Every startup has an origin story. Some begin with a market opportunity. Ours began with frustration.

In the additive manufacturing research labs at TU Berlin, we were working with material extrusion 3D printers — FFF machines — trying to produce parts with consistent, measurable quality. The goal was not unusual: characterize a material, optimize print parameters, produce reliable results. The reality was that two prints with identical settings on the same machine could yield different outcomes, and we often had no way to determine why.

The printer was a black box. Material went in, a part came out, and everything that happened in between was essentially unobserved.

The Problem That Started Everything

Anyone who has worked seriously with FFF 3D printing knows the scenario. You develop a parameter set that works. You print ten parts, and eight are good. Two are not. You examine the failed parts, look for obvious causes — maybe a layer shift, maybe delamination, maybe subtle dimensional drift — but the printer gives you nothing to work with. No record of what happened at layer 150 when things started going wrong. No data on whether the extrusion rate drifted, whether the chamber temperature fluctuated, whether the filament feed was inconsistent for thirty seconds.

In a research context, this is maddening. You cannot publish results you cannot reproduce. You cannot isolate variables you cannot measure. And you certainly cannot build a process qualification argument for industrial use if you cannot demonstrate that the process is under control.

We tried the approaches available at the time. External sensors bolted onto machines that were never designed for instrumentation. Camera systems that could detect some defects visually but provided no quantitative geometry data. Post-process inspection that told us a part had failed but not when or why the failure originated during the build.

None of it was sufficient. The instrumentation was an afterthought. The data, where it existed, was disconnected from the process. The machines themselves were designed to execute G-code and nothing more.

The Insight

The realization was straightforward, even if the execution would prove anything but: the fundamental problem with FFF was not the hardware. Plenty of machines could reach the right temperatures, move with adequate precision, and extrude material consistently enough. The problem was that the process ran open-loop. Print, hope, inspect. If something went wrong, you found out after the fact — if you found out at all.

What the process needed was real-time monitoring with actual feedback. Not a camera pointed at the build plate for timelapse videos. Not a temperature sensor that logged to a CSV file. A system that measured each layer as it was deposited, compared the result to the intended geometry, and adjusted parameters for the next layer accordingly.

Closed-loop control. The kind of feedback that is standard in CNC machining, injection molding, and virtually every other mature manufacturing process. The kind that, somehow, did not exist in FFF.

That insight was the seed of TRACK3D.

From Research to Company

The path from a research concept to a registered GmbH is not as linear as it looks in retrospect. There were feasibility questions: could a 3D laser profiler operate fast enough to scan each layer without adding prohibitive cycle time? Could the software process point clouds in real time and generate meaningful correction signals? Could all of this be integrated into a machine that was still practical to manufacture and operate?

We spent considerable time at TU Berlin answering these questions before we were confident enough to pursue commercialization. The academic environment gave us the freedom to experiment, to fail, to iterate — and crucially, to validate the approach with rigorous methodology rather than investor-driven timelines.

When we founded TRACK3D GmbH, operating out of TU Berlin as our base, we had already demonstrated the core technical feasibility. But a working lab prototype and a product are very different things, and building hardware as a startup is a particular kind of challenge.

The Milestones

Several early recognitions gave us both credibility and momentum. Receiving the TU Berlin StartUp Label validated that the university saw commercial potential in the research. Taking second place at the Berlin-Brandenburg Business Plan Competition in 2023 confirmed that the business case resonated beyond the lab.

The Pro FIT funding from Investitionsbank Berlin, co-financed by the European Union, was a critical enabler. Developing an industrial 3D printer with integrated monitoring is capital-intensive. Sensors, precision mechanics, high-temperature components, embedded computing, software development — the bill of materials alone is substantial, and that is before you account for the engineering hours required to make everything work together reliably.

The recognition at Wista Potenzial further connected us with Berlin’s broader innovation ecosystem and provided visibility among potential partners and early customers.

With funding secured, we focused on two parallel development tracks. The TrueFormer™ 600 is the hardware platform: a fully enclosed industrial FFF printer with a 600 x 487 x 570 mm build volume, a 250 degrees Celsius chamber, a 500 degrees Celsius nozzle, dual direct drive extrusion, and more than 25 integrated sensors including the 3D laser profiler. It is designed from the ground up for process monitoring — the sensors are not add-ons but integral to the machine architecture.

SituGuard™ is the software that makes the sensor data actionable: real-time 3D in-situ monitoring, digital twin creation for every print, and adaptive layer-by-layer parameter control. The hardware captures the data; the software interprets it and closes the loop.

Building Hardware Is Hard

We will not pretend otherwise. Developing an industrial machine from scratch is an undertaking that tests every assumption about timelines, costs, and complexity. Thermal management at 250 degrees Celsius chamber temperature creates engineering challenges that do not exist at lower operating points. Integrating a laser profiler into the motion system without compromising print speed or accuracy required multiple design iterations. Making 25+ sensors communicate reliably in an electrically noisy, high-temperature environment is not a trivial firmware problem.

Every hardware startup learns some version of the same lesson: physical products do not iterate like software. A mechanical redesign means new parts, new tooling, new assembly procedures, and weeks of lead time. We learned it too, sometimes the hard way.

But there is an advantage to building the machine and the monitoring system together rather than retrofitting sensors onto an existing platform. When the hardware is designed for observability from the start, the integration is cleaner, the data quality is higher, and the system as a whole is more robust. It is harder upfront, but it is the right architecture.

The Vision

Our founding thesis has not changed: industrial 3D printing needs to be transparent, controllable, and certifiable. Transparent means you can see what is happening during the process, quantitatively, at every layer. Controllable means the system responds to what it observes, maintaining process stability without operator intervention. Certifiable means the resulting documentation meets the requirements of regulated industries.

We chose the tagline “Open the Black Box of 3D Printing” because it captures the core problem we set out to solve. FFF has enormous potential as a manufacturing technology — large build volumes, engineering-grade thermoplastics, reasonable material costs, geometric freedom. But that potential is locked behind a lack of process visibility and control. Our job is to remove that barrier.

We are still early. There is more to build, more to prove, and more customers to serve. But the foundation — a technically sound approach validated in academic research and developed into an integrated hardware-software platform — is solid. And we are building it in Berlin, at the intersection of deep technical research and an ecosystem that supports exactly this kind of engineering-driven company.