Abstract Argon–argon dating (a variation of potassium–argon dating) of ordinary chondrites is being used to reconstruct the collisional impact history of their parent bodies. However, due to the fine-grained, multi-mineral, highly shocked nature of chondrites, the sources of potassium (K) in these meteorites have not been fully identified. By locating and isolating the different sources prior to analysis, better ages can be obtained. To distinguish between possible sources, we have analyzed Chico and Northwest Africa 091 (both L6 chondrites) via K mass balance, Raman spectroscopy, and argon (Ar) diffusion studies. In accordance with previous studies on other ordinary chondrites, the Ar in these two chondrites is nearly equally split between two releases, and the lower temperature release is identified as sodium-rich feldspar. Various scenarios for the higher temperature release are investigated, but no scenario meets all the required criteria. The Ar activation energy of the higher temperature release is the same as pyroxene, but the pyroxene has no detectable K. The K mass balance shows feldspar can account for all the K in the chondrite; hence feldspar must be the ultimate source of the higher temperature release. Raman spectroscopy rules out a high-pressure phase of feldspar. Neither melt veins, nor feldspar inclusions in pyroxene, are abundant enough to account for the higher temperature release in these meteorites.