Recall from the previous post, that effective culling is economically better for both customer and supplier than are customer complaints. Start culling now rather than wait for some magic bullet cure so the benefits are achieved sooner rather than later. Culling can also help with possibly finding a cure. Because you frequently separate questionable materials you might in the process notice (observational or statistical) correlations to events on your raw materials or machines.
However, there are two serious limitations to this approach. The first is that roll hardness might not be practical or predictive for your material. The second is even more universal: it is a bit late. By time you measure a problem on the roll, you are at least two roll time cycles from making a remedial move (presuming you even knew what action to take on the raw material or machine). That kind of lag between measurement and correction is not desirable, (even though much better than the lag of months between manufacturing and customer plaints).
What about (basis weight, thickness, etc ) scanners you say? Weren’t they intended to prevent the profile troubles that we are using the delta of roll hardness to pick up? Yes, perhaps, but my experience and your experience clearly show that this is not always so effective. Bagginess and poor roll appearance are almost as common on machines with scanners and automatic gage profile than without, especially in the more challenging applications such as thin film. Furthermore, some of the differences might be attributed to non-profile control reasons, such as better materials, mechanical designs and maintenance on newer machines. The reason for possible poor performance for scanners is simple. It is not a matter of averaging, alignment of measurement and control or any other of a variety of excuses. Rather, the caliper profile variation that could ruin a wound roll (and thus web) is on the order of 1-10% for most applications. Thus, in order to ‘fix’ the problem, the scanner must at least be able to reliably resolve something like a couple of percent or even tenths of a percent of thickness. So, given a 1 mil product, the scanner would need to be able to reliably ‘see’ 0.01 mils and then make appropriate corrections; a pretty ambitious goal given what is available in today’s equipment.
You may not accept the above reasoning, so let me give you something much more pragmatic. If scanners and automatic profile controls were so good, they should be able to flag areas and label them as having the potential for bagginess or corrugations, ridges, valleys etc because we have long known exactly the type of caliper profile pattern poses risk for each of these. When have you ever seen defect labels like that on scanners? Yet, with optical defects, a much younger field, we do just that. We can label a variety of defects with decent reliability (especially in paper where they can afford the high end systems). Defects such as cuts, holes (with various sub cases), spots (including sub cases such as bugs), streaks (with, you guessed it, subclasses) and many others. Incidentally, defect positions can be marked in 2D and transmitted digitally to the next winder so that it automatically stops at the defect (to the nearest couple of wraps) for inspection or culling. I am aware of nothing even resembling this helpfulness with thickness scanners, even in the paper industry that tends to lead with digital goodies.
Am I concluding that in general scanners are useless? Of course not, though in more than a few cases they can be shown to be little more useful than random number generators. Am I concluding that in general automatic profile controls are useless? Of course not, though I am aware of several cases where the controls are destabilizing. Also, even in the more moneyed and mature paper industry, the ends of the machine are usually under manual control as operators can out-perform these otherwise digital wonders. So, much as we might wish otherwise, sometimes the wound roll is the best measure of gage variation. Those of you interested in the subject might look at my “Secrets of a Level Process and Product” that you can download from my website.