This is only a preview of the August 1999 issue of Silicon Chip. You can view 38 of the 96 pages in the full issue, including the advertisments. For full access, purchase the issue for $10.00 or subscribe for access to the latest issues. Items relevant to "Remote Modem Controller":
Items relevant to "Daytime Runnings Lights For Cars":
Items relevant to "Build A PC Monitor Checker":
Items relevant to "An XYZ Table With Stepper Motor Control; Pt.4":
Items relevant to "Making Negatives From Positives":
Articles in this series:
Purchase a printed copy of this issue for $10.00. |
Looking for something to control temperature accurately
and easily? This switching temperature controller can either
heat or cool and can hold a temperature constant. Best of all,
it’s cheap and easy to build.
There are many processes which require temperatures to be closely maintained, from film and
photo developing through to home brewing and
even egg incubation. When temperatures change,
things go haywire: colours change, the brew goes
off, chooks cook . . .
It’s fairly easy to control temperature when things
get too cold – simply heat them up until they get to
the right temperature and turn off the heater.
It’s not quite so easy if things get too hot and require cooling. But this controller can do either – heat
or cool – depending on your application.
SWITCHING
TEMPERATURE
CONTROLLER
Heats OR
Cools
Design by Branco Justic*
Article by Ross Tester
54 Silicon Chip
T
he heater can be any standard
resistive heating device such
as a jug or electric heater element, an incandescent globe or even
a resistor. The specifications depend
on the application: more on this anon.
Cooling, on the other hand, is
done via a Peltier-effect device. For
those who haven’t come across these
before, see the separate panel for an
explanation on what they are and
what they do.
Suffice to say at this stage they are
a solid-state device which absorbs or
gives off heat when a current passes
through their junction.
Normally, designers of semiconductors go to great lengths to minimise
the effect. However, in a Peltier-effect
device the action is exploited. The heat
can only be absorbed from, or released
to, the area surrounding the device. So
the device either cools, or heats, the
surrounding area.
Provided certain precautions are
taken, they can be quite effective
coolers or heaters. They do require
significant current (several amps) but
typically can raise or lower the temperature by 50°C or more.
There are three Peltier-effect devices
specified for this project; you choose
which one you want. They are rated at
42, 60 and 75W and draw 4A, 6A and
8A respectively for a ∆T(or difference
in temperature between the two sides
of the device) of 65°C.
Unlike most temperature controllers
which simply switch a heater on or
off to maintain temperature, this controller switches the heater or cooler by
varying the duty cycle.
This form of control is not only very
accurate; in this application it’s also
a requirement of the Peltier device
which must be switched on and off
at a minimum of 2kHz. Repeatedly
switching on and off DC would result
in mechanical stress to the device and
its possible damage or destruction.
That’s not to say you cannot use a
Peltier device on DC. If the device is
turned on and left on for relatively
long periods, DC is fine.
It’s only when used in a temperature
control application where the device is
switched on and off many times over
relatively short periods to maintain a
constant temperature that mechanical
stress really becomes a problem.
And while on the subject of Peltier
devices, there is nothing to stop you
using one as a heater, if you wish. But
A typical Peltier-effect device. Actual size is 40mm square and about 4mm
thick. When connected to power, one side of the device becomes about 65°C
warmer than the other. This can be used for heating or cooling.
given their cost and the low cost of
a resistive element, we know which
we’d prefer!
The circuit
Operation is most easily understood
if you break the circuit down to its basic functions. Fig. 1 shows the circuit
diagram for the controller.
Transistors Q1, Q2 and Q3, along
with ZD1 and associated components,
form a series voltage regulator supplying a reference voltage of about 7.5V to
op amp IC1. Q1 and Q2 are in parallel
and are both driven by Q3, effectively
forming a Darlington transistor. The
stable voltage at the emitters of Q1
and Q2 is important because it gives
the controller its accuracy.
Temperature setting and sensing is
performed by a preset pot and thermistor connected to the input of IC1a,
one of the four op amps in an LM324
quad package.
The voltage at the inverting input
(pin 13) is set by the preset pot and
held stable by the 10µF capacitor. The
voltage at the non-inverting input,
though, varies with temperature due
to TH1, a negative-temperature-coefficient (NTC) thermistor. While
nominally 68Ω in resistance, the NTC
thermistor decreases its resistance
with an increase in temperature (and
conversely, of course, increases its
resistance with a decrease in temperature).
Therefore, if the temperature rises,
the thermistor’s resistance falls and the
voltage at pin 12 of the op amp will
drop slightly. If the temperature falls,
the voltage will rise.
The op amp has a gain of roughly
221, set (mostly) by the 220kΩ and
1kΩ resistors and the slight change in
voltage at the input results in a much
larger change in voltage at the output.
For example, if the voltage rises just
10mV at the input, the output voltage
will rise by more than 2V.
This voltage is applied to the pin
9 inverting input of IC1c and to the
non-inverting input of IC1d, pin 5.
You will note, though, that there is
another op amp in the circuit, IC1b.
It is connected as a sawtooth oscillator, with an output voltage varying
between about 1/3 and 2/3 of the rail
voltage at about 2.2kHz.
When power is applied, the + input
of IC1b is held at nearly 4V via the
voltage divider (R8 & R9) across the
regulated supply. C6 is discharged but
immediately starts to charge via R11.
When C6 reaches the input threshold
voltage of the of IC1b, it discharges
via R11 and the whole process begins
again. This sawtooth waveform is
applied to IC1c and IC1d.
IC1c and IC1d are comparators – that
is, they compare the voltage between
their + and - inputs and turn their high
or low accordingly. If the voltage is
higher on the + input than the - input,
the output goes low. If it is higher on
the - input than the + input, the output
goes high.
Perhaps this is most easily explained by referring to Fig. 2.
The output of each comparator then
is a pulse waveform at 2.2kHz with a
duty cycle (or on to off ratio) which
is in inverse proportion to the output
voltage of IC1a.
Depending on whether heating or
cooling is required, this waveform is
used to switch Mosfet Q5 or Q6.
A short duty cycle means very little
power is applied to the Mosfet gate
while a long duty cycle means it is
being powered most of the time. Hence
it cools (or heats) for most of the time.
In the cooling circuit (IC1c), a green
LED (LED2) connected to ground gives
a visual indication of the degree of
cooling. Even though the circuit is beAUGUST 1999 55
NOTE: INSTALL
EITHER HEATER
OR COOLER BUT
NOT BOTH.
ing driven at 2.2kHz (essential for the Peltier-effect
device) you cannot see the LED turning on and off
this quickly.
The heating circuit is slightly more complex, due
in part to ensuring that the gate of the P-channel
Mosfet (Q5) is not over-driven; however, it operates
in much the same way – the main difference being
it is opposite in effect. The 9.1V zener and series
diode ensure that the gate cannot be taken more
than about 10V below the source, when transistor
Q4 is turned on.
A red LED (LED3) in series with the base of Q4
shows the degree of heating.
One area not yet mentioned is the power supply. This depends to a large extent on the amount
of heating or cooling required – naturally, this is
limited when you use a 12V supply.
PULSE-WIDTH MODULATION EXPLAINED
Op amp 1 is connected as an oscillator, producing
a sawtooth waveform across the capacitor. This is
connected to one of the inputs of op amp 2. The
other input has its voltage fixed at a certain level by
the voltage divider across the supply.
As the sawtooth waveform voltage rises, it reaches
this threshold voltage and the op amp output goes
high until the sawtooth waveform voltage again falls
below the threshold.
If the threshold voltage is high, op amp 2’s output is high for a very short period compared to its
low-time each cycle. If the threshold voltage is low,
the op amp output is high for a significantly greater
length of the cycle.
The difference between high and low time is called
the “duty cycle”.
OP AMP 1
OP AMP 2
Fig. 1: the temperature controller is capable of either heating or
cooling, depending on which device is installed.
Fig. 2 (right): how pulse-width modulation works. At top is a
simplified circuit which you can see corresponds to IC1b and IC1d
in the circuit above. Below are the waveforms showing the inputs
and the output for a high voltage and a low voltage. The duty
cycle, or on time to off time, is in inverse proportion to the input.
56 Silicon Chip
Parts List
1 PC board, 114 x 77mm
1 plastic case#, with label,
85 x 120 x 28mm
1 14-pin IC socket
1 U-shaped heatsink,
32 x 28 x 13mm
2 3mm x 10mm screws & nuts
4 lengths figure 8 cable (see text)
Virtually the same size as the finished project, this photo shows how and where
all components are placed. Note the electrolytic capacitor at the bottom of the
PC board – it is a PC mounting type but is mounted lying down. We have shown
both Mosfets & heatsinks installed for clarity: normally there would be only one.
The circuit as shown is suitable for
supplies up to 50V or so with only
one resistor change (R15). The voltage ratings of C3 and C6 should also
reflect the higher supply voltage – they
should be at least 30% and preferably
about 50% higher than the supply. As
with most circuits, you can use higher
voltage rated capacitors if you wish
but these tend to be more expensive.
Regardless of the supply voltage,
it needs to be fairly well filtered. Remember, too that a heating element or
Peltier device will each draw significant current – quite a few amps, in fact.
The switching Mosfets (Q5 & Q6)
are both rated at 12A with a maximum dissipation of 88W. The heater,
or P-channel Mosfet could therefore
be used to control loads up to 1kW
with adequate heatsinking (certainly
much larger than the heatsinks specified). For even higher loads, higher
rated Mosfets could be used or even
paralleled.
A heating element of about 4Ω and
a 12V supply would be acceptable
with the heatsinks supplied. Larger
heatsinks would allow a 2Ω element
(72W). With a higher supply voltage,
much higher load powers can be produced while maintaining the same dissipation in the Mosfet. A 50V supply
and a 16.6Ω heater element would be
about 150W; an 8.8Ω load would be
about 300W.
In practice, a 36W heating load
(12V<at>3A) would produce acceptable
heat dissipation from Q5, mounted on
the PC board and using the heatsinks
specified.
What type of heating element?
That’s up to you: series or parallel
combinations of low voltage light
globes are one idea.
Or perhaps you could use an electric jug element stretched out to full
length and cut to a suitable length. A
1kW jug element is about 60Ω in water
– cut in half (30Ω each) and twisted
Resistor Colour Codes
Value
220kΩ
100kΩ
47kΩ
10kΩ
2.2kΩ
1kΩ
680Ω
10Ω
4-Band Code (1%)
red red yellow brown
brown black yellow brown
yellow violet orange brown
brown black orange brown
red red red brown
brown black red brown
blue grey brown brown
brown black black brown
5-Band Code (1%)
red red black orange brown
brown black black orange brown
yellow violet black red brown
brown black black red brown
red red black red brown
brown black black brown brown
blue grey black black brown
brown black black gold brown
Semiconductors
1 LM324 quad op amp (IC1)
4 2N5551 NPN transistors
(Q1, Q2, Q3, Q4)
1 Power Mosfet – either
IRF9530 P-channel (Q5) or
BUK453 N-channel (Q6)
1 GIG power diode (D1)
1 1N4148 signal diode (D2)
2 9.1V zener diodes (ZD1, ZD2)
1 4mm yellow LED (LED1)
1 4mm green LED (LED2)
1 4mm red LED (LED3)
1 Peltier-effect device (see panel)
Capacitors
1 1000µF 25VW# electrolytic (C3)
1 220µF 16VW electrolytic (C4)
1 100µF 25VW# electrolytic (C6)
2 10µF 16VW electrolytic (C1,C2)
1 0.1µF 16VW ceramic or
polyester (C5)
1 .0022µF 16VW ceramic or
polyester (C6)
Resistors (0.25W, 1%)
2 10Ω (R2, R3)
2 680Ω (R1, R16)
1 1kΩ (R6)
6 2.2kΩ (R4, R5, R12, R13,
R15, R17)
1 10kΩ (R10)
3 47kΩ (R8, R9, R11)
1 100kΩ (R14)
1 220kΩ (R7)
1 100Ω horizontal trimpot (VR1)
1 68Ω NTC thermistor (TH1)
A kit, not including Peltier device, is
available from Oatley Electronics for
$15 plus p&p. #Some components in
the Oatley kit may be recycled from
existing equipment.
Capacitor Codes
Value
IEC Code EIA Code
0.1µF
100n
104
.0022µF 2n2
222
AUGUST 1999 57
together would give 15Ω;
cut in quarters (15Ω) and
all twisted together would
give about 4Ω, and so on.
The cooler, or N-channel, Mosfet should be more
than adequate to handle
any of the specified Peltier
devices. If you want to use
more Peltier devices (in
parallel) you will probably
need better heatsink-ing
and perhaps a higher rated
Mosfet as well.
Fig. 3: the PC board component overlay. Compare this to the
photograph when assembling the board and you shouldn’t
have any problems. Again, both Mosfets are shown installed –
you choose the one you want for heating or cooling.
Construction
All components with the
exception of the thermistor
(TH1) are mounted on a
PC board measuring 114
x 77mm. This is designed
to fit into a small plastic
case measuring 120 x 85
x 32mm.
The cases supplied in
the kit are recovered from
surplus stock so are not
new but still perfect for the job. A label
fixes to the front of the case with the
power, cool and heat LEDs showing
through. This label is printed on paper
and will need some covering to protect
it. (We use adhesive plastic).
Begin construction by checking
the PC board pattern for any obvious
defects. If so, either correct or replace
the board. There are six holes on the
PC board which may need to be enlarged – the four mounting holes (in
the corners) all need to be drilled out
to 5.5mm (7/32in) while the two holes
for the Mosfets (in the middle of the
large copper areas) should be 3mm
(1/8in).
Start by inserting all resistors in
their appropriate positions, soldering
as you go. The three links on the board
can be made from cut-off resistor
pigtails.
There are seven capacitors to be
inserted, of which all but two are polarised electrolytics. One of these, the
100µF electrolytic (C6), is a PC board
type (ie, both leads emerge from the
same end) but is actually mounted lying down on the board. A dab of super
glue or silicone sealant underneath it
would help keep it in place.
If you need to fit a higher voltage
rated capacitor here (which will
normally be larger), there is plenty of
room to do so.
Next mount all the small semicon58 Silicon Chip
ductors, taking special care with the
diodes ZD1, ZD2 and D2. Sometimes
they look almost identical to the naked
eye – you may need a magnifying glass
to properly identify them.
Fortunately the power diode, D1,
normally looks quite different!
Solder in the pot (VR1) and the IC
socket but don’t insert the IC just yet.
Then solder in the three LEDs so that
their tops are 25mm above the surface
of the PC board. The yellow LED is
LED1 (power), the green LED2 (cool)
and the red LED3 (heat).
The last component to mount is the
appropriate Mosfet, Q5 or Q6. Again,
these look virtually identical so be
careful. It mounts flat onto its heatsinks with the legs bent down. Before
mounting, hold its three legs with a
pair of needle-nose pliers and bend
the ends of the legs down 90°, 5mm
away from the Mosfet body.
Check a second time that you have
the right Mosfet in the right spot: the
BUK453 is for cooling, the IRF9530
is for heating. Before soldering, slip
the heatsink underneath and secure
both the heatsink and Mosfet with
3mm screws and nuts. No insulation
is necessary between the Mosfet and
heatsink but a small dob of heat transfer compound wouldn’t go astray.
You could, of course, install both
Mosfets and install either the heater
or cooler (but not both). Conversely,
if you will only ever require cooling
(or heating), all components after IC1d
(or IC1c) could be left out.
Solder in a suitable length of figure-8
cable (or two individual wires) for the
thermistor, the heating element and
the Peltier cooler, along with suitable
red and black wires for power connection. Ensure that the cables have
a high enough current rating to cope
with the current drawn.
To complete the PC board, insert the
LM324 IC into its socket, making sure
it is the right way around.
Put the project aside for a while.
Enjoy a cup of coffee before you check
over all your component placement
and soldering.
Checking it out
Don’t connect your Peltier cooler or
the heating element just yet. However,
you will need to connect the thermistor to its leads.
Apply power and confirm that
the yellow LED comes on. Measure
the voltage across C4 – it should be
around 8V – and if you have either an
oscilloscope or frequency meter, check
that there is a 2.2kHz output from pin
1 of IC1b. You can also check that the
heating and cooling LEDs come on as
you vary VR1 over its travel.
If everything checks out OK, turn
off and connect the heating element
or Peltier device to their appropriate
leads. Note that the heating element
should not be polarised but the Peltier
device is: the black lead connects to
the Mosfet drain for correct use.
Now you can check that the appropriate devices really do heat or cool
as they should. You will probably find
that it takes a lot longer for a Peltier
device to cool than a heating element
to heat – that’s the nature of the beast.
Finishing off
As mentioned before, the case
supplied with the kit was intended
for another device. It has a number
of holes and cutouts down one side
which are handy to take the external
leads through.
You will need to drill three 4mm
holes through the lid of the case (and
the label) for the three LEDs to poke
through. It’s easiest to do this with
the label fixed to the case – we used
spray adhesive. The label itself might
need some protection – we use plastic
contact on our projects (see the article
in the April 1999 issue).
With the LEDs soldered in place as
noted above, they should just poke
through the holes in the front panel
when the PC board is mounted in
the case lid. The board sits on small
rebates in the case mounting posts and
does not require any further securing.
Take all of the external wiring
through any suitable holes in the
side of the case and pop on the bottom, securing it with the four screws
provided.
The thermistor needs to be mounted
in very close contact with the item
being temperature controlled but away
from the Peltier device.
If it’s a liquid, ideally the thermistor
needs to be actually immersed in it but
this is often impractical or dangerous
WHICH PELTIER DEVICE?
As well as the kit of parts, Oatley
Electronics currently have three Peltiereffect devices available to suit this project.
All measure 40mm x 40mm and have a
∆T of 65°.
4 Amp – Qmax 42W $25.00
6 Amp – Qmax 60W $27.50
8 Amp – Qmax 75W $30.00
Contact Oatley Electronics on (02) 9584
3563, Fax (02) 9584 3561 or email oatley<at>world.net (or visit their website,
www.oatleyelectronics.com)
* Branco Justic is the Manager of Oatley
Electronics.
(the metal leads could contaminate or
be damaged by the liquid). The thermistor could be “potted” for protection
but this could inhibit its ability to
detect temperature changes. This part
SC
is left to you!
WHAT IS A PELTIER-EFFECT DEVICE?
The “Peltier effect” occurs when current flows across
the junction of two dissimilar metals or semiconductors.
In one direction, heat is absorbed into the junction; in the
other direction, heat is given off. This effect can be used
to make a solid-state heater or cooler. They are usually
called Peltier-effect devices or Peltier devices but you may
see them referred to as thermoelectric modules.
A typical Peltier device is composed a number of P-type
and N-type Bismuth Telluride dice “sandwiched” between
two ceramic plates. While both P-type and N-type materials
are alloys of Bismuth and Tellurium, both have different free
electron densities at the same temperature. P-type dice
are composed of material having a deficiency of electrons
while N-type has an excess of electrons.
As current flows through the module it attempts to establish a new
equilibrium within the materials. The
current treats the P-type material as
a hot junction needing to be cooled
and the N-type as a cold junction
needing to be heated.
Since the material is actually at the
same temperature, the result is that
the hot side becomes hotter while the
cold side becomes colder.
Typical Peltier devices draw between 4A and 10A <at> 12V but there are “industrial” types
drawing 100A or more.
In a resistive load, the heat created is proportional to
the square of the current applied (I2R). In a Peltier device,
the heat created is actually proportional to the current
because the flow of current is working in two directions.
Therefore, the total heat ejected by the module is the sum
of the current times the voltage plus the heat being pumped
through the cold side.
Typically, the difference between hot and cold sides can
be 65°C or more. The ability to add or remove heat is mainly
a function of the current-handling capability of the dice.
With no moving parts, Peltier devices are rugged, reliable
and quiet. They are typically 40 x 40mm square or smaller
and approximately 4 mm thick. The industry standard mean
time between failures is around 200,000 hours or over 20
years for modules left in the cooling mode.
While not polarised in the true sense, most devices have
a red and black lead attached, signifying the positive and
negative connection. The convention is that with the device
lying flat and the leads pointing towards you with the red on
the right side, the lower plate is the “hot” side.
Reversing the power connections has no effect except
for swapping which of the two plates
becomes the “hot” side.
The Peltier device works as a heat
pump. In a cooling application it takes
heat from the surrounding area (or
more correctly anything in intimate
contact with the cold side) and passes
it through to the hot side. Normally the
hot side is itself thermally bonded to
a heatsink, often fan-cooled, to disperse the heat into the atmosphere.
Because the two ceramic plates of
the device are bonded together and one side expands as
it gets hot while the other contracts as it gets cold, thermal
stresses occur. If cycled on and off too often, damage or
failure may occur.
For this reason, where Peltier devices are to be turned on
and off repeatedly, they are fed with a pulse-width modulated
waveform instead of DC.
To finish, some trivia: heat one side of a Peltier device and
you’ll generate a tiny electric current – the “Seebeck” effect.
AUGUST 1999 59
|