What Is VMC in Aviation Multi-Engine Flight? A Shapeways Quality Review Story

It was a Tuesday in late October when the order showed up in my review queue. A small aerospace firm—five engineers, a rented hangar, a multi-engine test rig with a hard deadline—had uploaded three parts to Shapeways. The instant quote came back in minutes. The design files looked fine at first glance. It was the note in the order’s comment field that stopped me:

“All parts must not affect VMC in multi-engine configuration.”

Most people in manufacturing read “VMC” and think Vertical Machining Center, that workhorse CNC machine tool you see on every shop floor. The order-routing logic thought that. A junior reviewer on my team thought that. To be fair, so did I, for about ten seconds. But the phrase “multi-engine configuration” didn’t fit that reading. A machining reference is written as a process instruction—“machine per VMC-xyz spec”—not as a performance constraint. This read like a flight-test requirement.

I’m the quality lead on Shapeways’ order review team. In practice, that means every order gets looked at by a person before it goes into production. About 300 unique parts pass through my queue in a given year, and I turned back roughly 9% of first submissions in 2024 for unclear or conflicting specs. Custom manufacturing is full of small ambiguities, and the cost of guessing wrong is usually higher than the cost of asking.

Here’s the backstory, because it matters. The customer was a startup building a hybrid-electric flight demonstrator based on a twin-engine airframe. They’d been working with a traditional machine shop, and the work was fine—nothing wrong with the parts. The problem was pace. Getting a quote took a week, and for a small team spending grant money on an hourly test window, a week of waiting was a week of runway burned. To be fair, the shop wasn’t the villain here. It ran a process built for a different kind of relationship. The customer just needed something faster, which is how they ended up on Shapeways. What they were really buying was certainty that the test date would hold.

The order was a nice example of why a multi-process platform makes sense. Part one was a sensor mount for the outboard wing section—an organic shape, low stress, low volume, exactly the kind of thing that belongs in the Shapeways 3D printing service queue. Part two was a bent aluminum bracket: 1.5 mm sheet, a 90-degree bend, a cable pass-through. That’s sheet metal fabrication territory. Part three was a thin-wall tube, about 24 cm long, with a mounting flange on one end. At production volume, a part like that normally goes to a laser cutting machine for tubes—fast, repeatable, cheap per unit. But we were making four prototypes on a deadline. Additive was the practical call, and the customer didn’t care about the process as long as the parts worked in the test window.

The trouble started in the routing review, at the step where a human is supposed to catch what the software can’t. My teammate flagged the VMC note and recommended re-routing all three parts to CNC machining, on the theory that “VMC” meant the customer wanted Vertical Machining Center work. I get why they made that call. In a machine shop, VMC almost always means vertical machining center. It’s been that way for decades. If you work in this industry, it’s the default reading.

But the default reading didn’t sit right. The note said “must not affect VMC”—that’s a constraint on the part’s effect on the aircraft, not a process instruction. The routing logic, on the other hand, was unambiguous: machining would technically satisfy the note. Machined parts won’t affect anything aerodynamic if they’re made to shape. It was the safe choice. It also would have added five days to the lead time and roughly three times the cost. The numbers said accept the re-route. My gut said ask the customer first.

And there was time pressure on top of it. The customer’s test window opened in ten days. In my normal process, I’d have routed the order that afternoon and let the checks run. Asking, and getting the answer wrong, would burn a day and blow the window. Not asking, and being wrong, would deliver parts that were late, expensive, and based on a faulty assumption. I’ve seen what that does to a customer relationship. In a previous job, a quality issue like that cost us a $22,000 redo and delayed a launch by a month. I didn’t want a repeat.

I emailed the customer’s lead engineer with a short, slightly embarrassed question: “Can we get ten minutes on the phone? I want to confirm the VMC reference on your order.” He called back within the hour.

I asked him straight: “What is VMC in aviation multi-engine operations?” He laughed, the way engineers do when a non-engineer asks something that’s obvious in their world, and gave me the pilot’s version.

VMC stands for minimum control speed. Basically, it’s the slowest speed a multi-engine airplane can fly and still hold a straight heading after one engine fails, with the remaining engine at takeoff power. Below VMC, the rudder can’t produce enough force to counteract the yaw from asymmetric thrust. The airplane starts turning toward the dead engine, and no amount of control input will stop it. It’s a certification speed—one of the V-speeds marked on the airspeed indicator and drilled into every multi-engine pilot during training.

The connection to our order took a moment to land. They were modifying the wing section of the demonstrator. The sensor mount would sit outside the original profile, and it had to be aerodynamically neutral at low airspeeds. If the mount added asymmetric drag, it could raise the aircraft’s minimum control speed in the low-speed regime—and on a test aircraft, that’s the last thing you want. The VMC note wasn’t telling us to machine anything. It was flagging the part as flight-relevant so we wouldn’t mess with the surface finish or geometry without checking first.

So we kept the sensor mount on the Shapeways 3D printing service. The organic shape was a poor fit for subtractive machining anyway, and the printed part met the drag requirement at a fraction of the cost. The bracket still went through sheet metal fabrication. The tubes were printed as prototypes, with a note in the project file: at production volume, tube laser cutting would give better wall-thickness consistency and a lower per-unit cost.

The bracket had its own little lesson. The spec called for a 90-degree bend with an inside radius tighter than our standard sheet metal tooling could hold in 1.5 mm aluminum. Our DFM feedback flagged it and suggested a revised radius. I knew the material would behave, because earlier that year I’d audited the sheet metal side of our supply chain—we were qualifying a Magnum CNC press brake machine supplier for a share of our production bending, and their test cuts gave me a bend allowance table I actually trusted. We adjusted the radius by 0.4 mm, documented it, and sent the revision to the customer. They approved it in under an hour. A tighter bend would have looked cleaner on paper. The radius we specified actually bent without cracking.

The order went into production on Thursday. Eight days later—two days before the test window opened—the customer had all three parts. The sensor mount came out clean, the bracket survived its bend, and the tubes passed inspection. The engineer’s follow-up email said something I still think about:

“Honestly, we’d have made either version work. But the fact that you asked instead of assuming is why you’ll get the next order too.”

That interaction changed how I review annotations. There are three things from it I’d pass along.

First: acronyms are context-dependent. VMC means vertical machining center in a machine shop and minimum control speed in multi-engine aviation. Both readings are correct; neither is universal. If you’re putting a note on a design file, spell out the meaning once—“VMC, i.e., minimum control speed”—and you’ll save everyone a round of phone tag. (This definition was current as of our late 2024 audits. The FAA’s Part 23 certification rules were reorganized back in 2017 and have been evolving since, so if you’re designing to a specific regulation, verify the version that applies.)

Second: quality is bigger than the part that shows up. It’s the whole interaction—the quote, the response time, the questions you ask, how you handle ambiguity. A part that arrives on time but was made on the wrong assumption is a quality failure, because it changes how the customer sees you. The cost difference between two processes matters less than the trust you build by catching a context issue before it becomes a delay.

Third: know your own boundary. A human review on every order works for us because we’re a mid-size platform with a volume we can actually watch. If you’re a huge enterprise moving thousands of parts a day, you’d need rules-based checks that catch acronym ambiguity automatically, not a person with a hunch. Your mileage may vary, and that’s fine.

I still get a little buzz when I see “VMC” in an order. It reminds me that the cheapest thing we do—reading a note carefully—is often the most valuable. The parts we shipped that week were good. The trust the customer placed in us afterward was better.