Multipactor in WR-284-like geometries is measured utilizing local and global detection techniques. To emulate conditions one may find in a waveguide filter structure while maintaining the fundamental microwave mode, a standard rectangular waveguide geometry with the reduced waveguide height set to 2.1 or 5.5 mm was adopted. Two high power RF sources were used to investigate a large range of input power (few kWs to MWs): a solid state source using GaN HEMTs allowing for larger pulse widths than standard magnetrons (100 μ s as opposed to ∼4 μ s) and a MW level S-band coaxial magnetron for the high power end. Particular interest was taken in capturing the lower and upper limits of multipactor threshold. Lower multipactor thresholds for finite pulse duration are governed by the appearance of one or more electrons in the multipactor gap during the applied pulse as well as a minimum power (electric field) level that affects a secondary electron emission yield above unity. As shown, such initial electrons(s) may easily be seeded via an external UV source illuminating the gap. However, wall collisions of excited metastable molecules may be another source of electrons, an observation based on the experiment and prior research. A multipactor upper threshold was non-existent in the experiment, even at powers over 200 kW within a 2.1 mm test gap, which numerically yielded a gap transit time significantly shorter than one half-period of the GHz wave. This is attributed to the electric field distribution within the waveguide structure, which results in the multipactor's spatial position moving to more favorable locations within the test gap.