The creation and use of models in science is of great importance for knowledge production and communication. For example, toy models are often used as idealized explanatory models in physics education. Models can be a powerful tool for exploring phenomena in ways that facilitate learning. However, careful consideration of instruction and explanations needs to be considered to guide how students relate models to real-world phenomena in subject-correct ways. A design experiment was conducted to investigate how upper secondary school students can use models for learning in the physics laboratory. The intervention used in the study was a laboratory exercise developed over three phases where students worked with a mechanical Paul trap and a simulation to understand the principle behind a real Paul trap. Each phase of the study consisted of three to five laboratory sessions. The data were analyzed using thematic analysis and the learning process was understood using the theoretical framework of variation theory. From the results, it was possible to identify patterns of variation for successful lab groups and critical aspects and features students need to discern to effectively modelize the mechanical Paul trap. The findings also indicate that having students work with models can be a meaningful clarificatory process to develop a deeper understanding of the use and limitations of models in science.