Build an insulated box and save Nallo’s ice. Then turn the box into a refrigerator, one part at a time.
What keeps Nallo’s supplies cold?
Your next step and Nallo tips are in the workbench.
Your build is ready. What do you predict?
Choose a mission above or build freely. Test a build, save A and change only one thing if possible.
Build an insulated box and save Nallo’s ice. Then turn the box into a refrigerator, one part at a time.
Follow the next step in the workbench. You can also tap the panels in the picture.
Build or change one thing, discuss your prediction, then start the trial. “Undo” restores recent build changes and restarts the trial.
Follow the workbench: fit the floor, walls and lid, then test the box. Compare two equally long trials, changing one panel. Add the four cooling parts and connect their pipes to make a refrigerator. Individual parts and settings remain available in the expanded workbench. The yellow Nallo band shows the mission and goal; the workbench gives the next step. “How the experiment works” contains three instructions for each mission.
At low pressure, refrigerant in the evaporator absorbs heat from inside. The compressor compresses the vapour and needs work. The outer heat exchanger, the condenser, releases heat to the surroundings. After the expansion restriction, pressure and temperature drop again. Moving marks follow the closed circuit; they do not represent air circulating through the refrigerator.
The floor, walls and lid are independent panel regions with their own material and thickness. A region may remain open. The heat model assigns 20% of the area to the floor, 60% to the walls and 20% to the lid. Open regions are simplified strong thermal bridges. Active cooling requires all four parts and all four pipes. Compatible sockets define the circuit; this is not an assembly guide for a real refrigerant system.
The thermal resistances of two layers add in this simplified model. Reversing layer order while keeping each material’s thickness does not change steady heat conduction. A larger box has more surface area. More supplies have more heat capacity. Change only supply mass to investigate that effect. The second layer applies to all fitted panels; an open region stays open. Material and thickness of the first layer affect only the selected region. A zero-thickness second layer has no effect.
Predict which panel insulates better. Save a completed trial as A and change only material or thickness. Keep room temperature, door and starting temperature equal. Both operating modes normally start at 8 °C. Increase starting temperature deliberately for warm supplies. Compare the same target temperature with clear and obstructed ventilation behind the outer heat exchanger.
A thick wall does not create cold; it slows heat transfer. A refrigerator does not destroy heat. It transports heat and releases additional heat from its electrical input. Cooling power in watts is different from accumulated electricity consumption in watt-hours.
Compare equal ice cubes in two equal closed containers, with and without added insulation. Put them in the same place and observe after the same time. An adult can show where a real refrigerator releases heat. Do not open covers or alter pipes.
The model balances wall conduction, simplified door heat input, heat capacity and the latent heat of 120 g of ice. An idealised temperature-dependent coefficient of performance includes an efficiency factor. Larger heat exchangers and better ventilation reduce the modelled temperature difference of the cooling cycle. Refrigerant-specific pressures, humidity and three-dimensional airflow are not simulated. Values are for learning, not refrigerator design. Red heat-flow marks show power, not molecular speed. Box volume changes area and heat capacity. Wall layers do not have separate temperature histories. The ice and inside space start with the same conditions in both modes; warm supplies may also release heat outward.
Write a prediction before starting. Save a completed trial as A, change one thing and test again. The result keeps your prediction and names the changed parts. Mission feedback identifies conditions that must match. Free building is always available.
A learning simulation with simplified, transparent models. Save your build as a plan. A/B comparisons last until the page reloads. Your plan, prediction and results are not sent to a server.