutron energies, but most urgently in the low-energy (below 1 MeV) and high-energy (above 6 MeV) portions of the outgoing spectra where theoretical model results differ greatly. We present the design considerations (and some characterization results) of the two Chi-Nu neutron detector arrays: one array of 6Li-glass detectors and one array of liquid-scintillator detectors. These detector arrays are being constructed to meet the challenge of measuring the prompt fission neutron spectra (for a few common actinides) to a higher accuracy and precision than achieved previously and over a larger incident energy range than has been covered by previous experimenters. We see a significant reduction in neutron-scattering backgrounds with our new array designs.