What I've Learned Buying Ultimaker, CNC Machining, and Laser Cutting: An Admin Buyer's FAQ

Below are the purchasing questions I get asked most often—by engineers, by our CFO, and by people who've just been told to “handle the workshop setup.” I manage office operations and buying for a machine shop with about 40 employees, so I order everything from Ultimaker 3D printers to BMT 75 tool holders. These are my practical, no-jargon answers.

Are Ultimaker 3D printers worth the premium for a business?

Short answer: yes, for us. We bought our first Ultimaker S5 in 2021 and added an S7 in 2024—plus a Factor 4 later. I demoed it myself and the purchase cleared this year, but I'm getting ahead of myself. When I say “premium,” I mean you can absolutely buy a cheaper open-frame printer. The reason we stick with Ultimaker isn't the hardware specs. It's that we don't have to babysit the machines. The S5 prints, Cura slices predictably, and Digital Factory lets me check job status without walking to the workshop.

That reliability became a client-facing issue. We make jigs and fixtures for actual production lines. When a print fails, it's not “we'll reprint tomorrow”—it's a delay that a customer notices. The first time a cheaper printer's print head jammed two hours before a vendor demo, I understood why quality is a brand issue. I don't have hard data on hours saved by switching, but my sense is the reliability premium paid for itself in about eight months.

What Ultimaker 3D printer filament should I order?

It depends what you're making. For our customer fixtures, we use Ultimaker Tough PLA—it's stiff enough and doesn't warp as much when a part sits near a warm motor. We go through maybe 60 spools a year. 50? No, 55, I'm mixing it up with our TPU usage. Most of it is Tough PLA; the rest is TPU for soft-touch pads and PVA for supports.

To be fair, you can buy generic filament that claims the same specs. We tried that once (circa 2022). The dimensional variance was just enough that our press-fit inserts stopped fitting—we measured a 0.3 mm difference on a 20 mm boss. That's fine for a decorative part, not fine for a fixture. Put another way: the material cost difference was maybe $12 per spool. Rejections were costing more than that in staff time alone.

One thing I wish someone had told me earlier: store filament dry. We bought a dryer after a spool of TPU sat open over a humid weekend and printed like wet spaghetti. That's a $100 piece of gear that eliminates a very frustrating start to the day.

When does it make sense to use a large part CNC machining shop instead of 3D printing?

When the part needs to not be plastic, mostly. We outsource to a large part CNC machining shop for anything that has to carry real load, see coolant, or hold tolerances tighter than FDM can manage. Our rough rule: if it needs better than ±0.25 mm or a surface finish that looks injection-molded, we call the CNC shop.

That said, CNC has its own headaches. Lead times run 2–3 weeks; one shop we use quotes “10–15 business days” and that's optimistic. Minimum order quantities sometimes force you to buy 20 when you need 3. And large-part machining gets expensive quickly—one aluminum fixture housing cost us $2,300, where a 3D print would have been $80 in material. The most frustrating part: deciding the route is rarely a clean technical call. It's cost, speed, quantity, and what the client expects to see. I've had engineers insist on metal when plastic would have passed every test. I've also had to push back the other way and explain that client perception matters too—sometimes they simply want to see machined aluminum.

What is a BMT 75 tool holder and why would I need to buy one?

BMT 75 is a mounting standard for live tooling on CNC lathes. We had to order a set when we brought some turning work in-house instead of sending it out. The holder fits a specific turret interface—if I'm being precise, it's the style used on a range of Doosan and similar lathes. (I had to verify the spindle orientation twice because we got one clamp backward in our initial order.

Do you need them? If your shop has a lathe with BMT 75 positions, yes, and buy from a reputable supplier. We bought a cheap “compatible” version once from a marketplace listing. The tool ran hot, left chatter marks, and the next holder we ordered was the name-brand equivalent—about 35% more expensive. Which, honestly, I should have budgeted for from the start. Quality perception doesn't just apply to external clients. It applies to your own machinists trusting the setup enough to hold tolerances.

I don't have hard data on lifespan differences. What I can say anecdotally is that the name-brand holders have held their alignment for over a year of daily use, while the cheap one wobbled within three months.

How accurate is laser cutting, actually?

Laser cutting is accurate—but “accurate” depends on what you're measuring. A modern fiber laser can hold ±0.1 mm on features up to a certain size, if the sheet is flat and the operator knows what they're doing. For thinner material like 3 mm mild steel, you get clean, repeatable cuts. For thicker plate or reflective materials like copper, it gets trickier.

Take this with a grain of salt: I've seen test reports from two laser vendors claiming ±0.15 mm on 10 mm steel, but the actual parts we measured showed a slight edge taper. The beam isn't perfectly vertical, so the top of the cut differs from the bottom by a few tenths. That matters for mating parts. It didn't matter for a simple bracket.

We use laser cutting when we need flat metal parts quickly and edge squareness isn't critical. If it is critical, we machine them. If we need a prototype this afternoon, we 3D print it. Every process has a best-use zone.

Which process makes us look best to clients?

Whatever produces a part that fits and works on the first try. I say that half-jokingly, but it's the real answer. When a client watches a fixture clamp down on the first go, they walk away impressed. When they see a printed part with rough layers in a high-visibility spot, they question everything else you've done. That's not fair to the technology, but it's true.

Per FTC guidelines on advertising substantiation (ftc.gov), any performance claim needs to be backed by evidence. I've applied that to our own vendor qualifications: if a machining shop or filament supplier can't provide test data for their specs, they don't make our approved list. You'd be surprised how many sales reps talk about “holding tight tolerances” but can't produce a single inspection report.

My practical advice: match the process to the part, but also match it to the impression you want to make. A 3D-printed tool that does the job is fine for internal use. If it goes to a client site, post-process it or machine it. The $50 difference in part cost is cheap compared to the cost of a client thinking your work is sloppy. Quality in, quality out—and the purchasing department is where that starts.