In last month's column, we talked about machine quality. Machine reliability is more difficult to define in practice. This is because in normal operation, you can usually only talk about system reliability. Let's illustrate the concept using the example of web breaks. The reader can easily adapt this example to their own process.

Just because a web breaks on a machine does not mean that you can point the finger at the machine. It is not the machine that is breaking. It is the web. One might say there is nothing wrong with the machine if you run material that is not so ill behaved, such as carpeting. This may not be comforting to those who do not make carpeting, but the concept is still sound. Similarly, one can say there is nothing wrong with the brittle, cracked paper if you run it on a machine that did not load the web so brutishly. This may not be comforting to those who do not know how to process their web on a conveyor belt, but the concept is still sound.

Thus, web breaks are always a system issue, not a material or a machine issue—as the simple-minded manager might want to conclude. Only when you run that material on that machine do you find that they don't get along together. It does no good to point to another machine and say "that one doesn't have so many problems." That just reinforces what we already knew—machine details are involved. It does no good to point to another material and say, "that grade doesn't break so often." That just reinforces what we already knew—material details are involved.

What we must do is list all aspects of machine and material that bear on the issue of runnability. Only after both lists have been thoroughly constructed do we enter the next step: decision making.

This system nature makes it difficult to quantify "fitness for use" when specifying machinery. Machine builders seldom guarantee that the machine will run even well-known grades without defect. If the raw material is troubled enough, such as bagginess, even our best machines may turn the baggy lane into wrinkles. Builders may be more willing to guarantee "minimal defect" performance if they partner to share profits above that already guaranteed by the purchase.

So how do you protect yourself for reliability? What you should do is to specify reliability in terms of machinery only, not material or process. You could require that the machine run without mechanical breakdown or without controls fault for a period of time. An example would be to run 200 consecutive hours without a mechanical or electrical fault (unrelated to webs). The builder can have as many attempts as they wish to get the machine to that level. However, the clock restarts every time the machine shuts down, where the shutdown is not related to the web or operator. Merely show them the broken part and you have all the evidence you need. Merely show them the drive fault not triggered by the web (such as a web break detector) and you have all the evidence you need.

On flying splice unwinds and winders with turnups, we use a slightly different approach. Instead of time, we use turnups. Turnups need to be 99.x percent reliable where the x depends on your tolerance for risk and the cost for downtime. Thus, we might require 200 consecutive turnups (99.5 percent) without a miss. Here we would need to make sure the operator correctly prepped the roll. We cannot blame the equipment if the operator puts the tape on incorrectly.

Caveat emptor—let the buyer beware. Some machinery does not conform to good design (or maintenance) practice. To protect yourself, you merely need to write quantitative minimum time to failure standards such as these into the purchasing specs and include a last payment penalty clause as an incentive.

920/725-7671, DRroisum@aol.com, www.roisum.com