MSL Technical Guide 41

Using Peltier Devices for

Temperature Control
Version 1, July 2019

Introduction
Peltier devices are small solid-state heat pumps. They may
be operated in heating or cooling mode and are well suited
to temperature-control systems ranging in power from a
few watts to a few hundred watts. However, Peltier devices
are not as easy to use as engineers might expect, and some
care and understanding is required to avoid pitfalls.
This technical guide describes an electrical-analogue
Figure 2. An example of a single-stage Peltier device.
model of Peltier devices as an aid for the design of small
refrigeration and temperature control systems. The model
is easily incorporated into spreadsheets, where the perfor-
Peltier devices (Figure 2) consist of many, sometimes
mance of a system under different operating conditions can
hundreds of, junctions connected in series and stacked be-
be explored easily. Peltier devices can also be used to gen-
tween two ceramic plates, so that when a current is passed
erate power, but this application will not be discussed. The
through the junctions, heat is moved from one plate to the
discussion is also limited to single-stage Peltier devices
other.
and requires some basic mathematics.
In addition to the Peltier effect, there are two other ther-
mal effects occurring, both of which detract from the per-
Principle of Operation formance of the devices. Firstly, the passage of electrical
The Peltier effect is a thermoelectric effect (involving heat current through any conductor causes Joule (ohmic) heat-
and electricity) that occurs when an electrical current ing proportional to the electrical resistance of the conduc-
passes through a junction between dissimilar electrical tor and the square of the current.
conductors. Figure 1 shows a diagram of a pair of junctions Secondly, once a temperature difference has been es-
with an electrical current flowing through them (remember tablished between the two sides of a Peltier device, heat
that electrons flow in the opposite direction to the current). from the hot side leaks through the conductors to the cold
Now imagine that the electrons in one of the conductors side. This heat flow is proportional to the temperature dif-
(red) are in a high-energy state, while the electrons in the ference between the two sides. Although Peltier devices
second conductor (blue) are in a low-energy state. When can be manufactured from almost any pair of dissimilar
the high-energy electrons cross the junction, they drop into conductors, manufacturers use materials specially de-
the low-energy state and dump the excess energy as heat. signed and selected to optimise the balance between the
Similarly, when the low-energy electrons cross the junc- Peltier effect and the electrical and thermal conductivity.
tion between the two materials and enter the high-energy
state, they must take up energy from their surrounds and The Model
cool the junction. The practical effect of the current flow-
Figure 3 shows an electrical-analogue model of the thermal
ing through the pair of junctions is that heat is moved from
behaviour of a Peltier device, including a heatsink con-
one junction to the other. The direction of the heat move-
nected to the hot side of the device and the system that the
ment can be reversed simply by reversing the direction of
Peltier device is to control connected to the cold side of the
the current.
device. In the electrical-analogue model, the current
sources represent heat sources and the resistors represent
thermal resistances. A thermal resistance relates heat flow
to temperature differences in a similar manner to Ohm’s
law: temperature difference divided by heat flow equals the
thermal resistance. The SI unit for thermal resistance is
K/W (or °C/W). The heat source on the cold side of the
Peltier device is a source of negative heat (cooling). The
heat source on the hot side of the Peltier device is a positive
source that dumps heat into the heatsink.
Figure 1. A single pair of junctions in a Peltier device.

Measurement Standards Laboratory of New Zealand Page 1 of 5

 Qh = sITh + I 2 Re 2 − (Th − Tc ) Rhc . (4)

where:
I is the current through the Peltier device [A],
Re is the electrical resistance of the device [V/A], and
s is the Peltier coefficient for the device [V/K].
Each of equations (3) and (4) has three terms. The first
term in each equation describes the heat moved by the Pel-
Figure 3. Simplified electrical analogue model of the Peltier tier effect, which is proportional to the absolute tempera-
system. ture (hence the need to use the absolute temperature scale),
and the electrical current, I, passing through the device.
The constant of proportionality, s, is the Peltier coefficient.
The parameters in the model are: Note that the sign on the first term for the cold side is op-
• Ta is the ambient temperature [K], posite to that for the hot side (the heat goes in one side and
• Tc is the temperature of the cold side of the Peltier de- out the other side). The equations are completely symmet-
vice [K], ric, so the two sides of the Peltier device can be inter-
• Th is the temperature of the hot side of the Peltier device changed simply by reversing the current.
[K], The second of the terms in equations (3) and (4) is due
• Rh is the thermal resistance of the heatsink [K/W], to the Joule heating, which is the electrical heating occur-
• Rc is the thermal resistance between the controlled ring within the Peltier device. The heat is apportioned
space and ambient [K/W], equally between the hot and cold sides, so that half of the
• Rhc is the internal thermal resistance between the hot Joule heat appears on each side of the device.
and cold sides of the Peltier device [K/W], The third term in each equation is the heat leaking
• Qc is the (usually negative) heat exiting the cold side of through the Peltier device between the hot and cold sides
the Peltier device [W], due to the thermal conductivity of the materials forming
• Qh is the (usually positive) heat exiting the hot side of the conductors and junctions within the device. The heat
the Peltier device [W], leak is given by (Th − Tc ) Rhc , where Rhc is the internal
• QL is the heat generated by the load in the controlled thermal resistance between the two sides. Since the heat
space [W]. leaked from the hot side appears at the cold side, it has the
opposite sign in the two equations. The internal thermal re-
There are two pairs of equations that describe the rela- sistance Rhc turns out to be a major factor limiting the ther-
tionships between the variables in the model: the first pair mal efficiency of Peltier devices because it is very low – in
of equations describes the external effects of the Peltier de- fact, much like that of a good heatsink. This means that
vice on the load and heatsink, while the second pair de- much of the work done by the Peltier effect in moving heat
scribes the internal operation of the Peltier device. from one side of the device to the other is undone by the
Firstly, the hot-side temperature depends on the thermal heat flowing backwards through the thermal resistance.
resistance of the heatsink, the heat dumped there, and the
ambient temperature:
Determining the Model Parameters
Th = Ta + Qh Rh . (1) Most data sheets for Peltier devices describe the device in
terms of five parameters:
There is a similar relationship for the cold-side temperature • ∆Tmax: the maximum temperature difference that can be
(note that Qc is usually negative): achieved across the device.
• Qmax: the maximum heat moved when there is zero tem-
Tc = Ta + (Qc + QL ) Rc . (2) perature difference across the device.
• Imax: the current that produces the maximum tempera-
Note that all temperature variables used in this guide are ture difference ∆Tmax.
expressed in kelvin (i.e., on the absolute temperature scale)
• Vmax: the voltage across the device when ∆Tmax is pro-
for reasons that will become clear shortly.
duced.
The second pair of equations describes the heat pro-
duced by each side of the Peltier device. The heat flux at • Ts: the temperature at which the above properties are
the cold side (usually negative) is specified, usually 300 K (27 °C).
To determine the values for the parameters in our
Qc = − sITc + I 2 Re 2 + (Th − Tc ) Rhc , (3) model, we need to take these values from the device data
sheet, substitute the values and appropriate operating con-
and at the hot side (usually positive) is ditions into equations (3) and (4), and solve the equations.

Measurement Standards Laboratory of New Zealand Page 2 of 5

 Firstly, the maximum heat moved at the cold side is I maxVmax − 2Qmax
determined from equation (3) with Th = Tc at the specified s= . (12)
I max (2Ts + ∆Tmax )
temperature of Ts = 300 K. Hence,
Note that these parameters describe the properties of the
Qmax = − sI max Ts + I max
2
Re 2 , (5) materials from which the Peltier device is made. The pa-
rameters are typically temperature-dependent and not con-
which provides us with an equation that includes Re and s stant as assumed in the model. Therefore, the model is not
as unknowns. Note that Qmax must be negative for it to be exact, but usually it will be sufficient to give a good over-
consistent with the way we have defined the heat sources view of the performance and behaviour of the Peltier de-
in the model (see numerical example below). The maxi- vice in most applications.
mum temperature difference, ∆Tmax, is achieved when Table 1 shows a section of a spreadsheet with the input
Qc = 0. Hence, from equation (3), data and calculated values for a 60 W Peltier device. The
cells in yellow are the numbers taken from the device data
0 = − sI max (Ts − ∆ Tmax ) + I max
2
Re 2 + ∆ Tmax Rhc . (6) sheet, while the model parameters are the values of the de-
vice parameters calculated by the spreadsheet. This exam-
To get the next equation, we need to consider the total ple can be used to check the formulae in your spreadsheet.
power used by the Peltier device; i.e., the sum of equa- Note that the sign of Qmax is negative.
tions (3) and (4):

Q total = sI (Th − Tc ) + I 2 Re . (7) Table 1. A spreadsheet showing the results of equations (10),
(11), and (12), for the model parameters. The cells shaded yel-
This equation can be rewritten by recognising that the total low indicate data from the device datasheet.
device power is equal to the voltage across the device times Data from Datasheet Model Parameters
the current through it; i.e., the voltage across the device is
Ts 303.00 K s 0.05262 V/K
Vtotal = s (Th − Tc ) + IRe . (8) Qmax −60.00 W Rhc 1.865 K/W
Imax 6.10 A Re 2.0027 Ω
This equation has two terms: the last term gives the voltage Vmax 15.90 V
drop across the electrical resistance of the Peltier device,
∆Tmax 70.00 K
while the first term is the Seebeck voltage produced in the
conductors by the temperature gradients between the Pel-
tier junctions. The Seebeck effect, which generates this Investigating Device Performance
voltage, is the same as that responsible for the voltage gen-
Now that the model parameters have been determined, the
erated by thermocouples. When Peltier devices are used in
model can be used to investigate the behaviour of the de-
the power-generation mode with given hot- and cold-side
vice in different situations. First, we must solve equa-
temperatures, this is the voltage generated. When in cool-
tions (1) to (4) to determine the cold- and hot-side temper-
ing mode, the current times this voltage is the thermody-
atures. The cold-side temperature is
namic work done in moving the heat from the cold side of
the Peltier device to the hot side.
2Ta ( Rhc + Rh + 2 Rc − sIRh Rhc )
In order to generate the maximum temperature differ-
ence, the maximum voltage must be applied. Hence + 2QL Rc (2 Rhc + Rh − sIRh Rhc )
+ I 2 Re Rc (2 Rh + Rhc − sIRhc Rh )
Tc = ,(13)
Vmax = s∆Tmax + I max Re . (9) 2  Rhc + Rc + Rh − s 2 I 2 Rc Rhc Rh − sIRhc ( Rh − Rc ) 

We now have three equations (5), (6), and (9), from and the hot-side temperature is
which we can determine the three parameters in the model.
The solutions are 2Ta ( Rhc + Rh + Rc + sIRc Rhc )
+ 2Q L Rc Rh
2 ( I maxVmaxTs + ∆TmaxQmax )
Re = , (10) + I 2 Re Rh ( Rhc + 2 Rc + sIRhc Rc )
2
I max (2Ts + ∆Tmax ) Th = .(14)
2  Rhc + Rc + Rh − s 2 I 2 Rc Rhc Rh − sIRhc ( Rh − Rc ) 
∆Tmax (2Ts + ∆Tmax )
Rhc = − , (11) These equations can be written into a spreadsheet to in-
Qmax (2Ts + ∆Tmax ) + ∆Tmax ( I maxVmax − 2Qmax )
vestigate the effects of different operating conditions, as il-
lustrated below. Using the model in this way is more effi-
and
cient that doing the real experiments with real devices. You

Measurement Standards Laboratory of New Zealand Page 3 of 5

may need to do some fine tuning afterwards, but the with a current of only 2.5 A, much less than the 6 amps
spreadsheet will give you a good head start on the thermal maximum current indicated by the datasheet. Note too, any
design of your system. current higher than 2.5 A causes the cold-face temperature
to increase instead of decrease. If the Peltier device is used
The Heatsink in a feedback system, it is essential to limit the current so
that it never exceeds the value giving the minimum tem-
The heatsink thermal resistance is the most important char-
perature, otherwise the negative feedback will change to
acteristic of any system using Peltier devices. The loss in
positive feedback, causing the control system to run away
performance due to a high heatsink resistance occurs be-
and lock in a hot state at full power.
cause of the thermal feedback within the Peltier device.
The more heat dissipated on the heatsink, the higher the
Cold-Side Heat Load
hot-side temperature becomes, and the greater the amount
of heat leaking back to the cold side through the internal Figure 5 plots the effect of the cold-side heat load (i.e., heat
thermal resistance Rhc. In effect, Rhc is the output resistance dissipated in the cold zone) on the performance of the sys-
of the Peltier heat source (analogous to the output re- tem. The curves are plotted using equation (13) with
sistance of a current source), and just as a current source Rc = 1,000,000 K/W, Rh = 0 W, Ta = 300 K, and a range of
with a low output resistance is a poor current source, Peltier values of QL. The effect of the cold-side heat load is to raise
devices are also poor heat sources. They require heatsinks the curves, almost uniformly, at all currents. Note too that
with a very low thermal resistance to work well. there is a slight increase in the current where the tempera-
Figure 4 plots the cold-side temperature versus current ture is minimised.
for the 60 W example device, for a range of different heat
sink resistances from 0 K/W to 1 K/W. The curves were
generated using equation (13) with Rc = 1,000,000 K/W
(approximating infinity), QL = 0 W, and Ta = 300 K. Note
that the curve for zero heatsink resistance, the bottom curve
in Figure 4, replicates the unrealistic curve shown in most
data sheets.
The internal leakage resistance for the 60 W device is
Rhc = 1.865 K/W, and the heatsink resistance should ideally
be much smaller than this value. The range of heatsink
thermal resistances in the Figure 4 spans a practical range
from low-cost aluminium heatsinks 100 mm × 125 mm ×
50 mm (~1 K/W) to very large heatsinks 200 mm × Figure 5. The effect of cold-side heat load on the system temper-
200 mm × 200 mm with forced-air cooling and heat pipes ature versus current.
(~0.2 K/W). The best heatsinks are liquid cooled or the
heat-pipe heatsinks commonly used to run computer pro- Cold-Side Thermal Resistance
cessors at faster than normal speeds (overclocking).
Figure 6 plots the effect of thermal resistance between the
cold controlled space and ambient. The curves are plotted
using equation (13) with Rc = 0, Rh = 0 W, and Ta = 300 K.
The effect of the thermal resistance is to reduce the effec-
tiveness of the device for both heating (negative current)
and cooling (positive current). You should aim to provide
the best insulation between the cold zone and ambient that
you practically can.

Figure 4. The effect of the heatsink thermal resistance on cold-

side temperature versus current.

Operating Current
Figure 4 shows that when a Peltier device is used with any
realistic heatsink (i.e., Rhc > 0), the minimum temperature
is achieved with a current less than the specified maximum
for the device. When our example device is operated with
Figure 6: The effect of poor insulation around the controlled
a 1 K/W heatsink, the minimum temperature is achieved
space.

Measurement Standards Laboratory of New Zealand Page 4 of 5

Use in Control Systems In Figure 5, showing the effects of cold-side heat load,
the gain remains much the same for all heat loads. In Fig-
Figure 4, which shows the how the cold-side temperature ure 6, showing the effects of cold-side thermal resistance,
depends on the device current, emphasises the importance the gain is reduced by the thermal resistance between the
of limiting the current through the device when it is used load and ambient.
in a control system, in order to avoid accidentally operating Note that in all operating conditions, the controller gain
the system with positive feedback and having it run away is higher when heating than when cooling.
and settle in a locked-up state with the highest drive current
and in a very hot state. Further Reading
One of the important characteristics of the Peltier de-
vice when it is used in a control system is the gain of the MSL Technical Guide 10, “A Simple, Narrow-Band
controller. For the Peltier device, the gain is the rate of Thermistor Thermometer” (available at http://www.meas-
change in cooling with input current, dQc dI . In Figure 4 urement.govt.nz/).
the slope varies a lot depending on whether the system is MSL Technical Guide 20, “Introduction to Temperature
operating above ambient, near ambient, or below ambient. Control” (available at http://www.measurement.govt.nz/).
For a control system operating near ambient temperature,
the gain will be the slope on the curve near zero current. In
Figure 4, the slope is unchanged by the heat sink (although Prepared by D R White.
the range of control currents is very much reduced with a
poor heatsink). If operating below ambient, the heatsink
has a significant effect on the gain.

Contact Details
Postal address: Measurement Standards Laboratory, Callaghan Innovation, PO Box 31-310, Lower Hutt 5040,
New Zealand.
Website: www.measurement.govt.nz
E-mail: [email protected]

The Measurement Standards Laboratory is New Zealand’s national metrology institute, operating under the authority of
the New Zealand Measurement Standards Act 1992. The Measurement Standards Laboratory is part of Callaghan
Innovation, a New Zealand Crown entity.

This work is licensed under a Creative Commons Attribution-No Derivatives 3.0 Licence
(CC BY-ND 3.0 NZ, see https://creativecommons.org/licenses/by-nd/3.0/nz/).

This licence allows for redistribution, commercial and non-commercial, as long as the work is passed
along unchanged and in whole, with appropriate credit.

Measurement Standards Laboratory of New Zealand Page 5 of 5