AbstractsEngineering

The fiber manufacturing has traditionally had three different process phases: Preform Manufacturing, Fiber Draw and Proof Testing. This thesis focuses on combining draw process and proof testing, which requires catching the fiber end after break at full production speed without disturbing the draw process. Proof testing means applying a specified tensile load to continuous lengths of optical fiber. The tensile load is applied for as short time as possible, yet sufficiently long to ensure the glass experiences the proof stress. The proof test cycle is divided into three steps: load, dwell and unload. Nowadays the industry commonly accepts that the dwell time has minor effect on the final minimum strength of fiber, but the strength decreases during the proof testing cycle. The unload time is considered to be a machine property by capstan design, but new approach is suggested in this thesis, where the unload time is a material property. The two-region crack growth theory states that depending on the unload rate the crack growth may happen in two region. The effect of the coating was also studied. It was found that at high load rates the coating carries a substantially higher part of the load. Additionally theories suggest that the strength of the fiber is significantly higher just after draw than approx. 1 hour later, because of water penetration. An approach for modeling tension behavior in combined draw and proof testing process mathematically was introduced and a universal simulating tool was generated. Experiments were carried out to verify the theory. The first was the draw tension experiment. The second part was the comparison of the conventional proof tension measurement and the new method needed for the combined processes. After this the effect of the different process elements were evaluated. Two different methods to survive proof testing break were introduced and tested. A combination of tubes and belts turned to be the most reliable. Since the winding quality is important, a new winding algorithm was developed. Several trials focused on preventing whipping phenomenon. An optimal whipping guard and auxiliary blade minimized the whipping. The transfer reliability was tested with a separate take-up module. The result was that the transfer reliability is at satisfactory level at the existing production speeds. The future work will include more experimental testing, which will combine the critical components into one machine. The overall reliability depends on how well these critical components are integrated to each other.