It is well known that a telescope with an unobstructed circular pupil delivers a smaller diffraction pattern than one centrally obstructed by its secondary mirror. Spaceborne dark energy investigations require measuring targets over a wide range of redshifts, with the most distant galaxies being the reddest, faintest, and smallest. For any given signal-to-noise (SNR) requirement, these highest redshift targets are the most demanding in terms of mission cost (time, aperture, etc), not only because they are faint but also because the diffraction pattern is largest at the longest wavelengths being observed. At the same time, a telescope's field of view must be large --the order of a square degree --to survey the entire extragalactic sky in reasonable time. The large field of view imposes a minimum requirement on the size of the secondary mirror baffle. For a centrally obstructed telescope, an enlarged secondary mirror baffle further enlarges the diffraction pattern. Previously published JDEM telescopes were centrally obstructed. Here, we explore unobstructed telescope designs because these can have a nearly ideal Airy diffraction pattern, avoiding both the central obstruction and the supporting spider legs, limited only by optical manufacturing and alignment errors. They therefore can deliver the best possible SNR for a given aperture. Simulations show that a 1.1m unobstructed aperture can deliver about the same cosmological constraints as a 1.4m aperture that has a 50% linear central obstruction.