How to calibrate the laser and its optics in a metal 3D printer

Calibrating the laser on an LPBF machine isn't a single operation; it involves four distinct checks that can become miscalibrated for different reasons and result in different defects: the actual output power, the focal point, the accuracy of the scan field geometry, and—in multi-laser machines—whether all lasers are aligned. Each check has its own specific symptom, and recognizing the symptom saves weeks of blindly adjusting parameters.

1. The real power

What is being checked. That when you request 300 W, 300 W actually reach the work surface, not the fiber optic output. Between the two, there are lenses, mirrors, and a protective window, and all of these absorb some power.

With what. A thermal power meter placed on the platform plane. The working range is scanned —at 20, 50, 80 and 100% of setpoint— and the setpoint versus delivered power curve is plotted.

Why does it fail? Almost always through the protective window: it gets dirty with condensation from the process and absorbs more and more light. It can also be due to diode drift over time, and to the degradation of the mirror coatings.

What defect does it cause? Progressive lack of fusion. Parts that were coming out well now start coming out porous without any changes, with the irregular and elongated pores characteristic of insufficient energy. It's deceptive because it develops gradually.

2. The position of the focus

What is checked. That the narrowest point of the beam is exactly in the plane where the powder is spread. If the focus is above or below, the point reaching the bed is wider and the energy density drops, even if the power is correct.

With what. The usual method is a focus ramp: a series of lines are drawn on a metal plate, varying the focus position by a few tenths of a millimeter for each line, and then the narrowest and most defined line is measured. Equipment with a beam analyzer measures this directly.

Why it fails. Thermal expansion of the optics when heated, changes in window thickness when replaced, and mechanical displacements of the print head.

What defects does it cause? Thicker outlines than necessary, dimensions that run in the same direction across the entire part, and a poorer edge finish. If the blur is significant, it indicates a lack of blending.

3. The geometry of the swept field

This is the most laborious and the one that causes the most silent problems.

What is being checked. That when the software requests a point at coordinate (100, 100) the beam falls on (100, 100) in reality, and that this holds true across the entire work surface, not just in the center.

With what. A known grid—a matrix of points or crosses spread across the entire field—is marked on a plate, measured with a measuring machine or by vision, and compared to the theoretical standard. The deviation of each point feeds into a correction table that the control system applies in real time.

Why does it disappear? It's intrinsic to optics: two rotating mirrors don't project a perfect reticle, and the lens introduces its own distortion. That's why it's not eliminated, it's compensated for. It also changes with temperature and when any optical element is replaced.

What defect does it cause? The unmistakable symptom is a part that's correct in the center of the build plate but out of tolerance at the corners, with the deviation increasing towards the edge. Also, parts that come out fine but then stop coming out when placed in a different spot on the build plate, which is the most frustrating way to waste time.

4. Laser coincidence

This only applies to multi-laser machines, and it is the reason why multiplying lasers does not automatically multiply productivity.

What is being checked. That two lasers that have to stitch the same area point to the same spot with an error of a few microns, and that the overlap between their fields leaves neither a gap nor a double bead.

With what. Reference marks manufactured by each laser in the shared area and measured afterward. Each one is corrected against a common origin.

What defect does it cause? A line of weakness in the seam, visible when cut, right where the two sections meet. In parts with fatigue resistance requirements, this is a serious problem because it is a flat, oriented defect.

How often

There is no universal answer, but there is a reasonable criterion:

What's indisputable is what needs to be done before a long or expensive project: check the power and clean the window. It takes twenty minutes versus thirty hours of machine time.

The daily check that everyone does

Before launching, three things that aren't calibration but prevent most discards:

• The protective window, clean and unmarked. A stain absorbs energy, heats up, and eventually breaks the window.
• The flatness of the coating in the first layer. If the first layer isn't uniform, none of the subsequent layers will be.
• The oxygen level in the chamber, below the set point and stable before starting, not yet decreasing.

What to ask a manufacturer

1. Which calibration procedures can I perform myself, and which require a technician?
2. Is the sweep field correction table accessible, or is it locked inside the control panel?
3. Is a record of each calibration, with date and values, exportable?
4. Does the machine notify me when a calibration is due to expire?
5. What is the cost of a complete calibration service, and what is the turnaround time?

The second and third points are the most important. Without an exportable record, nothing can be rated, because calibration is precisely one of the things that needs to be frozen. And a closed correction table means depending on the manufacturer every time a lens is modified.

How we see it

The optics, controls, and software of the ALBA are all proprietary developments, and this has a practical consequence in this matter: calibration procedures are executed from SamyStudio, directly on the machine, and the results are recorded and can be exported. There is no black box that can only be opened by a technician with a password.

It's not a flashy brochure advantage. It's the kind that becomes apparent the day you have to show a client why the part turned out well.

—

What exactly is being calibrated, in what is a laser and why does a fiber laser melt metal. The defects that appear when some of this is missing, in porosity and density. And the difference between monitoring the process and correcting it, in closed-loop control.