Thermoelectric (Peltier) Cooling Analysis
Theory: Peltier effect versus Joule heating
Overview
I want to cool a sensor slightly below room temperature and am considering a Peltier device. Does more current always mean more cooling?
It doesn't work that way. Current through a Peltier device pumps heat at the junctions (the Peltier effect), but it also generates Joule heat inside the device. Pumped heat is proportional to current while Joule heat rises with its square, so cooling peaks at a certain current and falls beyond it. Heat also conducts back from the hot side. The basics are to view cooling capacity, input power and hot-side rejection as functions of current and temperature difference.
Basic equations
$\alpha$ is the module Seebeck coefficient, $R$ electrical resistance, $K$ thermal conductance, $T_c$ and $T_h$ cold- and hot-side temperatures and $\Delta T = T_h - T_c$. Half the Joule heat is taken to return to the cold side.
$Z$ is the figure of merit; materials with larger $ZT$ reach larger temperature differences and higher COP. Bismuth telluride near room temperature has $ZT$ around 1.
A 150-year wait for practical use
The Peltier effect was discovered in 1834 by the French watchmaker Peltier, who found that current through a junction of dissimilar metals heats or cools it depending on direction. In metals the effect was small, overwhelmed by Joule heating and conduction, and impractical. That changed in the 1950s, when semiconductors such as bismuth telluride were found to give far larger effects. Today, with no moving parts, silent operation and precise temperature control, Peltier devices serve in laser temperature control for optical communications, small refrigerators, wine cellars and PCR instruments.
Worked example: a 127-couple module
Assume a 127-couple bismuth telluride module with $\alpha$ = 0.0508 V/K, $R$ = 2.0 Ω and $K$ = 0.5 W/K ($ZT$ = 0.77 at 300 K). With the hot side held at 300 K, the maximum temperature difference is 68.9 K and the maximum cooling at zero difference is 58 W (7.6 A).
Example 1: pumping 20 W
| ΔT | Current needed | Input power | COP | Hot-side heat | Cooling at COP-optimal current |
|---|---|---|---|---|---|
| 10 K | 1.96 A | 8.7 W | 2.31 | 28.7 W | 5.8 W (COP 3.65) |
| 20 K | 2.58 A | 15.9 W | 1.26 | 35.9 W | 9.9 W (COP 1.51) |
| 30 K | 3.39 A | 28.1 W | 0.71 | 48.1 W | 12.2 W (COP 0.80) |
At 30 K difference, pumping 20 W takes 28 W of electricity.
Yes — and the hot side must reject the 20 W pumped plus the 28 W consumed, 48 W in all. Peltier efficiency falls sharply as the temperature difference grows. The right-hand column shows that running at the COP-optimal current is efficient but pumps little heat per module. So if efficiency matters, use several modules at lower current each; if cost or size matters, use fewer modules at higher current.
Example 2: heat-sink influence
The cold side was held at 15 °C with 25 °C ambient, varying heat-sink resistance. The hot-side temperature rises with the heat rejected:
| Heat-sink resistance | Maximum cooling | Hot side then | Cooling at 3 A | Hot side at 3 A | COP at 3 A |
|---|---|---|---|---|---|
| 0.1 K/W | 41.2 W | 38.6 °C | 27.5 W | 29.8 °C | 1.36 |
| 0.2 K/W | 35.1 W | 48.2 °C | 25.3 W | 34.2 °C | 1.21 |
| 0.5 K/W | 22.3 W | 61.3 °C | 19.4 W | 46.1 °C | 0.85 |
A poor heat sink raises the hot side, widening the temperature difference and cutting both capacity and COP. Peltier performance is often set more by the heat sink than by the device itself.
Design workflow
- Set the heat load, target temperature and ambient temperature.
- Assume a heat-sink resistance and solve together with the hot-side temperature.
- Choose module count and current by trading efficiency against size.
- Check cold-side insulation and condensation protection (when cooling below the dew point).
- Check contact resistance (thermal interface materials) and solder-joint life under thermal cycling.
“More current, higher temperature”
A Peltier-cooled device wasn't reaching its target, so the current was raised — and the temperature went up instead. Investigation showed a small heat sink with the hot side already above 60 °C. More current meant more input power, a hotter hot side, a wider temperature difference and less cooling. With a larger heat sink, the original current reached the target. When cooling falls short, measure the hot-side temperature first.
Common mistakes
Mistakes and fixes
| Mistake | Effect | Fix |
|---|---|---|
| Sizing the heat sink for cooling load | Hot side overheats, cooling fails | Size for load plus input power |
| Assuming more current cools more | Temperature rises instead | Check current for peak cooling |
| Designing to catalogue maxima | Real capacity drops with ΔT | Calculate at operating ΔT |
| Ignoring condensation | Electrical failure | Insulate and seal |
| Ignoring contact resistance | Extra temperature difference | Evaluate interface materials |
I'd like to try related calculations.
Try the Peltier calculator, ZT figure of merit calculator and heat sink designer for device and heat-sink calculations. Related topics include thermoelectrics, heat sink design, thermal interface materials and thermal resistance.
Related Topics
Experience the theory firsthand with the interactive simulator for this field
All Simulators