How to choose a metal 3D printer: the twelve questions that really make the decision

The way to choose the wrong metal 3D printer is to compare technical specifications. The way to choose well is to start with the part itself and end with the cost of running it for five years, leaving the laser power for last, where it belongs.

These are the twelve questions, in the order in which you should ask them. Most of them aren't directed at the salesperson: they're directed at yourself.

First, your parts

1. What is the largest part you have to manufacture, with its three dimensions?

It's the only question that definitively rules out certain machines. And there's a nuance: simply fitting isn't enough. A part that occupies 95% of the volume doesn't leave room for proper orientation, and orientation is the variable that most influences the outcome. As a rule, the largest dimension of your biggest part should be around 70% of the available space.

Also, pay attention to the shape of the volume. A cylinder with a diameter of 200 mm and a prism with a diameter of 250 mm and a diameter of 250 mm and 250 mm and 300 mm cannot accommodate the same parts, even if the volume in liters is similar.

2. How many per year?

It makes clear whether the conversation is about productivity or flexibility. With twenty parts per year, a faster machine doesn't pay for itself; with two thousand, each minute of cycle time is multiplied by two thousand.

3. With what material, and who decides that material?

If you work with catalog alloys, almost any machine will do. If your business involves qualifying your own material, or a material that isn't currently listed, this question becomes the most important of the twelve and reappears as number 7.

4. What tolerances and finishes does the drawing specify?

If the answer includes machined surfaces afterward, the machine's raw accuracy matters less than it seems. If the part has to be finished, it matters a great deal, and it's wise to be wary: an LPBF part doesn't come out with machining tolerances, no matter who makes it.

Next, the installation

5. Where will it be located, and what is already there?

Here are the costs that no one quotes. Available power and type of connection. Space to open the machine, not just to install it. Extraction. Argon or nitrogen supply, and whether it will be a cylinder or a tank. And whether the opening where it will go has a doorway large enough—we've seen all sorts.

A machine that runs on 230V single-phase and fits through a standard door can be installed in a day. One that requires dedicated three-phase power, external cooling, and construction work takes months and costs more than what was included in the original offer.

6. Who will be managing it, and how much time can they dedicate to it?

If it's a part-time worker who also does other things, the machine needs to be easy to start and clean. If there's a dedicated team, more complexity can be accepted in exchange for more control. There's no single right answer.

And now, the machine

7. Is the parameterization open or closed?

The question with the most long-term consequences, and the one that's asked the least. Specifically: can you modify power, scan speed, layer thickness, and fill strategy, save the result as your own parameter set, and export it?

If the answer is no, you're stuck with the manufacturer's materials catalog forever. And it's important to understand why many people can't open it even if they want to: they integrated the machine with catalog components, and they bought the slicing and control software from a third party. Those who didn't write the code can't open what that code doesn't expose.

8. Does he force you to buy the powder from him?

The equivalent in consumables to the previous question. Captive powder is a recurring, non-negotiable cost that grows with your volume, precisely when it matters most. Ask about it explicitly, and ask if using powder from another supplier voids the warranty.

9. What does it record, and what record comes out of the machine?

Applied parameters, layer-by-layer history, oxygen, gas flow rate, actual power. And the forgotten part: whether that can be exported in a usable format. Without an exportable record, a part cannot be qualified because there is no evidence to show.

10. What happens when something goes wrong?

Response time, technician location, availability of critical spare parts and their country of origin, and what can be resolved remotely are all important factors. A machine down for three weeks waiting for a replacement part costs more than the price difference between two parts of equipment.

Asking about the real origin of key components is not chauvinism: it determines the lead time for a spare part.

11. How much does it cost to keep it running for a year?

This is separate from the price. It includes electricity consumption, inert gas, filters, protective windows, powder handling consumables, calibrations, maintenance contracts, and the operator's labor hours. In many installations, this figure approaches 15-20% of the equipment price each year.

12. And now, laser power and volume.

At the end, not at the beginning. Because by then you already know what part you have to make, how many, and with what material, and those three things have already narrowed down almost everything. More power allows you to go faster or use thicker layers, but it also provides more energy per unit of length, which can lead to keyhole mode and fill the part with large pores if the parameters aren't right. More watts isn't better: it's different.

The test that's worth twelve

All of the above can be faked in a meeting. This cannot: bring your part and ask them to manufacture it.

Not a manufacturer's sample part, already selected and adjusted. Yours, with its drawing and tolerances. And ask for three things: the parameter file they used, the process log, and the actual build time. Their willingness to provide these three things tells you more about a supplier than any product sheet.

Even better if they let you watch the machine in operation with someone who uses it daily. Half an hour next to the operator answers questions you didn't even know you had.

What you shouldn't use as a criterion

• The number of machines installed. It tells you who has been selling the longest, not which one is right for you.
• Awards and trade shows. They're nice, but they don't manufacture parts.
• Comparing scan speeds in the abstract. Without layer thickness and strategy, a number of meters per second means nothing.
• Density "up to 99.9%". Without specifying the material, parameters, and measurement method, it's just a statement, not a fact.

Where are we

We manufacture LPBF machines, so this isn't neutral, and we're not going to pretend it is. What we can do is say which questions on that list we're a good answer to, and which ones we're not.

We are a good answer in 5, 7, 8 and 9: the ALBA operate on 230V power and enter through a normal door, the parameterization is completely open because the slicer and the control are ours, we do not force you to buy the powder from us, and the process history is exported.

We're not the answer if your part doesn't fit within a 250 × 250 × 300 mm volume, if you need multiple lasers working simultaneously, or if you're looking for the largest installed base in Europe. We prefer to say this beforehand.

And if you want to try before you decide, you don't need to buy anything: you can bring a part and have it made on one of our machines.

—

What's inside one of these machines, in what is a metal 3D printer. The cost of setting up the entire installation, and ways to surround it, in the real CAPEX of entering additive manufacturing. And the comparison between the two platforms, in machines.

And the chapter that almost never appears in the comparisons —what gas each material uses and what installation it requires— in argon or nitrogen.