This article was first published in 2008.
Many people see tracking the airflows over the car
- and especially through an underbonnet (or top mount) intercooler - as a black
art. They instinctively know that the dyno room is completely the wrong place to
test underbonnet intercooler airflows, but at that point they’re stymied.
Apart from putting on a bigger core and ever-bigger scoops, what else do you do?
Some people think air director plates under the scoop are needed – but others
see these as a waste of time.
So what do you do if you’ve got a turbo car with
an underbonnet intercooler and want better results?
The most important step is to understand that
airflow doesn’t always occur as pictured.
Plenty of cars use underbonnet intercoolers, with
two different approaches taken to forcing air through the intercoolers when the
car is moving.
The most common is by means of an external bonnet
scoop. As the car moves forward, air flows over the surface of the bonnet. This
air flows in through the open mouth of the scoop and builds a high pressure on
the upper face of the intercooler core. The Impreza WRX is probably the best
known of cars using this approach.
The other approach is to use an air duct
integrated into the underside of the bonnet, making it invisible from the
outside. This duct picks up high pressure air from the front of the car and
again causes the air pressure on the upper face of the intercooler to be
increased. The (old model) Mazda 3 MPS uses this approach.
But these descriptions tell literally only half
the story. Outside air will only flow through the intercooler core if
there is a higher pressure on one face than the other. In other words,
there needs to be a pressure differential across the core before any air
movement through the core will occur. So the fact that there’s a forward-facing
scoop over the intercooler, and that the air pressure builds up on one face (as
described above) is actually not the real point. It’s quite easy to have a
forward-facing scoop that connects to the intercooler, but to still not have any
flow through the intercooler – or to even have a backwards flow, where hot air
from under the bonnet passes out through the intercooler!
Huh? How could that be?
Actual versus Imagined Flows
A previous series that we’ve done – starts at
Undertrays, Spoilers & Bonnet Vents, Part 1 – tells some of the story.
The car in that series was a Nissan Maxima V6 turbo with an added underbonnet
intercooler (as standard, the engine is not intercooled). A large,
forward-facing scoop was placed over the intercooler, with the mouth of the
scoop near the front of the car on one side.
With the car travelling at 80 km/h, the pressures
each side of the core were directly measured. (We’ll cover the techniques used
to do this in a moment.)
These measurements showed that when no engine
compartment undertray was fitted, the pressure on the bottom (engine) side of
the intercooler was 0.1 inches of water higher than the pressure on the
front face of the intercooler. In other words, there was a negative pressure differential - air was flowing from inside the engine bay and out
through the intercooler core.
But with the forward-facing scoop, how could this
be so? In short, the amount of air flowing in through the radiator grille
was sufficient that the pressure within the engine bay was higher than was being
achieved on the upper surface of the intercooler by the action of the scoop. In
other words, the intercooler exit was completely blocked – in fact, more
than blocked... the air was flowing the wrong way!
Various different plastic undertrays were
experimented with and by altering the airflow characteristics in the engine bay,
a positive pressure differential was created, where the pressure on the upper
surface of the intercooler core was higher than the pressure on the lower
surface. Further tweaks to the undertray resulted in the final pressure positive
differential shown in this graph – 0.3 inches of water at 80 km/h.
As you’d expect, intake air temp dropped a lot
(and the engine coolant radiators also worked better).
In a car with a factory-fitted underbonnet
intercooler, you can be fairly confident that the intercooler core has at least
a small positive pressure differential across it, with air from outside flowing
through the intercooler.
But that’s about all you can be confident of.
the road speed at which that airflow starts to
whether the action of the electric radiator and
air con condenser fans influence intercooler flows
how airflow grows (or doesn’t grow) with speed
the maximum magnitude of the airflow
the even-ness of that airflow across the width and
depth of the intercooler core
Each of these is critical in optimising
This discussion of pressures all sounds very
scientific – “I am sure he’s right, but what the hell?” However, when I tell you
that it’s very easy to actually measure these pressures on a moving car so you
can see what’s really going on, it all becomes a heap more relevant.
There are two instruments that can be used to
measure these pressure variations. One is a Magnehelic gauge and the other, a
manometer. (You can’t use a normal pressure gauge like a turbo boost gauge
because the pressures are very small.)
Magnehelic gauges are made by the US company,
Dwyer. They are designed to measure both positive and negative pressures, and so
have two measuring ports. By using both ports simultaneously it’s easy to
measure pressure differentials – just what is wanted in this application.
Magnehelic gauges can be bought new from Dwyer, or
alternatively, secondhand. eBay is a good way of buying these gauges very
cheaply – expect to pay about US$15-25 for one.
When buying a Magnehelic gauge, select a gauge
that measures up to a maximum of about 3 inches of water. (The 3-inch gauge lets
you use it in other applications as well – see below. If you intend using it
purely for aerodynamic work, buy a 0-1 inch gauge like the one shown here.)
gauges are extremely useful in car modification. They can be additionally used
to measure flow restriction throughout the intake system, including pressure
drops across intercoolers and the air filter. For more on these techniques, do
an AutoSpeed site search under ‘Magnehelic’.
Instead of a Magnehelic gauge you can use a water
manometer. A manometer simply consists of a U-shaped clear plastic tube, partly
filled with a liquid (usually water with food colouring in it). You can easily
make your own by using some plastic hose and a plywood or particle board
Each arm of the manometer is connected to the
pressures being measured. The fluid in the manometer then moves in response to
this pressure difference – the more it moves, the greater the pressure
difference. The actual pressure change can be indicated by measuring the
difference in height of the two fluid columns. For example, if their levels are
vertically 1 inch apart, you are measuring a pressure differential of 1 inch of
To make the manometer more sensitive, you can
incline it at a fixed angle. If the manometer is angled at 30 degrees from the
horizontal, a difference in level of 1 inch (measured along the tubes) becomes
an actual ‘inches of water’ measurement of 0.5 inches. In this way, very small
pressure differences can be easily read off, even in a moving car. (Of course,
you should use an assistant to read the manometer.)
Note that while I have described manometer
measurements in inches of water, it’s usually not worth making the measurements
in actual units – it’s a lot easier to just put arbitrary makings on the
manometer backing board so you can see relative changes.
The only downside of the home-built manometer is
that its orientation must be kept fixed (eg vertically or at a constant angle)
and very small pressure differences are hard to measure.
The purchase of a Magnehelic gauge, or the making
of a manometer, might seem like a lot of stuffing around. It isn’t.
More on Manometers
A home-built manometer can be a very sensitive
instrument, capable of showing pressure differences of just 0.01 psi. So despite
the simplicity of the instrument, don’t think for a moment that it is a poor
The first thing to do with your manometer or
Magnehelic gauge is to find out what’s actually happening! Run a pressure tube
from one arm of the manometer to the front face of the intercooler. Secure the
end of the open tube in place so that you know where it is measuring pressure
(eg use some masking tape). Run the tube that connects to the other arm of the
manometer to the other side of the intercooler core.
If you’re using a Magnehelic gauge, connect the
tubes to the low and high pressure ports on the instrument, the high pressure
port connecting to the tube that goes to the upper face of the intercooler. (If
you find there’s a negative pressure differential, you’ll be able to see the
needle swing a little the wrong way before it comes up against the stop. If
that’s the case, you’ve got problems! Reverse the connections to the gauge to
see how bad the problem is...)
Make sure that the tubes you are running around
the engine bay cannot foul the throttle or get tangled in a rotating fan.
Also ensure that the hoses are not crimped closed when the bonnet is shut
(detach each tube from the manometer or Magnehelic gauge and check you can suck
Pick a speed that you can safely (and legally)
achieve in a nearby test area, and do all testing in this area, in the one
direction and at the same speed. A practical minimum is 80 km/h. The reason you
do it at the same speed is obvious, but why in the same area? If you watch the
gauge, you’ll see that the instrument is sensitive enough to detect changes in
ambient airflows caused as you pass the sheltering affect of buildings, tress,
etc. These changes are particularly noticeable when there’s a breeze blowing. If
you change areas mid-way through testing, you could easily think that the
flickering of the needle is due to changes you’ve made, rather than the test
In the same way, you’ll also see differing
behaviour of the test gauge when following other vehicles. Yep, we’re talking a
very sensitive gauge able to detect real-world changes!
have a hugely important place in tuning and checking power outputs. But not even
the most enthusiastic dyno room operator in the world wouldn’t suggest that a
dyno replicates exactly what occurs on the road. In the case of aerodynamic
flows through radiators, oil coolers and intercoolers, a typical dyno room is
simply nothing like the on-road reality.
are often tested with the bonnet open, making a radical difference to the way
air gets to an underbonnet intercooler. And even with the bonnet closed, a dyno
room is not a proper wind tunnel, an environment that’s needed to statically
test cooling air flows. Some dyno rooms use a dedicated fan or force-air ‘sock’
(pictured) to cool an underbonnet intercooler. This will certainly keep the
intercooler working well – but the on-road results may be completely
Here’s what was able to be found in just 5 minutes
of on-road testing.
The car was a Peugeot 405 SRDT turbo diesel that
uses a duct integrated into the underbonnet sound insulator to feed the
intercooler positioned towards the rear of the engine bay. Where the duct picks
up air from isn’t all that clear, but it appears to gather air from the gap
between the bonnet and its locking platform. However, as indicated above, it’s
the pressure differential that matters, so even if the feed duct is
pretty poor, perhaps the engine bay is optimised to create a low pressure below
The pressure probes were placed in the middle of
the top and bottom faces of the intercooler. At a test speed of 100 km/h, there
was a pressure on the top surface of the intercooler of positive 0.4 – 0.5
inches of water. (The 0.1 fluctuation being caused by wind gusts and the
presence of other vehicles.) Under the intercooler, at the same speed and in the
same conditions, there was a pressure of 0.4 inches of water. (The under-bonnet
pressure fluctuates less as wind gusts and the presence of other vehicles have
Now that means the difference in pressure above
and below the intercooler was just zero to 0.1 inches of water. It doesn’t sound
like much of a pressure difference - and it sure isn’t. As a comparison, the
pressure difference across the radiator / air con condenser at the same speed
was a relatively constant 0.25 inches of water – 2.5 times as much.
To put this another way: at speed, the airflow
through the underbonnet intercooler was terrible. Therefore, installing larger
intercooler core would have achieved little – in this car, the first step in
improving intercooling efficiency would need to be the creation of a greater
pressure on top of the core (eg by an external bonnet scoop) or the reduction in
pressure under the core (eg by experimenting with different shaped undertrays).
After the pressure differential had been improved, the next step would be to measure the pressure differences across the face of the intercooler. Guides could then be made to make the flow through the core even across its full area.
With readily obtainable (and potentially very
cheap) instruments and some simple on-road testing, it is possible to directly
measure the pressures affecting intercooler airflows. With that information
available, making effective intercooling improvements becomes much simpler. In
short: don’t waste your money making random changes when it’s so easy to find
out exactly what is going on.