This is the second installment in our 2012 Demo Part story. We began the series in Part One by introducing my new role, which supports scripting with our industrial products, and we quickly framed the story behind our new demo part. Today we will look at the Demo Part project in greater detail, follow the transformation of the part's design, and share with you the indispensable role Protomold played in bringing this part (actually parts) to injection mold, through production, and on to market
So where were we? I believe I left off with likening our situation to that of a fire drill...actually, as pointed out by our COO and in-office Sonic the Hedgehog, Tom Kurke (and by"Sonic" I mean tear-assing through the office no really, expeditious by nature, unstoppable work ethic, legal degree, and hidden arsenal of assorted Nerf weaponry yes Tom, we know about them...he's scary fast I mean fleet, the man is one nimble bas ). But I digress. The point is my career is at stake with this next sentence, or perhaps the last. He said that this was not a fire drill, and he's right, it was not a drill it was a full-fledged #@$! four-alarm fire. But to him? Biz as usual. I'm just saying tear-assing if that's a word.
We had six weeks to create and deliver parts around the world, starting with a wealth of purpose, murky design, and a CAD model detailed for rapid prototyping. Not to mention practical budgetary constraints and no roadmap to get us there. However, we had promised that our long-overdue demo part would be included in our sales kit this year, along with a thumb drive replete with the Geomagic industrial products' trial-ware and the part's CAD model just in time for the 2012 release. "Not a drill" indeed.
Why do we need one? How do you go about getting one? Others have them. This fact alone gives rise to the question;"Where's ours?" We had been threatening to"get" one for years.
Perhaps we need to begin with;"What exactly IS a demo part?" In short, a demo part (in our case) is simply a physical object that is used to demonstrate our software in concert with our sales-channel partner's scanning and probing devices. Primarily to be used as a test subject in our canned sales pitches and training. In broader scope, a demo part should also, by appearance alone, extend or represent the business. It should reflect who we are and effectively demonstrate what we do.
According to Kevin Scofield, our Industrial Product Manager and human Geomagic industrial product encyclopedia, we needed something small and light so it packs easily, travels well, and looks cool, but heavy enough to feel"substantial," and it probably should be scan-able...(breathe)...and it needed to withstand the repeated over-clamping by cheater-bar carrying tech's, being dropped from unknown heights, and stand up to the caustic chemicals most assuredly in its future. Most importantly, it needed to incorporate features that can be used to demonstrate the probing metrology side of our business when teamed with our hardware partners....and can we get it in Geomagic blue? Again, we had a wealth of purpose.
It all started, or at least gained traction like all good ideas, with sweat equity and necessity. Nick Chwalek and Gregory George, venerated Regional Managers and our customer-facing tip-of-the-sword, were not satisfied with all the talk, and like all our staff that pound-the-pavement, they have skills beyond a pleasant demeanor, steely resolve, and tactical command of our products, partner hardware, and industrial processes. They're techno-geeks. Especially Nick. Do you know anyone else that can control their RC truck with Twitter or that integrated their laptop with their Honda Element? I rest my case.
Anyway, the story goes these two guys got a new toy from our partnership with MIT's FabLab event at the American Tobacco Campus in Durham. Ping (also a techno-geek...but with an artistic side) had borrowed a VFlash printer from 3D Systems and asked us to"do something cool" with it. Nick and Greg didn't waste any time and were probably trying to hook it up to Facebook or something when they had that proverbial eureka moment. "We can make a demo part!"
SolidWorks in hand, Nick dove in and created our first official CAD model and called the team together to bounce ideas around. Nick, Greg, Mike Blazek - our professional services guru - (he provides industrial professional services I haven't seen anyone better in my 30 plus years well, perhaps Lu McCarty I'll let them fight over who's better at what), Alex Yefimov, our in house Geomagic Qualify 2012 expert and Senior QA, and John Lockhart (our recently departed Qualify Product Manager), and of course Kevin Scofield, all being rather busy, had little time to mash something together. We are a software company after all, but they somehow found the time.
They came up with the wish list and created what is affectionately known as, frankly I don't know what they called it but it certainly had merits, all those nooks and crannies were engineered with a specific GD&T callout in mind and a prototype is a damn sight better than a napkin sketch. It's functional.
When they did present the idea to upper management it gained momentum. Having a part in hand seemed to make the difference. Tom"the Sonic" Kurke, Esq. gave the go-ahead and asked the tough questions: How many do we need? How much will this cost? And whatever we do it needs to be ready for 2012 release in November.
Initially the part count was low; I hear it was hovering around ten, twenty five, or fifty. It depended on who you talked to, but generally rapid prototyping or machining was the production method of choice. High-polish aluminum would be sexy but could it be scanned without treatment no. Perhaps we could use a ball-peen, sandblast, or a brushed finish? Perhaps but the quotes coming back were long in delivery and a bit pricey, especially since the quantity was starting to grow.
What? Oh yeah. Marketing gets involved about now. Marketing always seems to have the knack to find useful things to do with well everything including my time. "We could use the part in our sales package! We need five hundred. Thanks. Make it so."
So now, rapid prototyping is out of the picture, and machining as well. Rapid prototyping would not possess the industrial strength and life expectancy we needed. Both would be too expensive and would require a great deal of post-processing to achieve the look and feel we desired, which would also lengthen our schedule out of range. I think it was just a natural gravitation to injection mold.
We had a CAD model and now we had Tom Simon. Net savvy and knowing how to type (really he does more than that) he found a website that promised injection molded parts from a CAD model in one day. Simple. He uploaded the model to the site and got back not only a quote, but a litany of issues with the design as well. One hundred percent expedite surcharge?! For injection molded parts in one day? Yeah I'd charge that too, in fact I thought it was cheap, or maybe outright impossible.
I didn't hear from Tom, but I did get an email from Kevin Scofield.
"Richard, you've worked in injection molding right? Can you look at something?"
"Yeah sure" < and what's this got to do with testing our code six weeks from release?>
Long story short, we needed to redesign the part for injection mold, it's what I did for many years, and the hard working QA team, graciously led by a closet F1 and Go-kart Drift champion, our own Arun"Big Chief" Kunjur, gave the OK and cut me loose for a few hours. I still think he and the team need to be properly thanked. So for my small part in this thank you again!
I have to say my list of concerns about the part's design matched the list from this web-o-sphere portal, nearly word for word. The part was too thick, had no split, no draft, no ribs, and no fillets, there would be wells, sinks marks, and some severe warping. Who was Protomold? Their website actually analyses the model and warns you of these issues (and many others) automatically? This I had to see.
I browsed to Protomold's site with little confidence that it could provide this service honestly. But it does with a browse-able account history, dynamic 3D markups, quotes, material and color pickers, mold polish and texture selections, and their Mold-ability Advisor just incredible. You can view a demo of the ProtoQuote system here (and you will need to install their ProtoView plug-in to see the 3D.) So, if you use CAD to design plastic parts you probably use some kind of mold-ability wizard. But I bet it's not hooked up to a team of tooling engineers in a state-of-the-art mold fabrication and production facility in the heart of Minnesota (great sausage by the way.)
Based on the report and on my gut feeling about cost (have I mentioned Rachael's flesh removing gaze?) I was looking for a quick fix. I was a QA engineer with four weeks to go before our code release, for gosh sake don'tcha know? I tried hollowing the part out from the backside, but that eliminated two of our reverse engineering workflows. In particular, Exact (NURBS) Surfacing and freeform Parametric Surface fitting.
Hollowing out the other side (the"front" with all the probing features) just looked horrifying and it would also create huge sink marks and wells on the freeform surface side. Besides, it would probably sound like a cheap toy if slid across a table. I delivered the first attempt with expressed reservations, but also with a certainty that we would do something different. I think it was unanimous it was ugly.
Well that's the pretty side, here's the back side.
While talking to Karl Mathews hmmm Karl, what do we call Karl? What nickname do you give a nine hundred pound gorilla that owns you? Right whatever name he wants. So Karl, with his customary ever-so-damn-clever thought process, and the worst Australian accent I have ever heard (not surprising since he is English), suggested we make two parts and fit them together. Great idea.
The team had discussed this possibility in passing and had toyed with the idea of a living hinge and the USB thumb drive stuck inside but what about cost? The living hinge or a family-mold for two parts would increase the shot size (the amount of plastic injected per cycle) by another third. They would also nearly double the mold size for a part already in the 8.0" x 4.0" x 1.0" range and really the two part solution would probably just end up as two molds. Either way it would lead to either a larger press or two presses in production bottomline it would just cost more. Regardless Karl wants so Karl gets and I agreed, it was the simplest and quickest solution weighing all the factors.
We received approval to research the two part approach and I hunkered down and got to work, with a slight increase in artistic license and enough time to do it right. We came up with a two part clamshell with press fit connection pins. The first part I completed was the"shell" or receiver (here with the"insert" shown semi-transparent):
And the second, the"insert", shown here from the inside with the shell transparent:
Even with the final design complete the job was not finished. Tooling issues still existed, and I knew from past experience that it is best to leave these up to, or work closely with, the tooling engineers and fabricators to get the job done right the first time. What it all boils down to is time. And time being equal to, and in our case more costly than money, we had to work out the finishing touches without wasting either.
The part's exposed surface finish needed to be matte. High-gloss would not scan well...or perhaps not at all. Textured surface finishes drive the need for increased draft angles to promote the part's release from the mold. Protomold's website, followed up by Protomold Service Technician Gene Rupp, not only helped us with texture selection and draft configuration, they did so while minimizing design rework and sorting out the ejection pin placement.
Protomold enabled us to clearly define which surfaces were"customer facing" which is an important concept in injection molding. It plays a major role with the final aesthetics. The pictures above and below are from our actual correspondence with the Protomold team showing the final ejection pin locations and "split" between core and cavity surfaces.
Gene and their tooling engineer Dave Hichens helped us formalize this"split" for the insert part and the necessary geometry (the twenty rings seen on the inside surface of the insert part) to facilitate part"grip" to the core side (blue in the image below.) We had intentionally designed the insert"backwards," so to speak, to force what would usually be the cavity side in a"normal" design to be the core side. The texture (exposed to the customer) is applied to the cavity side of the part (the green surfaces). We did this reversal due to the GD&T geometry, but also around the edge of the insert to minimize the "beauty gap" between the mating parts.
This image below is a static screen-grab of the website, but would be interactive (in 3D) if you were logged in and working with their team.
Those issues by themselves would make this job complicated enough. But there was one more problem; the gripper pins (the pins and bosses that would hold the parts together.) These were my chief concern as they are difficult to get right, the first time. Too much interference and you would need a hydraulic press and custom tooling to seat the parts correctly without ruining the finish. Too little and they would just fall apart. I asked for a"light press fit," designed the interference into the CAD model that Dave called out, added the gripper rings as prescribed, uploaded the final design, crossed my fingers, and waited for the call.
These issues, taken out of context, are an everyday part of injection molded part design. What's extraordinary is that they were analyzed, iterated, designed, negotiated, quoted, and implemented while the molds were being fabricated and without losing a single day of schedule by web and phone.
Our final choices on these matters, from the numerous engineering solutions presented to us, were based on our needs and limitations. Protomold worked within our timing constraints, budgetary limits, and went, what I consider to be, well-above-and-beyond to give us the most for our dollar while satisfying our most important criteria.
I'll finish this up this is turning into a blogumentary and gush-fest. I flew to Minnesota to pick up the parts and hand-deliver them to our beloved Tom Simon, who was now near death. I knew it was a gamble when it came to the aesthetics and assembly. I drove by Protomold's headquarters and snapped this picture still not knowing if we had parts that would actually work I had just hung up the phone with Tom.
We were both a little nervous, with Rachael and Tom Kurke waiting waiting...who could blame us.
I was pleased to finally meet Brittany Tamm and Dave Hinrichs (the faces behind the website.) Unfortunately Gene was busy helping others. They had two, sealed, ready-to-ship, boxes containing what I imagined to be our parts. They offered me water or coffee, but all I could think about was opening those boxes and pressing the parts together. After a brief but close inspection of the parts (they were flawless...and after digitally surfacing Barbie for ten years I think I would know flawless surfaces) I slowly pressed them together. I felt a slight resistance followed by a positive"seating" and then, only then, did an effortless smile appear on my face.
After many thank you's and handshakes, I had to send everyone pictures immediately! Alas, no network signal...I'm switching carriers. So I had to again wait...well only until the parking lot:
I don't even remember returning the rental car .nor the quick flight home I slept the whole way. I was and I still am floored by the success. Protomold truly knocked it out of the park. And for those that are curious, the material we chose is ABS Lustran 433 in an elegant Cornflower Blue. Our next demo part will have the GD&T engraved into the mold or perhaps just Tampo (pad) printed we haven't decided.
Speaking of"next"; what's next in this series? The plan all along was to repurpose this project for as many tasks as we could. My part, beyond telling the story, will be to show the reverse engineering of the parts in our next installment, followed by their inspection against the actual CAD model in the final post of this series.
Thank you again Protomold!