Our bodies use counter-current flow to help keep us warm in a really efficient manner. It's done by putting the ingoing blood vessels next to the outgoing blood vessels, with the system set up so that heat is always transferred from an outgoing vessel to an ingoing vessel, right through to the core.

FWIW, counter-current flow is also used in our kidneys to help shift chemicals in and out of the urine.

These models are the connecting parts for a multi-stage Peltier-based counter-current heat pump which is designed to cool down (or heat up) the air inside a “core”, while keeping temperature gradients at each stage small.

The current design expects 40mm PVC pipe (outer diameter 42.78mm, inner diameter 38.54mm, as measured from a Bunnings 40mm PVC pipe), and TEC1 12706 Thermoelectric Cooler modules (0-12V, 6A). The Peltier devices should be placed inside ~35mm sections of 40mm PVC tube with additional sealant to stop air leakage, and then attached to the caps, with holes drilled into the pipe for wire access. Two 1-sided caps go on each end of the multi-stage array, and the two-sided caps go between sections within the multi-stage array. The end caps are designed with a fitting for a 13mm garden hose or flexible PVC pipe (e.g. Boston 13mm Clear PVC Hose).

The 40mm pipe will need to be filed slightly to accommodate the Peltier devices (about 0.6mm on each side for a friction fit that uses PVC elasticity to hold the devices in place). Alternatively, the pipe can be distorted into an oval shape; this distortion method is my current plan.

As the number of stages increases, the energy benefit of the Peltier array increases. My estimate is that it will take at least 5 stages to do better than a single stage with a 2°C temperature gradient at each stage (i.e. 5 x 2°C vs 1 x 10°C). As Peltier inefficiencies increase with increasing temperature differentials, the benefit should increase substantially as the temperature gradient increases at each stage.

Note: this is mostly untested; I have no idea if it'll work.

My initial experimentation suggests an AA battery is good enough to set up at least a 4°C gradient across a TEC1 12706 Thermoeletric Cooler. Assuming linear increases, four AA batteries providing 6V should be good enough to set up a 4°C gradient across each of 4 stages, leading to a 16°C gradient overall (e.g. from 21°C room temperature down to 5°C).

My current working model uses plasticine instead of glue to connect the stages and provide some degree of leakage protection. I have a 1kg yoghurt container on the core end, and ~20mm of PVC tubing on the hot (positive) side of the outer end; this will be connected up to a 120mm funnel which is also connected up to a 120mm CPU fan for air suction / heat removal.

For more theory / working behind this, see here:

https://gringer.gitlab.io/presentation-notes/2018/10/03/two-lines-on-everything/#2020-jun-17