A decade ago in this column I tried to answer this complicated question and did not do a very good job. I can now answer this question definitively: It depends.

The first thing it depends on is whether you're looking at ingoing or outgoing tolerances. Ingoing tolerances apply when you decide to move a roller. A part of a hairsbreadth (0.002 in.) level and square would be appropriate because you can usually hit that target in but a single move using precision measurement such as optical tooling. We have done this for the better part of a century and know that hairsbreadth tolerances are rarely difficult. The exceptions have to do with specific parts of machinery that may have looseness, be bolt-bound or where the alignment measurement is compromised. The exceptions allow relaxed standards if we know our best simple effort is good enough for that specific location.

Outgoing tolerances decide if you're going to move a roller, and that's where the discussion gets interesting. Just because you can hit 0.002 in. doesn't mean you need to maintain outgoing tolerances of 0.002 or 0.02 or even 0.2 in. Like everything else, our roller maintenance efforts must pay off by reducing alignment-sensitive causes of waste and delay. Sensitivity in turn depends on material (thickness, modulus, etc.) and roller (width, span, etc.) factors. We're now able to calculate how close alignment needs to be to avoid wrinkling and can make some estimates to avoid increasing the number of web breaks. These calculations are the result of decades of work at the Web Handling Research Ctr. at Oklahoma State University but are only available to their sponsors. Now, the public has access to similar tools. TopWeb can estimate the tendency to wrinkle as well as the stress riser that would exacerbate web breaks.

Something else that has a huge influence on the answer is whether the misalignment is in the parallel or skew direction. In-plane, parallel or tram misalignment is by far the worst. It is about 100X as damaging as out-of-plane, scissors or skew direction. Guides, for example, make use of large but controlled misalignment to move the web without damage. The good troubleshooter would make use of that knowledge by only moving where it counts.

Thus all things considered, it depends. For very thick and tolerant webs, we might be able to find enough with a precision level (in a velvet lined case) and precision Pi tape (with a vernier that can read 0.001 in.). However, even more important is the machinist's skill in making the measurement. They must have experience, the best tools and ample time. And they must have the web-handling and process knowledge to know what's important and what's not. For intolerant webs, such as typified by wide paper, the answer can vary enormously. On the one hand, 0.002 in. might not be close enough. Another roller on that same machine may need only be within 2.0 in.

So where does all of this leave us? We align every roller to a hairsbreadth because it can usually be done without costly downtime. It's the recurring alignment that depends. In general, paper has overdone alignment, and film is still catching up to best practices. Textiles and rubber are so hopelessly behind they can't see the people ahead of them.

920/312-8466, drroisum@aol.com, www.roisum.com