Beruflich Dokumente
Kultur Dokumente
Four common approaches to troubleshooting press-shop problems are 1) ignore the problem, 2)
emotionally order random corrective actions, 3) blame the metal supplier and 4) undertake a
logical data-based analysis of the problem. This true case study incorporates features of all four
approaches. The higher-strength-steel problem part is a deep-drawn, closed-end channel with
welded flanges. Separate dies in separate presses form symmetrical left- and right-hand
channels. Severe sidewall scoring in the right-hand channel is a perpetual problem that no longer
can be ignored.
Attempts to eliminate the scoring have focused on pursuing a long list of traditional modifications
such as polishing the die, changing punch-die clearance and playing with different lubricants.
One could characterize this approach as emotionally ordering random corrective actions. These
attempts to correct problems almost always begin without a valid definition of the problem or a
final goal.
One day, two different coils of steel produced right-hand parts with different amounts of scoring.
Now the problem is obvious, the steel is the root cause. The steel supplier receives a call to
attend a demonstration at the stamping plant. First, a section of coil B (for bad) produces parts.
Then a section of coil G (for good) produces parts. Yes, the G-coil parts visually show somewhat
less scoring than the B-coil parts. In the traditional way of doing business, the representative
from the steel supplier returns to his company with a mandate to supply better steel. Blaming the
supplier is completed.
However, all participants have ignored the rules of good troubleshooting. First, a numerical
definition of the problem is required to provide a base line or reference point. Relating back to
this starting condition allows one to track progress toward the final goal or problem solution. The
final goal also must be a numerical and realistic target. A goal of zero defects forever is not
realistic. It looks, I think, I know, I assume, I feel, its probably and I am comfortable
with are not numerical statements of either the problem or final goal.
How does one numerically define scoring of a part? In this problem, the area of the rectangle
encompassing each distinct patch of scoring is measured. No scoring for 0.25 in. means that one
patch has ended and another patch has started. The total area of all patches is the scoring index
for that part. Measurement of multiple parts not only provides an average scoring index, but also
a measure of repeatability.
The graph shows the results from this study. The results for the right-hand part (1) show two
distinct lines confirming that coil G produces
Source: http://www.metalformingmagazine.com/magazine/article.asp?iid=40&aid=4986