I don't think the .capx in the original post reproduces the problem, but I did notice a problem with the spritesheeting code in r169 which could have caused the oddly small sprite in the later screenshot. I think I've fixed it for the next build now.