Behind the scenes – Injection molding
June 4, 2012 11:50 AM   Subscribe

Before each program segment for the CNC machine is executed, a laser scanning probe verifies the tool geometry and compensates for wear and tear of the tool.

Just.... wow.
posted by RolandOfEld at 11:52 AM on June 4, 2012 [1 favorite]

posted by DU at 11:56 AM on June 4, 2012 [1 favorite]

In the 80's, I worked second shift at a plastic injection molding company for a summer. Mostly, we made keys for computer keyboards. There weren't any lasers, just a lot of first degree burns (from trying to remove the piece too early) and tiny cuts (from using an exacto blade to slice the little extra bits of plastic from around the sides.

So, I guess what I'm saying is that when I looked at this link, the first thing I thought (in my best 'Four Yorkshiremen' voice) was 'LUXURY!"
posted by Joey Michaels at 12:08 PM on June 4, 2012 [2 favorites]

These machines are so accurate, they are subjected to export restrictions by the Japanese government under International Traffic of Arms Regulations (ITAR). The reason being is that these machines have the accuracy required for machining the complex 3D curvatures required by modern nuclear weapons.

Hah, cool.
posted by Think_Long at 12:10 PM on June 4, 2012 [1 favorite]

so, I skimmed their blog and their main page (and defunct web store page), and it still ain't clear: wtf are they making/selling ? I'm guessing it's a modern erector set/lego type of dohickey ?
posted by k5.user at 12:14 PM on June 4, 2012

Back in the late 80's, I worked at an injection molding plant -- and that rack of molds looks very familiar to me. The hardest part was getting a new mold design to run well. We'd get a new one, mount it up (usually in Machine #7, which was closest to the shop door) and see what it did.

Usually, what they did was not eject. We'd look at why, and then the mold would go back to the Tool and Die firm to add or move ejection pins. The second problem was interference -- the ejectors were too long, or sticking, where they're too short and the plastic flows into the pin space. Occasionally, they'd be in the wrong place and just break. Worst, though, was a misbalanced mold, which was typically only an issue for a mold making multiple non-identical parts.

An unbalanced mold was one where the plastic took much longer to reach one part than another, and you then had to find a melt temp and shot length that would manage to fill the far part, while not overfilling the other. So, you were trying to find the sweet spot between not having the short part buried in flash from the overshot, and not having the far part come out incomplete, or a short shot.

In some cases, there were no sweet spots, and the mold went back for recycling, and a new one was built.

Another big problem was cooling -- if the cooling passages were in the wrong place, you would have the part cooling unevenly, and in the worst cases, having the mold open while some of the plastic was still too soft, and then distorting on ejection. You could mitigate this somewhat by just lengthening the cycle time, but that's very expensive -- cycle time x number of cavities = number of parts per time, and thus, misplaced cooling could make for a much slower mold.

So, a new mold design would go into #7, we'd figure out what was wrong, and then decide what had to be fixed and what we'd just work around. Cooling was always the dicey one. If you were going to be running the same parts for a year, you'd get the cooling dead on to run as many parts as possible, but if you were only running a few thousand, the question changes -- how much cycle time are you going to add, and how long will it take to get the mold retooled to cool better? If you were only going to be running a couple of weeks, it was faster to eat the cycle time.
posted by eriko at 12:20 PM on June 4, 2012 [18 favorites]

k5.user, here's their Kickstarter page, which has a bit more about what they're doing.
posted by aaronbeekay at 12:21 PM on June 4, 2012

Wow, Eriko, you brought back a bunch of (suppressed?) memories for me. Yeah, that's what it was like.
posted by Joey Michaels at 12:36 PM on June 4, 2012

Legos is said to bury worn-out molds in the concrete of its new buildings. :-)
posted by Cranberry at 12:40 PM on June 4, 2012

k5.user: They're making a thing that is competition for slotted aluminum framing.
posted by thewalrus at 12:42 PM on June 4, 2012

You mean Lagos, Nigeria?
posted by weapons-grade pandemonium at 12:42 PM on June 4, 2012

You know you're a dork when someone typing "Legos" bugs you.

('Cause it's LEGO.)
posted by toekneebullard at 12:55 PM on June 4, 2012

I worked as a plastics molding machine operator for a temp job for about a month. Was payed minimum wage. Would not recommended, really - it's boring, often solitary, and slightly hazardous. Lots of better stuff to do for minimum wage.
posted by victory_laser at 1:02 PM on June 4, 2012

The only reason you're not supposed to call LEGO™ Bricks legos is due to some trademark paranoia, Like how you're not supposed to call photocopies made on a Xerox™ copy machine copies "Xeroxes" that you "Xeroxed".
posted by delmoi at 1:17 PM on June 4, 2012 [4 favorites]

Was payed minimum wage. Would not recommended, really - it's boring, often solitary, and slightly hazardous. Lots of better stuff to do for minimum wage.
Eh, probably better then working in fast food.
posted by delmoi at 1:18 PM on June 4, 2012 [1 favorite]

In the 70's I had a summer job as a graveyard shift operator on injection molding machines making golf balls. I preferred operating two machines at a time as there was more of a rhythm to the machine cycles and the motions of placing cores and then retrieving and degating the molded tree of four balls. The one four-machine station was too busy to enjoy.

While the salient concern was having fingers crushed I never really even had a close call; the greater hazard was that after the hand flourish that deposited four cores into each of the molds one of the cores would continue to spin or wobble -- like a basketball shot rolling around a rim -- and rather than settling safely down into the mold it would roll itself up and out at the exact moment the top of the mold came stamping down. Best case scenario, the core was crushed unto dust (if it were a solid core ball). Worst case, the core would roll out to the edge of the mold, to be pinched out at hellacious velocity, ricocheting off stuff.

That would wake us up for a minute or two.
posted by cairnish at 1:21 PM on June 4, 2012 [3 favorites]

Eh, probably better then working in fast food.

I'd like to see you eat leftover plastic moldings.
posted by cavalier at 1:29 PM on June 4, 2012 [1 favorite]

You don't call them Legos because Lego (at least in the UK) is a mass noun.
posted by pharm at 1:40 PM on June 4, 2012 [2 favorites]

eriko is dead on.

This was better than minimum wage, because we were on 12 hour shifts, it wasn't that hard, and except for one incident wrt cleaning the dies properly, relatively easy work. Cranes to lift the dies, awesome "art" made from spent plastics that could not be reground, just above minimum wage because of the 12 hour shift thing, alternate four day weekends, and you don't have to ask if you want fries with that. If you worked the overnight shift, you could keep a notepad in your pocket and write lyrics or poetry (I wonder whatever happened to that guy).

We were making car parts, so nothing much bigger than my fist (and more often the size of my thumb). My group manned the experimental die section; I don't know if anything we ever made that year even actually made it into vehicles, but when I look at the insides of that year-vintage and brand car, I look for "my" work ...

But the upshot of working experimental is that your machine duties always changed. Never the same thing two "weeks" in a row.

What got me was the noise. Oh the noise. I didn't have to wear ear gear (I wish I Had Done So) but it likely wouldn't have mattered. It got into your bones.
posted by tilde at 2:01 PM on June 4, 2012 [1 favorite]

We do this every time.
posted by not_that_epiphanius at 2:45 PM on June 4, 2012

The only reason you're not supposed to call LEGO™ Bricks legos is due to some trademark paranoia, Like how you're not supposed to call photocopies made on a Xerox™ copy machine copies "Xeroxes" that you "Xeroxed".

This photograph of the new Xerox machine has been Digitally Enhanced with Adobe Photoshop(tm).
posted by thewalrus at 2:49 PM on June 4, 2012

Before each program segment for the CNC machine is executed, a laser scanning probe verifies the tool geometry and compensates for wear and tear of the tool.

Arrrgh. Having spent two years thinking about moving flat, bullnose, and round-end CNC bits over computer generated geometry. I don't see how this can work. Unless they know the position on the tool that is doing the milling at each point. Or are doing 5-axis as opposed to 3-axis milling. But in general, no.
posted by benito.strauss at 3:12 PM on June 4, 2012

I would love to see Legos* being made. I could watch that for hours.

I wish some of this technology was available when i was casting and machining aluminum at the foundry. Our ancient version of AutoCAD didn't automagically create CNC code for us. I coded CNC G-codes by hand on a little screen on the turret mill from a blueprint made by a real HP plotter.

I destroyed a lot of scrap castings and mill heads when I started, but it started a love and life of programming.

* - I've always called them "Legos" unless I'm referring to a single Lego brick. I plan to continue doing so and refuse to bow to peer pressure from the internet to change. I plan to keep using most of my Americanisms of speech and writing. I have nothing against English speakers of the Commonwealth** but I find Americans who purposefully write with British affectations to be annoying and pretentious.

** - See also "Two peoples separated by a common language"
posted by double block and bleed at 3:35 PM on June 4, 2012 [1 favorite]

Yeah, memories of the injection molding plant in Pleasant Prairie. I was IT manager, but the president decreed that every employee in the office spend a week on a machine every year so we'd all have a freaking clue what the company actually did. Trimming flash, taping boxes... good times. And by "good times" I mean dreadfully dull times. I don't think I could ever do that for a living because it was so dull. But as factory jobs at that level of pay went, there were worse things to be doing.

And God, laser CNC machines? As if.
posted by fatbird at 7:00 PM on June 4, 2012

I enjoyed the juxtaposition of the old school machine shop tools and benches next to the alien spaceship machines.
posted by gjc at 5:39 AM on June 5, 2012

pharm: "You don't call them Legos because Lego (at least in the UK) is a mass noun."

Sez you.
posted by Reverend John at 8:16 AM on June 5, 2012

Of course, we only get close-up photos of the molds because it's an open-source project. Otherwise, the company would most likely be closely guarding their design so if someone tries to bite their style, they've at least got to go through the process eriko mentions above.

It might also explain why Megablocks never quite get the same feel as Legos. ( I love the story of them burying their old molds in concrete. )
posted by RobotHero at 2:02 PM on June 5, 2012

This is a huge unseen industry. The story of injecting molding must surely rank up there with the transistor as one of the giant advances in the last 75 years.

The two-part molds described here are not the only way to mold a complex part. If you have a radio with a battery compartment, (or, probably the battery compartment cover on a remote control), have a close look at the latch; you'll see that it would be impossible to completely mold with two parts sandwiching together because part of it is undercut. Further examination will reveal an extra mold parting line around the undercut part. This is a multi-part mold, and there is a section of the 'wall' that is a separate part that probably slips out sideways before the main halves separate. This ingenuity can be found on stuff throughout civilization.

If you ever built a plastic model, you have became intimately aware of the limitations and compromises of injection molding. The molds for plastic models were pantographed (traced in three dimensions into another scale) from wooden masters that were hand-built by craftsmen working from photos, and often a full-scale original object. This high labour input (as has so very much work) has been supplanted by CAD, from scanned objects or even manufacturers' original files. The CAD drives the cutters which drive the steel work and the mold creation.

The result, over 50 years, has been an emancipation of the model; from being the product of a few high-tech firms in Europe, Japan and the U.S., to a wild variety of small firms, many in Russia and the former east-bloc, and a rising output from China, naturally, all using CAD to design and deliver highly-detailed models of vast varieties of equipment. Military models are always popular, and such is the demand that now you can get a 1:35 scale model of prototypes of German armored vehicles that never went into service!

Peter Jackson, of Rings fame, grew up an unsatisfied modeler of WWI aircraft, due to the generally crappy quality of the few aircraft that were being offered. He founded Wingnut Models with some of his Rings proceeds and today Wingnut offers stunning large-scale (1:32) models of many WWI aircraft, researched and designed in NZ and produced in China.

It is a golden age for plastic modelers in many fields, and the revolution in injection molding is one of the main causes.
posted by bc_fred at 9:02 PM on June 5, 2012 [1 favorite]

« Older "And what were they serving at El Bulli? Water!"   |   Christmas in June Newer »

This thread has been archived and is closed to new comments