Typical peltier devices (approx
40mmx40mm) |
Peltier
devices are literally heat pumps, which have two sides;
a hot side, and a cold
side. When a voltage is applied (around 12V), heat
is 'magically' pumped from the cold
side to the hot side
through the semiconductor junction.
Peltier devices have different
power ratings, corresponding to how fast the cold
side is able to cool down an object. Another factor
is generally specified, the delta-T, which is the
maximum thermal difference in temperature between
both sides.
|
I managed
to pick up a standard 80W unit at 65 degrees delta-T
for $11. These things are getting cheaper and cheaper
so picking one up shouldn't be a huge financial burden.
The key
with peltier devices is that they don't function to
specifications unless there is something to help take
the heat out from the 'hot side'. Generally, a heatsink
is useful for this purpose.
The first
time I tried this with a wimpy heatsink and the heatsink
got pretty damn hot and the 'cold side' failed to
cool much at all.
|
Correct peltier setup with
heatsink on 'hot side' |
|
A 65° delta-T means
that if the 'hot side' is at 50 degrees (with a heatsink),
then the absolute minimum temperature achievable on
the 'cold side' is -15 degrees. Thus the cooler you
can keep the 'hot side' at, the colder the 'cold side'
will be.
|
After
much scrounging around, and even contemplating driving
to a pawn shop NYC
to find what I needed, this large heatsink that came
out of an old computer finally made something happen.
I
set this up on a drowsy sunday morning, hence the rubber
bands. Visible in the picture is the heatsink, a high
speed fan mounted to the heatsink, and an SLA battery
for the power source. A 10 cent coin sits on the peltier
ready for freezing... |
|
|
Here is some
frost and chunks of ice forming on the coin
about half a minute after power-up. It is evident from
the frozen water droplets that the peltier is still
cooling relatively slowly - given there is enough time
for condensation to form first before freezing over. |
Here's
another go at it, again with ice chunks forming where
water had condensed onto the block.
Also visible
in this image is the lower left hand side of the coin
beginning to defrost. This is caused by the 'cold
side' warming up as the 'hot side' struggles to dissipate
the heat.
|
|
|
I was curious as to what
would happen if the heatsink was placed in ice-cold
water. Yeah, cheating, I know. But hey - ice crystals
formed almost immediately and this image was taken
within 10 seconds of power up.
No chunks of ice this time.
Notice how the rubber bands are frozen too.
|
Very
self explanatory photo I think... to show the depth
of the ice/frost coating. |
|
|
And an
extreme close up of the ice crystals. Frozen so fast
that they don't even have time to merge and form a smooth
surface. |
| New
peltier setup! The previous setup proved to be a pain
in the arse in getting rid of the massive amounts of
heat from 'hot side' of the peltier.
The
local university decided it would throw away this gargantuan
of a heatsink, so I got it for free - the only price
I had to pay was the labour involved in lugging it back
home.
The
fan below is there to cool the heatsink. |
|
|
The results
are promising! Any water on the surface of the peltier
freezes within a few seconds(!) upon power-up. Here
is a small button cell battery being frozen to death,
and a sultana having the worst day of its life. |
A
closeup of the ice and frost covering the surface of
the button cell battery. The sides are covered in ice
which is the result of condensation beginning to trickle
down just before the whole cell drops to below zero
temperatures and freezes over.
Frost
starts to form out of control everywhere on the peltier
after a while and creates a slight short circuit hazard...
|
|
|
Another
shot from a slightly different angle showing the silver
heatsink paste used to provide thermal conductivity
between the 'hot side' of the peltier (bottom), and
the surface of the heatsink. This stuff is very expensive
and is normally only used for CPU mounting. I only use
it because I got it cheap and it is much more effective
than standard heatsink paste.
Each time
the peltier is turned off, the ice melts dramatically
fast, and the resultant water creeps into the peltier
unit... so be careful to dry it thoroughly before turning
it on again! I should really get around to sealing the
peltier element with silicone or something.
|
The recently
acquired 1-Wire data loggers were put to use to measure
the temperatures reached by the peltier device.
The setup
was as above, with the data logger pressed down against
the surface of the peltier by means of a weight (separated
via plastic insulation to avoid thermal conductivity
between the weight and the data logger). On the graph,
the switch on time corresponds to time = 70.
From the
resultant graph it is immediately obvious that the difference
between using fan cooling is minimal, although this
could of course been due to the angle of windflow.
Lowest temperature
(heatsink only): -5°C
Lowest temperature
(heatsink and fan): -6°C
|
Peltier
temperature vs. time
(Peltier
switched on at 70 seconds, switched off at 470 seconds)
|
Peltier
surface humidity vs. time
(Peltier
switched on at 70 seconds, switched off at 470 seconds) |
This
graph shows relative humidity of the surface of the
peltier, and is somewhat related to the temperature,
as the lower the temperature, the more water is going
to condense onto the peltier element. It is of no surprise
then, that the average humidity of the fan assisted
heatsink setup is slightly higher than the heatsink-only
setup. |
After
a long break and following the acquisition of a new
3 speed 150W industrial fan, the possibilities of revisiting
the peltier project were immediately obvious. The new
setup consists of the heatsink mounted directly on top
of the high powered fan, and the peltier element is
insulated by means of a foam box in order to minimise
the effects of the high velocity winds affecting output
temperature. |
|
Peltier
temperature vs. time
Peltier
switched on at 0 seconds
|
The results
were impressive, with all three fan speeds achieving
peltier temperatures of less than -10°C. The inverse
relationship between wind speed and temperature is clear,
but I suspect there would be less of an effect (perhaps
up to a point) as the wind speed is increased even further.
Lowest temperature
(Fan speed 1): -12°C
Lowest temperature
(Fan speed 2): -13°C
Lowest temperature
(Fan speed 3): -14°C
Also note
that there was never a distinct "minimum"
temperature reached, unlike the previous setup where
temperature would drop to a minimum then start to rise
as the heatsink warmed slowly. Presumably the fan has
such cooling power that it is able to maintain the heatsink
at a constant temperature. Tests on this later showed
the heatsink temperature to rise slowly and stabilise
at around 35°C over time. |
I believe
with proper insulation and heat transferral, the minimum
temperature can be lowered even further. Future setups
may incorporate water cooling, but at the moment I am
still trying to maximise results using only air assisted
cooling. |
|
