Summary
A factory profile is a starting point, not a solution. The manufacturer builds it to work acceptably for as many users, materials and parts as possible. That is why it rarely fails badly, but it doesn't optimize anything either. For prototypes, visual models and training, that is perfectly fine. For accurate dimensions, repeatability and known mechanical properties, it is not. Tuning a profile only makes sense if the profile is then locked, documented and always used the same way. Tuning without that just replaces one unknown with another.
What the manufacturer doesn't know about your part
The manufacturer doesn't know your tolerances, your material or the conditions your part will work in. It doesn't know whether strength, surface finish or build speed matters most to you. So it makes a profile that covers everything and optimizes nothing. The result is consistent, but average.
Three sources of difference between paper and practice
The first is material. Every batch of material is slightly different, and the same generic profile is often used for materials from different manufacturers. With filaments there is another common cause, moisture. PA, PETG and TPU absorb moisture from the air, and wet filament gives a poor surface and weaker parts regardless of the profile. Dry the material before tuning the profile. The second is the machine. Your machine is never exactly the same as the machine in the documentation. Calibration, backlash, vibration, nozzle wear and component wear all change how it behaves in real conditions. The third is the environment. Room temperature and humidity, drafts, even whether the machine sits in a cold workshop or an office.
How this differs by technology
In FDM printing, the profile is almost everything. Temperature, flow, retraction, cooling and speed are set directly, and this is where tuning has the biggest effect. In resin printing, the key parameter is layer exposure time, which depends on the specific resin and even on its temperature. In SLS and MJF systems, the user changes fewer parameters. But part orientation, build chamber packing, the share of refreshed powder and the scaling factors that compensate for shrinkage still decide the outcome. In metal PBF, the manufacturer's parameters for standard alloys are developed through serious testing and are usually a good baseline. You don't change parameters for the sake of changing them. Developing your own parameters makes sense when you need a property the standard set doesn't deliver, or when you work with a material the manufacturer doesn't support. But even the best factory set doesn't know your geometry. Orientation, supports and behavior around overhangs are always your decision, and that is where most problems start. For a first check of a metal part, you can use the SLM DFAM Checker. In DED systems, parameters depend on geometry more than in any other technology. Heat accumulates, so the same parameters on a thin wall and on a massive block do not give the same result. Factory settings here are only for getting to know the machine.
When a factory profile is perfectly enough
For prototypes where you are checking shape, for visual models, for training and for parts with no requirements for dimensions or strength. There, tuning the profile spends time that brings nothing measurable.
How a profile is tuned in practice
First, dry the material and check the machine's condition, calibration, nozzle and build plate flatness. Second, set the basic parameters on simple test models, temperature, flow and retraction for FDM, or exposure time for resin. Third, print a test part with known dimensions and measure it. Use the result to set dimensional compensation. Fourth, if mechanical properties matter, print and test specimens in the orientation the real part will be built in. Fifth, lock the profile.
A locked profile is worth more than a good profile
The most common mistake is not a bad profile. The most common mistake is a profile that keeps changing. Every small "let's try" tweak means you no longer know why a part is good or bad. When a profile gives good results, give it a name and a version, save it, record which material, machine and batch it applies to, and don't change it without a reason and a record. That is the first step toward repeatable production. The whole process is described in What Process Qualification Actually Looks Like, and checking each individual part in How a Single Part Gets Qualified for Production. Software tools that optimize parameters automatically exist and are improving, and I wrote about where they really help in AI in 3D Printing: Where It Really Helps and Where It Doesn't. But they don't replace the discipline of locking and documenting either.
Frequently asked questions
Is a factory profile bad? No. It is made to work acceptably for everyone, so it is optimized for no one. Should I change the manufacturer's parameters for metal printing? For standard alloys, usually not. The manufacturer's parameters are a good baseline. Orientation, supports and overhang strategy are your decision. How often should a profile be rechecked? With a new batch of material, after service or a nozzle change, and whenever results change for no obvious reason. Why does the same profile give me different results? Most often because of moisture in the material, a different material batch or the machine's condition. Rarely because of the profile itself. Can one profile apply to several machines of the same model? As a starting point yes, but each machine should be checked with a test part before the profile is used in production.
Pilot to Production Program: from prototype to serial application →