Every now and then a good idea happens. The Modular head was a great idea and enabled the use of multiple types of AFP heads, ATL, ply cutting, part probing, etc. with the use of a single machine and machining cell. At the time the modular head was developed by Electroimpact circa 2004, the industry assumed (and accepted) that AFP was an unreliable process. It still isn’t as reliable as we’d like. One way of coping with this lack of reliability is to stage more than one head in the AFP cell so that a spare head of the exact same type is ready to jump into action if the head out on the floor has an issue. If the reliability of the AFP process were to increase 10x or 50x, would there still be a business case for the multiple AFP head system? The modular head may still win the day, but the metrics change. For instance, if there was only 20 minutes of down time for every head load, it may no longer be advantageous to have 2 heads of the exact same type in the cell. It is our goal to eliminate AFP process unreliability to the point where this discussion has real meaning.
To address the #1 cause of reliability issues experienced in 777x we invented the Modular-Servo-Creel head. We built a full working prototype of this machine and demonstrated it to Boeing and others over the past year. What we learned was indeed we did fix the #1 cause of reliability issues that we see in production of the 777x spar (the loss of tension during large speed changes during the zero degree ply). In the process of using this head other causes for unreliability also came into view. They actually had nothing to do with the theory of operation of the head as we previously experienced with the old creel system, but more to do with preparation. These items are:
Head Cleanliness
Blade Sharpness
A valve that is failing or leaking
A seal that is worn or leaking
A spring plate that failed
Each of these items caused an error on a part that we were trying to make and diagnosis took longer than acceptable causing even more issues on the part until the correct diagnosis and remedy was made. Because we identified these items as potential causes of mistakes on the part, we created a system described in detail in this paper. This section describes a process and method for cleaning the AFP head using a dishwasher, a method for checking blade sharpness and finally a method for checking the module functionality. We demonstrate this this system running our prototype AFP head building 16 plys of the 50’ spar, some stringer charges and then a hexagonal test part placing 100,000 individual tow strips without a process error, not even a slipped tow.
