In March 2024, 36 hours before a line restart, a production manager called us at 2 PM. She was not calm. The original supplier had a laser cutter, but the shop’s CO2 laser lens (2 inch) was cracked and on backorder. No lens, no laser cut. And searching for “projects for laser cutter” at midnight wasn’t going to fix a problem that actually needed a machined bracket with a pocket and a bend, not a flat acrylic shape. Oh, and their replacement lens was on a truck somewhere, but not in time.
We’re a small prototyping shop: three staff, one Ultimaker S5 and an older Ultimaker 3D printer. In my role coordinating rush orders, I get calls like this more often than you’d think. Last quarter alone, we processed 47 rush orders with 95% on-time delivery. Maybe 46, I’d have to check the system. But this one had to be in that 95%. Normal turnaround for a custom bracket is two or three days. We had 36 hours. Most of our rush orders range from $500 to $15,000; this one was at the bottom of that range, but urgency doesn’t care about price.
I remember thinking we could print it. The bracket wasn’t structural; it just held a sensor and stopped it from wobbling. FDM is perfect for that. At least, that’s been my experience with deadline-critical plastic parts. I checked the Digital Factory queue, and both Ultimaker 3D printers were free until Thursday. “We can definitely do this,” I said. Then I felt my throat tighten.
The first mistake I almost made
In my first year, I made the classic specification error: assumed “standard” meant the same thing to every vendor. Cost me a $600 redo. So this time I asked a lot of questions. What material? What temperature around the sensor? What hole spacing? The answers: sensor runs at 60°C, PETG is fine, hole spacing within 0.5 mm. That’s doable on an Ultimaker S5 with a 0.4 mm nozzle at 0.15 mm layer height. I imported the model into Cura and checked the orientation. The original part had a bend in it, so we printed it standing up with thicker walls at the bend. It took 15 minutes in Cura to get right, and it saved us from a part that would have snapped.
We set infill to 35%, four perimeter lines, and a first layer height of 0.2 mm for better adhesion. Cura estimated 9 hours 47 minutes. That left a buffer, but not enough to waste a spool. Then I almost skipped the overhang check. I knew I should double-check it, but I thought “what are the odds?” Then I remembered May 2023. We skipped the final review because we were rushing and it was “basically the same as last time.” It wasn’t. $400 mistake and a customer who almost didn’t come back. I told my assistant about it, and he went back to the slice settings. Our company policy now requires a 48-hour buffer on every deadline because of that one.
The printer, the filament, and the $55 overnight fee
Then the real problem hit. We had a nearly full spool of PETG on the shelf. “We have enough,” I said. That was another almost-mistake. After the purge and test block, the spool was at 32%, and the part was going to need 38%. I did the math and went pale.
We needed more filament fast. I called our account rep and ordered the same PETG from the Ultimaker store, paying $55 for overnight shipping. The filament itself was $38, so the total was $93 for a part that was already late. It wasn’t the cheapest option. It was the right one.
After I hit confirm, I kept second-guessing. What if the courier didn’t make it? What if the batch was different? The next 18 hours were stressful. I didn’t relax until the parcel arrived at 10:14 AM the next day.
While we waited, we used the remaining PETG to print the small alignment jig first. That used less filament and kept the printer ready. When the new spool arrived, we switched it out, purged the hotend, and started the main bracket. The first layer looked a little squished, so I leveled the bed again. You know that question people ask: “When you start a car, do you press the brake?” You do, even if you’re late, because you don’t want the car lurching into a mailbox. With a 3D printer, the brake is checking the first layer before you walk away. We slowed to 15 mm/s and watched the first loop. It was boring. It was also why the part came out usable.
Not every urgent job is a laser cutter project
I have nothing against laser cutters. To be fair, they’re great for flat templates and quick gaskets. We use one for those. But the customer’s CO2 laser lens (2 inch) was backordered, and even if it hadn’t been, a laser cutter can’t make a pocket or a curved bend. It cuts through. For a sensor bracket with a hole pattern and a bend, FDM on the Ultimaker was the right tool. PETG doesn’t crack like acrylic, and the S5’s build plate had enough room to print the whole piece in one go.
We printed through the night. I woke up at 4:30 and checked the remote camera. It was 87% done. I watched the last layers from the kitchen, coffee in hand. There’s something satisfying about a perfectly executed rush order. Seeing it delivered on time and correct, after all the stress, is the payoff.
What I’d do differently
The part shipped at 4:15 PM Wednesday, almost 14 hours before the deadline. The customer’s line restarted Thursday morning. We paid $55 in rush fees, but it saved a $600 order and, eventually, a $4,700 relationship. Later, she told me the plant was losing about $6,000 an hour during an unplanned stop. That put the $55 into perspective.
A larger shop might have turned them away because $600 is not a $10,000 PO. We didn’t, and that’s the point. Small doesn’t mean unimportant—it means potential. When I was starting out, the vendors who treated my $200 orders seriously are the ones I still use for $20,000 orders. This $600 bracket was the entry ticket to a much larger relationship.
Also, keep spare filament. Order from the Ultimaker store before you need it, not after. The $55 rush fee was our fault for not checking the spool earlier.
And on the technical side: measure twice, respect layer adhesion, and never skip the first-layer check. When you start a car, you press the brake. When you start a print, you check the first layer. The customers who need you in a panic won’t remember that you were fast. They’ll remember that you were right.
