Learners’ conceptions of program dynamics shape their reading, writing, and debugging of code. But, the invisibility of underlying program behaviors that transform code to outcomes challenges learners. Teachers adopt notional machines, defined as simplified notions about program dynamics, to support learners within computing classrooms. Researchers have gathered notional machine examples by interviewing post-secondary instructors. But, there is a need to capture notional machines in practice within computing classrooms, particularly introductory high school classes where teachers communicate program dynamics to learners with limited to no prior programming experiences. Through a qualitative video analysis of seven online class periods (80 minutes each) across 14 weeks of an introductory physical computing electronic textiles high school unit, this paper answers: (1) What notional machines did the teacher use in practice? (2) At what levels of granularity did they communicate program dynamics? And, (3) What representational forms did they take? The analysis revealed a three-tier framework to capture notional machines in practice. First, notional machines belonged to one of the five themes depending on the layer of abstraction simplified within electronic textiles. Second, they differed along the levels of granularity—individual atoms, program blocks, relations between blocks, or the entire program. Third, they took two distinct representational forms—verbal explanations and participatory roleplays. Overall, the analysis has two-fold contribution: provides a framework for future research to capture and study notional machines in practice, and, at the same time, presents one of the first accounts of notional machines adopted within a high school introductory physical computing unit.