Words of Wisdom: Pushing the envelopeWords of Wisdom: Pushing the envelope
July 3, 2003
Bill Tobin is a plastics processing consultant and president of WJT Assoc.Anne Bernhardt is president of Plastics & Computer Inc., process simulation specialists. |
In the 1950s test pilots coined the phrase ?pushing the envelope? to compare calculated performance limits to the actual test results in supersonic aircraft. Test pilots could do this because the science of aeronautics and space travel was not well defined. It still isn?t. There are still more exceptions than there are rules in aerospace. For example, according to traditional aerodynamics, the bumblebee can?t fly. Too bad nobody ever told the bee. As a catchy phrase, ?pushing the envelope? is now widely applied because, in the case of part design or material selection, the alternative description ?extreme? could result in a poor performance review and salary adjustment.
I (Bill) learned a lesson when working in the medical industry. I was merrily thinning wall stocks, making faster solvent welds, and pushing the envelope on statistical testing. It was in the ?70s when profit was the only goal of corporate behavior. My achievements came home to roost when I watched my wife being wheeled into surgery for an operation while hooked up to equipment that contained all my engineering changes. From then on, I concentrated my profit improvement energies on productivity, not product design.
Every plastics part designer knows that sharp intersections and corners increase stress. Basic engineering teaches that strength (rigidity) is a cube function of wall thickness. Every molder will tell you that molded-in stress is normal. But when the outside of the mold is sweating like a marathon runner, or the machine?s hydraulics are working at maximum capacity to push cool material into a cold mold in the quest of a few seconds cycle-time reduction, something is wrong.
Here are some examples of counterproductive efforts to push the envelope.
Thinner Wall Stocks
Thin implies a lack of rigidity and an increase in flexibility. These are good qualities, but flexibility must be tempered with a place to flex. If flexing is at a sharp corner, weld line, or abrupt transition, microfractures can quickly form. Continued flexing can cause the section to break.
Thin implies shorter cooling times. True, but you have to fill the cavity first. If the wall stock is thin, it freezes off, quickly impeding the flow of material to the rest of the part. Designers don?t care if the molder is smashing the material in, but this can result in high molded-in stresses that both warp the part and make it susceptible to premature failures. Extensive real-life testing is the only way to know you haven?t broken the envelope.
Bad Wall Stock Transitions/Small Radii
Abrupt changes in wall stock, nonuniform wall stock, and small radii anywhere on a part are a formula for disaster. However, a CAD program can?t see this. In many designs, a CAD designer leaves an area on a design with a print note such as ?blend to suit? or ?break all sharp edges? simply because he can?t or won?t finish the design himself. Since designers are now going straight from an electronic design to a cutter path, many of the notes are simply overlooked.
The flow analysis of a large industrial light housing helped identify and solve the root cause of an in-service cracking problem. The left image is the filling isochrones; on the right, the shear stress distribution at the end of fill. The high shear stress areas (in red) corresponded with the failure locations and the last areas to fill. Profiling the injection reduced the shear stress at the end of fill and solved the in-service problems. |
Supplier Information
Every engineering course teaches students to take the published data and divide by a substantial amount to build in a safety margin. Most people don?t understand a resin company?s data. Resin companies are the world?s best makers of test bars under ideal conditions in which dimensions, functionality, and cosmetics are not issues. Supplier data comes from tests under ideal conditions. Naïve processors commonly come back to the resin suppliers and whine that their tensile, impact, and other properties are never close to the published data. Why? Commercial molding is never done under ideal conditions.
Exotic Materials
Too many designers create part designs first, then go looking for a material that fulfills both the structural and functional requirements of the part. This leads to pushing the envelope in material specifications. While the $50/lb-materials usually work, their price is prohibitive. The more conventional materials such as ABS or PC, in any of their dozens of grades, alloys, or blends, are usually only marginally capable of meeting the physical specifications.
In some cases, poor design and pushing the envelope are of small concern. Look at the plastic CD case under polarized light. The rainbows from stress look better than a Grateful Dead tie-dyed T-shirt. Warp or breakage means nothing in this application. Have you ever heard of anyone taking it back for a new case because the hinge strap broke off? Look at VHS cassettes. Leave one on the dashboard in the middle of summer before returning it, and you?ll see it warped like ribbon candy. This is why the video store charges you an extra $.50 for insurance. Using materials that are only marginal for the application can work in some circumstances, as long as the upshot of product failures is inconsequential.
Believing Your Simulations
Some simulation programs for mold filling and cooling are based on highly false assumptions. The first is that the rheology developed for your inputs correctly reflects your material in a production situation. Just because the band and grade you plan to use are in the database it does not mean that the data are correct, and that they reflect the material that you use in actual production. The calculations are based on rheological and thermal equations. Because most designers won?t pay for custom material testing, generic rheological data are frequently used. Not all ABS is the same. Different grades have different properties. The material databases are based on testing one lot of material one time. There can be errors in the testing or a bad lot of material, and the data don?t reflect the effects of adding additives and regrind.
I (Anne) once had a client who was convinced his molder was incompetent because the mold wouldn?t fill, even though the simulation said it would. I looked at his simulation pictures and graphs. The only problem I could see was the injection pressures in his simulation were 80,000 psi at fill. I tried to explain that at these pressures, even assuming the machine could deliver them, the screw would have broken or something would have exploded. His explanation was that he couldn?t get the mold to fill in his simulation until he raised the pressures to that level. Oh well. At least the pictures were pretty.
Mold simulations reflect a perfect world in which the part design is perfectly cut into the steel; ejector pins don?t get in the way of cooling the ejector side; gates, runners, and cooling lines are all the exact size; and the steel alloy used has known and true thermal properties. Unfortunately, we don?t live in a perfect world. Go to your production source and find out what the real melt temperature, mold temperature, coolant temperature, and cooling flow rates are before you waste your time simulating a mold that already exists.Simulations are good but they require two things:
Perform a sanity check. Before and after doing a simulation, look at your results and ask a few questions. Are your inputs and the program?s output reasonable? If not, why?
Understand that a simulation gives you the luxury of simulating. In multicavity tools, first simulate filling the cavities to find out what conditions are required to get good parts. Then add your runner systems to make sure that you are delivering the melt under the proper conditions to each cavity. This is an eye-opening experience.
In same-part molds (not family molds), delivering the material at the same pressure and temperature fills all cavities in a uniform manner. You?d be surprised at the differences you get by moving the gate location, using a valve gate with a comparatively huge orifice, simulating use of high-conductivity metals (BeCu alloys) in areas that are slow to cool, and lower conductivity metals (stainless steel) in areas that cool too quickly. Molds are supposed to be a one-time expense, but a process mistake or mold design mistake makes you pay throughout the entire product life. Is it cheaper to get it right now or later?
Even without knowing the derivation of ?pushing the envelope,? the phrase creates an interesting visual?yet none more hilarious than ?extreme designing.? Experience has taught us a lot about product design and materials selection, but we?ve downsized. Now computer jocks in front of their flat panel displays are called designers. The mentors who learned from making every possible design mistake in their 20 to 30 years of experience have been laid off. In-house, free consulting is gone.
If you?re not doing a repeat of an existing design, go find an expert. Build on the expert?s knowledge, and you?ll make excellent products at economical prices. Using an expert is always cheaper than making engineering changes to correct mistakes. Benjamin Franklin said, ?Make haste slowly.? It is better to do your homework, make minimal changes, and bring a product to market with high confidence than to push the envelope, inflict a product on the marketplace, and hope for the best.
You May Also Like