The Project:
Our decision to take on this project influenced our entire business strategy, including the timing of our move, the selection of projects we would take on, and how we grew.
The Cliff Notes:
This part fits into an apparatus that processes blood from post-orthopedic surgery patients. The machine removes debris and clots, and allows the patient to self-administer anti-coagulant. Every single dimension of this part is critical, because it fits into a transfusion device with very small tolerances. Most people, it turns out, think it’s a bad idea to have post-op patients’ blood leaking out of the transfusion machine.
The Challenge:
Where do we start? First, of course, we had to acquire a machine capable of making the part. The material, PETG, is a difficult and finicky substance. It’s extremely hard, which makes it a royal pain to trim without cracking.
Each port must be the exact correct size, which is challenging because not all of them are created by blown air. The bottom port is trimmed off by, for lack of a better metaphor, an industrial-strength guillotine. The location of the cut affects both the inside and outside diameter of the hole, so we’ve got to be crazily precise here. Also, the guillotine is liable to crack the whole bottle if it’s not done just so (darn that PETG!).
If you rotate the part 360 degrees, you’ll see the parting line (the “seam” of the molding) zig and zag a couple of times. This “jawed” parting line creates complexity. As the pins and bushings that hold the the mold in place wear, misalignments can occur. This boggled us for quite a while; we’d get the thing going perfectly and then normal wear and tear would degrade the quality in unpredictable ways.
Luckily, our engineering manager came from an injection molding background, and innovated a trick (blow molding competitors, pay close attention here) using mold locks, in addition to pins and bushings, to hold the mold in the correct position and prevent misalignments. Since that discovery, we’ve added mold locks to several other projects with alignment issues.
As if all that weren’t enough, the part has to be completely smooth, with no jagged edges that could tear something or cut someone. So we had to add the step of blow-torching each part as it came off the belt to melt away any sharpness.
Finally, we discovered that sections of our all-aluminum mold were wearing out way too quickly. We switched to beryllium copper, but that wasn’t strong enough. We switched to stainless steel, but that wasn’t (can you guess?) strong enough. We finally settled on S7 tool steel for our mold inserts. This is the stuff they use for forcefully cutting other metals, in applications like industrial chisels, punches and shear blades. This is good stuff!
The Result:
We can now make 10,000 units per week-long run, going nonstop night and day (we try very hard to avoid shut-down and startup once a run is in progress). A successful run achieves 90% productivity or better.
Ongoing Quality Control:
Again, where do we start? First, this job forced us to bring quality inspections right to the machine on a constant basis. With most jobs, we pull a few units off the line every hour and fiddle with them in the quality control area. With this part, every single bottle gets checked with a port gauge, a metal plug that measures inside and outside diameters of the holes.
We use our good old magna-mic (magnetic micrometer) to check wall thickness. This customer has a minimum thickness requirement, and we have two reasons to keep close to this number: too thick wastes material, which drives up our costs, and too thick increases the likeliness of cracking by making it harder to cut.
We visually check each bottle for contamination, since it needs to be completely translucent and of a medical grade approved material for handling blood. Nobody wants to add little bits of loose contaminant to their blood as they get it filtered.
The big picture of quality control is the interesting part (to us, anyway). We engage an outside Six Sigma Black Belt consultant to help us keep track of the thousands of variables that can mess up a run. (You should meet him. He labels his eyeglass case, and labels the spot in his briefcase where the eyeglass case goes. He’s the only person who ever makes us feel a little sloppy.)
We’ve been continually improving the Six Sigma Black Belt process over the past seven years. We have pre- and post-run meetings every time we run this job. If you were to attend the pre-run meeting, you’d meet our Black Belt consultant, the entire engineering team, the production manager, the quality manager, shop floor workers, and supervisors. You’d see us poring over data and statistical analyses from our last few runs, with a focus on any variations that could affect quality and consistency. We look at the recent maintenance history of the machine, and try to anticipate what might have changed due to the maintenance. We complete a pre-run “fire drill.”
The post-run meeting can be fun if we’ve met our goal for that run. If something went wrong, we roll up our sleeves and troubleshoot for up to several hours, until we’ve solved the problem.
This part challenges our problem solving abilities like nothing else. It drove quality awareness to a completely new level in the organization. Everyone who works on the this job needs specialized training. They meet with the quality manager, study a huge sampling of what can go wrong (kind of like Edison’s 10,000 ways NOT to invent the light bulb), learn how to document the statistics we’ll be analyzing after the run, and – very importantly – how to use the airhorn. (When the airhorn sounds, everything stops and we go into very intense “What went wrong?” mode.)
It’s no understatement to say that this project has inserted quality processes, not just into the technical aspects of blow molding, but into our very DNA. Quality is now our philosophy of management and planning. We use LEAN and Six Sigma to map out company wide projects. Heck, there are probably people here who use Six Sigma to get their kids ready for school in the morning.