This article was first published in 2004.
In this introduction we’ll do what journalists never do – well, magazine
journalists anyway. And what’s that? We’ll tell you the punchline before we even
begin. The Peltier-based intercooler water spray covered in this story was
unsuccessful – it didn’t work. Well, not adequately anyway. So why tell the
story? Two reasons: firstly, why it
didn’t work is very interesting, and secondly, it’s also very good background
for those who wonder about using Peltiers for intercooling purposes.
It frequently crops up in web discussion groups: why can’t Peltier coolers be
used to build an intercooler? It seems reasonable enough: Peltier coolers are
designed to work off car-type voltages, in recent times their prices have been
dropping fast, and they do what you want an intercooler to do – remove heat.
(See the breakout for more on these fascinating devices.)
But the short answer regarding Peltier intercoolers is that with the current
level of Peltier technology, it’s not going to happen. A powerful Peltier device
like this one is rated at 80 watts of electrical power. Wile the relationship
between thermal heat movement and electrical input power is not 1:1, the power rating of a turbo car
intercooler (eg the B4 Subaru at 13.4 kilowatts) is so much higher that you’d
need something like 15 Peltiers to do the same job – and that’s ignoring the
very real problems of the actual heat exchange process and the power consumption
from the battery. It’s certainly possible that one day intercooling will head in
this direction, but so far it’s not been viable.
That’s intercooling – but what about cooling just the water that you spray
onto the intercooler core? There’re a few efficiency loss steps along the way
(the Peltier has to act through a heatsink to cool the water which in turn has
to act through the intercooler metalwork to cool the intake air) but taking this
approach has a large advantage.
The advantage is this: water can be cooled over a reasonably long period (eg
5-10 minutes) and this ‘assembled coolness’ then dumped in one hit. This is
potentially very effective because quite often a turbo road car needs a big
power squirt – but then nothing much more for some time. An overtaking move, or
a quick traffic light sprint, as examples.
Sometimes called ThermoElectric Coolers (TEC), Peltier modules consist of a
number of p- and n-type pairs (couples) connected electrically in series and
sandwiched between two ceramic plates. When connected to a DC power source,
current causes heat to move from one side of the TEC to the other. This creates
a hot side and a cold side on the TEC. A typical application exposes the cold
side of the TEC to the object or substance to be cooled and the hot side to a
heatsink which dissipates the heat to the environment. A heat exchanger with
forced air or liquid may be required. (As clever as TECs are, they can't eat
heat - only move it!) - www.melcor.com/faq.htm
Peltier module manufacturer Melcor has plenty more interesting and
detailed design information on their site: www.melcor.com
The First Design
I decided to put the theory to the test and bought two 80W Peltier devices
They cost AUD$18.50 each. In a typical application (eg a portable car fridge)
the Peltier device is sandwiched between two heatsinks, often with a block of aluminium
in there as well to give some clearance to the walls of the fridge. The inner
heatsink transfers the cold to the interior of the fridge (or more accurately,
removes heat from the interior) while the outside heatsink gets very hot and is
cooled by a fan. (Here the inner heatsink is shown – smeared with heat sink
compound to better aid heat transfer.)
So a Peltier device acts as the conduit, moving heat from one face to the
The design was for an interim water storage device that could be plumbed into
the intercooler spray line between the pump (attached to the main reservoir) and
the nozzle. The 250ml of water held in this ‘cooling container’ would be
directly cooled by the Peltier device. When the spray started, the water in the
container would be pressured by the pump and would flow out through the nozzle,
being replaced by more water coming from the reservoir. Of course if the spray
ran continuously, the ‘cool’ water would soon be used up, but in typical
discontinuous use I figured the cold spray would be good for power bursts.
I bought a high quality plastic box and two heatsinks, one a little smaller
than the other. The smaller heat sink mounted inside the box lid, its fins
projecting down into the box volume. In the lid of the box I cut two 40mm square
holes, in which two Peltier coolers sat, their lower surfaces in contact with
the inner heatsink. On top of the box - in contact with the upper surface of the
Peltiers - was an aluminium bar, while the second (larger) heatsink sat on the
strip. (The bar gave enough clearance for the Peltier power supply cables to
exit under the heatsink.) The sandwich assembly was bolted together with two
through-screws and nuts and all contacting surfaces were smeared with heatsink
To see how effective the system would be, I filled the box with water (it
took about 250ml) and then connected the Peltiers to a bench power supply. A
small but powerful fan was aimed at the upper heatsink.
Within a minute or so of power being applied, the upper heatsink was hot to
touch – heat was being drawn out of the water and dissipated through the
heatsink. However, even after 5 minutes, the water inside the box had barely
changed in temp. I then measured the actual voltage that the nominally 12V power
supply was providing and found that under the substantial 160W load it had
dropped to around 9V. Concerned that the lower voltage might massively decrease
the Peltier performance I then moved the test assembly to my car and ran it
straight off the battery, engine running.
This time the upper heatsink got even hotter, but the water inside the box
wasn’t getting any cooler. It was always going to take a certain amount of time
before the water became adequately cold, and I figured that 10-15 minutes was
about the realistic maximum this could take. But after 10 minutes the water was
– if anything – only a few degrees cooler than at the starting point, and after
15 minutes it was much the same.
Despite the Peltiers dragging heat out of the water, not much was happening
to the temp of the water.
Why Didn’t It Work?
There are a few reasons why the system didn’t work. One is the specific heat
capability of water. As we have covered in the past in water/air intercooler
stories, water has a massive specific heat – that is, even a small volume can
absorb a huge amount of heat with very little temperature increase. (That’s why
when you put a saucepan of water on a stove, it still takes quite a long time to
bring the water to boil – even though you might be pouring in a few kilowatts of
power.) And the same characteristic applies when trying to cool it – you can
remove a lot of heat without the temp dropping too much.
But the most important point relates to getting rid of the heat. The external
heatsink has to shed all of the heat that is being drawn from the water, and as
we said above, that’s a LOTof
heat. Despite the use of a larger
heatsink on the ‘air’ side of the cooler, the ability of air to take that heat
is much less than that of water, so in this case the heatsink was working
overtime... all of the time.
Simply, even with a relatively large external heatsink, not enough heat was
being able to be shed to cool the water. But what about using an even larger
external heatsink? As an experiment I bolted a huge 550 x 140 x 13mm alloy plate
straight to the Peltiers. That’s a helluva hefty piece of aluminium! And this
time the system started to work – the water inside the plastic box cooled down
rapidly to a temp of 15 degrees below ambient. Incredibly, even this size of
heatsink still got hot though – it needed fins to better shed the heat.
One point that many people will make is that the electrical load being placed
on the alternator by the Peltiers (180W in this case) will subtract from the
power available to drive the wheels. However, it’s easy to get around that –
simply switch off the Peltiers (using a relay and a boost pressure switch)
whenever the car’s on boost.
The Final System
Nothing daunted, I continued with the development process.
One reason I was happy to continue is that I managed to source a very large
aluminium heatsink. The unit, which is sized at 250 x 185 x 60mm, came from a
scrap metal dealer and cost just AUD$6. This looked like it would be very
effective as the external heatsink, shedding the heat drawn out of the water to
the air. However, its finned-on-both-sides design meant that an aluminium angle
bracket needed to be used to connect a mounting plate on which the Peltiers sat
to the main body of the heatsink.
Other than that, the design was much the same as before – an internal
heatsink projecting into the water with the Peltiers sandwiched between this
heatsink and the main outside mounting plate, which in turn connected to the
large heatsink. In addition I placed two 12V PC fans at the base of the heatsink
(which was positioned vertically), so that convective airflow up past the
heatsink would be encouraged. (By now this was going to be a boot-mounted
design!) In this view, the lid of the plastic box – which comprises the lower
half of the box – is missing.
I filled the box with water and connected power to the Peltiers. But again
the result wasn’t all that effective – the heat transmission path from the
Peltiers to the large heatsink was too high in resistance, despite thick
aluminium being used to connect the two. The aluminium plate on which the
Peltiers sat was very hot, and the angle connecting this to the heatsink was
hot. But the heatsink was only just warm to touch. It seemed that the heatsink
itself had to be in direct thermal contact with the hot side of the Peltiers if
it was to work well.
And without having access to a mill to shave off a whole bunch of fins, that
was too hard to do.
Car Body as the Heatsink?
Quite early in the development process I considered using the whole car body
as a heatsink. However, the Peltiers have to be very well connected thermally to
the heatsink and there no real way of effectively doing that if the car’s body
is being used as the heatsink.
It is probably possible to develop an effective Peltier-based intercooler
water spray. However, it would need a very large heatsink, cool a relatively
small volume of water and consume quite a lot of electrical power. And there’s
certainly no easy and cheap DIY approach that will work well.
At the time of writing Oatley Electronics has available a pre-made 1-litre
insulated water tank with attached Peltier cooler, heatsink and fan. It works
from 12V DC and costs just AUD$37. While this looks perfect for an intercooler
water spray application, we would suggest that based on the size of the fan and
heatsink, it is a low power device that would only cool the water adequately if
run continuously (ie 24 hours a day).