A multipurpose instrument for the measurement of reflectance and transmittance versus angle of incidence for both specular and diffuse samples in the solar wavelength range has been constructed and evaluated. The instrument operates in the single-beam mode and uses a common light source for three experimental setups. Two integrating spheres, 20 cm in diameter, are used for diffuse transmittance and reflectance measurements. The transmittance sphere can be turned around an axis through the sample to vary the angle of incidence. The reflectance sphere uses a center mounted sample and a special feature is the position of the detector, which is mounted on the sample holder at the center of the sphere. This way the detector always sees the same part of the sphere wall and no light can reach the detector directly from the sample. The third setup is an absolute instrument for specular samples. It uses a small averaging sphere as a detector. The detector is mounted on an arm which rotates around the center of the sample, and it can thus pick up both the reflected and transmitted beams including all multiply reflected components. The averaging sphere detector is insensitive to small side shifts of the detected beams and no multiple reflections between detector and optical system occur. In this report a number of calibration procedures are presented for the three experimental setups and models for the calculation of correct transmittance and reflectance values from measured data are presented. It is shown that for integrating sphere measurements, the geometry of the sphere and the diffusivity of the sample as well as the sphere wall reflectance and port losses are important factors that influence the result. For the center mounted configuration these factors are particularly important and special emphasis is given to the evaluation of the reflectance sphere model. All three instrument setups are calibrated using certified reference materials and nonscattering mirrors and substrates. The results are also compared to the results of a double-beam Beckman integrating sphere for near normal angles of incidence and Fresnel calculations. The results in this article show that good agreement is obtained between results from the different instruments if, and only if, proper evaluation procedures are applied to the measured signals.