Performance assessment of the bond between FRP as externally bonded reinforcement (EBR) and concrete requires establishing bond-slip characteristics using appropriate experiments, one of which is the double shear bond test. Different versions of the said test, ranging from its traditional form to improved versions, exist. Yet, the improved versions themselves are not without shortcomings. The aim of this paper is to present a simple and improved double shear test setup, customized to ease its application in FRP-concrete bond tests in addition to satisfying other important functional requirements. Designed and fabricated as a Y-hook-cylinder system to substitute half of the traditional double shear specimen size, the simple feature of traditional approach is retained in the other half. Performance of the developed setup is, first, verified against that of the traditional approach on a number of specimens. Subsequently, the proposed system is used to conduct an exploratory study on the influence of coarse aggregate type on FRP-concrete bond behavior. Two concrete types are produced; one containing fresh aggregate and the other consisting of steel slag aggregate substituting 30% of the maximum fresh aggregate size in the first concrete type. Typical bond-slip models of the two specimen types are developed and compared with several other literature sources. The results, both in terms of the debonding load and bond-slip characteristics, show improvement in bond performance when a substrate failure occur in the specimens containing steel slag aggregate.