Does Walter need another fan?
Test the complete head, including the cooler, covers, wiring and peripherals. Compare cooling changes under the same conditions.
Whole-build temperature prediction: not ready. This is an engineering workbench and test protocol, not laboratory certification.
What the current airflow tells us
The cooler currently blocks air in the numerical model but does not conduct heat from the Pi. Its fan is absent. An extreme color on that model cannot tell us that the printed head will overheat.
The cooling design to test
Review a real temperature run
Import the supplied CSV format or a Walter thermal-run JSON. Files stay in this browser tab. Missing channels remain unknown. No estimated readings are inserted.
Record references document provenance; typing a reference does not verify its contents. The maker must review the actual calibration and uncertainty records.
Open a recording to see its actual readings and missing evidence.
Proposed review criteria: at least30minutes, final10minutes settled, no gaps over15seconds, stable power, CPU target75°C and test-stop review at80°C. These are project targets, not battery, material or touch-temperature limits. All component limits still need selection and review.
A valid comparison needs matched hardware, workload, cooler, fan policy, power and ambient. A cardboard run cannot stand in for an open-bench run.
Whole-build test sequence
| Case | Setup | Decision |
|---|
Every modeled part needs a thermal identity
Inspect all33 modeled solids
| Part | Modeled size · mm | Thermal definition |
|---|
Heat-source evidence and missing loads
| Source | Available reference | Measurement needed |
|---|
Before the next simulation earns trust
Coupled solid-and-air heat transfer, actual fan performance, grid refinement and comparison with physical measurements are required. More colorful particles cannot replace those checks.