This paper outlines two projects concerned with the quality testing of wire and strip metal used for the formation of springs and attempts of in-line control of a (helical) spring coiling machine for wire. The Fracmat project was a highly successful attempt to test wire for its coilability, which is known to be due to sequential variability. Methods from chaos theory, notably the Packard -Takens plot, were used to extract information that had otherwise been ignored to produce a test that could successfully distinguish between good wire and poor wire which otherwise satisfied all other available quality tolerances.