This article was first published in 2008.
It’s often talked about but seldom actually done – putting together a new
airbox for the airfilter. The need for a new airbox can come about for a number
of reasons – most commonly because the factory airbox has been ditched in favour
of a pod filter, but sometimes also because other changes in the engine bay have
resulted in the standard filter box having to be turfed. As most people know,
the idea of leaving an exposed filter inside the engine bay is a definite no-go,
primarily because the engine then breathes hot air, which costs power through
more retarded ignition timing and lower density.
So an airbox is a must if you have a cone or other non-boxed filter under the
bonnet – but how do you make it?
The Guinea Pig
The guinea pig car was my Maxima V6 Turbo. The standard filterbox had
initially been modified by having most of its lid removed, with the air fed to
the exposed filter through a hole in the bonnet which was sealed to the filter
box by foam rubber strips. This modification had decreased the measured pressure
drop of the entire pre-turbo intake system from 30 inches of water to 20 inches
of water – an excellent flow improvement. Intake air temp had also marginally
dropped. However, for other reasons, the airbox as a whole had to be removed,
being replaced with two oiled cotton pod filters.
Interestingly, running these exposed twinned filters (still with the same
bonnet opening but without the foam rubber sealing strips) resulted in measured
intake air temps rising massively – from 30 degrees C to something like 50
degrees C on a 20 degree day! Note that was the temp measured after the
intercooler, which was working just as it was before...
The location of the exposed filters (deep inside the engine bay, just in
front of the strut tower) meant that they were inhaling very hot air. In fact,
popping the bonnet and feeling the temps of the filters confirmed that they were
very hot – hotter after the car had been moving only slowly, but still hot even
after travelling at faster speeds.
Simply, they needed to be mounted in an airbox.
An airbox has three main purposes:
1) To seal filters off from the engine bay so that air flowing
through them comes from outside the engine bay, where it is cooler
2) To allow a low restriction flow into the filter(s)
3) To allow a low restriction flow from the airbox into
the intake (eg the airflow meter)
Looking at the first point, what is more important than anything else is that
the airbox is airtight, or very nearly so. It’s easy to forget that under the
bonnet of a car there lots of hot air rushing around – air that has come through
the front-mounted heat exchangers (radiator, air con condenser, possibly an
intercooler and oil cooler), air that has been heated by contact with the
exhaust manifold(s), air that has passed the hot head and hot block. In a fully
warmed-up car it’s not uncommon to measure temps of 70 or 80 degrees C under the
bonnet – hotter in traffic or when the ambient temp is high. If the air flowing
into the airbox can possibly source itself from the engine bay, that’s often
what it will tend to do – so you don’t want leaks in the airbox.
Secondly, there has to be a plentiful supply of air from outside of the
engine bay and into the airbox. Simply, the ducts that supply this air cannot be
oversized. For example, on engines developing 250kW or so, even two three-inch
ducts are not oversized – a single 6-inch duct is even better! (However, more
important is the duct cross-sectional area – see the ‘Intake Area’ breakout
box.) Furthermore, the mouth of this duct should either be forward-facing or
positioned in an area of high aerodynamic pressure, and especially if it is
smaller than optimal, a bellmouth should be formed on the leading edge.
Finally, the air should be able to flow easily from the airbox and into the
engine. Most pod filters are designed to flow well when suspended in space, but
if for example you’re using a flat panel filter inside the box, a bellmouth must
be used on the exit duct.
Should an airbox be made out of insulating materials to try to exclude heat?
The answer to that is yes – and no. There are three ways in which heat can be
moved: conduction, radiation and convection.
Conduction occurs when surfaces are in contact, so in the case of our airbox,
if the bodywork to which it is attached is hot, that heat with be passed on to
the box. Making the airbox from an insulator (eg plastic) will reduce this heat
transfer, but conductive heat transfer is a relatively minor source of hot
Much more important is radiation. This is heat transferred by means of energy
waves – eg light from the Sun, or more relevantly, heat from a glowing exhaust
manifold. Any opaque substance will stop heat radiation – think of shading your
face from sunlight with a sheet of metal. Radiative heat transfer is very
important under the bonnet.
Convectional heat transfer is caused by the movement of a fluid – eg air.
Again, this is very important in the case of an airbox, but making the box
airtight will prevent most convectional heat transfer to the inside of the
airbox. (Although there will be some conduction through the walls of the box as
the convective air currents come into contact with them.)
So while an airbox made from an insulator will give best performance, in
actual fact even a metal airbox will provide almost all of the benefits. And the
best of both worlds? A metal box lined with insulating foam...
Building the Airbox
The trouble with airboxes is that they’re very rarely just a box. Instead,
they’re likely to need to conform to the shape of the inner guard (fender), to
work themselves around a battery or air-conditioner receiver/dryer, and to have
an opening for the airflow meter exit duct. So the idea that you simply fold up
a box out of sheet metal, place the cone filter inside and you’re done – well,
that approach is found only in articles where the author hasn’t actually had to
build an airbox and fit it in the car!
In the case of the Maxima, clearances were extremely tight – firstly there
were two cone airfilters, not just the one that would normally be used if the
factory airbox was removed; and secondly, an underbonnet intercooler was also in
a non-standard place... being positioned where the battery once sat. (For more on
these changes see DIY Budget Intercooler Fitment
and Relocating the Battery.)
The airbox design that resulted from these constraints (along with others
like cost and effort) is probably a good compromise of ‘do-ability’ and
Looking at the available space, it appeared that sheet and angle aluminium
could be used for half of the box, with the rest of the box formed by the inner
guard and strut tower. This approach would position the sheet alloy facing the
engine and turbo and would – with a few extra extensions – also provide a good
mating surface for the cold air feed, which was to come from a bonnet scoop.
In other words, what was to be formed was an open-topped box with two of the
walls made of sheet aluminium and the other walls by the existing metalwork of
the car. Sealing of these parts together would be by foam rubber strips.
However, when it’s said quickly like this, even this sounds easier to build
than it really was. There’s nothing in the construction that’s really hard – just time-consuming and a little
fiddly. It’s the sort of job to do when you have a full free day and want to
really get into the nitty gritty of working with your hands. It’s definitely NOT
the sort of thing to do when you’re in a tearing hurry and really want to be
The sheet aluminium was cut with a jigsaw (lubricate the blade with auto
trans fluid – it works brilliantly!) and pop rivets were used to hold the pieces
together. The edges were joined with alloy angle and a lip around the top was
also formed from angle. (Incidentally, all this material was bought from a scrap
metal dealer – easily the cheapest way of getting aluminium in a variety of
At this stage the half-box looked like this. Note that the seams can be
sealed with silicone.
With the half-box in the car I placed foam rubber around the right-hand side
and bottom-right (arrowed). (The left, bottom of the box is alloy sheet.) Note
the lip around the top of the box that will seal against the foam rubber of the
bonnet scoop when the bonnet is shut.
If the airbox is to be fed by ducts, the construction can be much the same
but a top panel can be attached by self-tapping screws and appropriate holes for
the fresh-air ducts can be made in the walls of the box.
With the foam painted black with a spraycan, the end result immediately looks
more professional. Note that the foam rubber can be held in place with contact
adhesive or (as here) simply pushed into place and held in position through its
With the cone filters inserted into the intake pipes, the view looks like
The Air Feed
The air feed was to be by a bonnet-mounted scoop. However, making things a
bit more complex was the fact that the scoop additionally had to feed an
underbonnet intercooler, mounted ahead of the airbox. To accomplish both tasks,
the inside of the scoop was organised as is shown in this diagram. The view is from above with the scoop removed.
Air entering the scoop can go in two main directions. Some enters two feed
tubes that connect the front of the scoop to the rear, supplying air to our
airbox. The air that enters in between the two feed tubes is directed down
through the intercooler. Air is prevented from flowing across the intercooler
and into the airbox opening by a foam rubber blocking pad positioned laterally
across the inside of the scoop.
So all air flowing to the airbox occurs through the two tubes that run from
near to the front of the scoop – so what’s the detail on these?
The flow of air into these tubes is critical in determining how much air will
get to the airbox. And since this is air (when mixed with fuel!) makes all the
power, you want as much to get in as possible... In this case, the vertical
dimensions of the scoop meant that 50mm was the maximum diameter of plastic pipe
that could be used – and considering the power of the engine (about 140kW), 2 x
50mm pipe makes for a little smaller total cross-sectional area than I’d like.
So to improve the flow of air into these tubes as much as possible, the
fronts of the tubes were (1) mounted back from the mouth of the scoop, so
positioning them in the high pressure area caused by the foam rubber blocker and
the limited outflow going to the intercooler, and (2) equipped with
Bellmouths – a curved edge all round the entrance to a pipe – can be formed
in a number of expensive ways but the approach used here is cheap and easy. By
cutting up a plastic cake dish normally used to cook a microwave cake (one that
has a large hole in the middle), an excellently radius’d bellmouth can be
obtained. (For more on bellmouth shapes, see the ‘Bellmouths’ breakout box
below. For more on making bellmouths from cake dishes, see the end of Driving Emotion.)
This photo shows the bellmouth that results from chopping away all the excess
plastic. With a little heating of the 50mm plastic pipe to soften it, it was a
push fit in one end, then being securely retained in place with glue.
The top and bottom of the bellmouth was sanded away a little to get adequate
clearance inside the scoop, and the other end of the pipe was heated slightly to
allow it be to be flattened a little on the top and bottom. After these
processes, here’s what one of the airbox feed tubes looked like. Remember, two
of these sit inside the bonnet scoop.
As usual, the application of black paint from a spraycan improves the
Here’s the external view. Looking into the front of scoop you can just see
the two airbox feed tubes.
It’s easy to get hung up on intake pipe sizes to airboxes – and with good
reason. The flow of a pipe is roughly proportional to its intake area, so
working this out for various pipe diameters gives you a quick feel for flow
area. If the pipe is circular in cross-section, its area is found by: diameter x
diameter x 0.785.
So a 7.5cm diameter pipe has an area of 44.1 square cm. Two 5cm diameter
tubes have an approx combined area of 5 x 5 x .785 (x 2 tubes) = 39.3 square cm.
From this you can see that two 5cm tubes provide less cross-sectional area than
a single 7.5cm tube (39.3 vs 44.1 square cm). For comparison’s sake, a single
10cm tube has a cross-sectional area of 78.5 square cm!
Yep, it pays to go big...
Sealing the Scoop
The air feed tubes direct air into a bonnet opening cut above the airbox.
However, for all of this air to reach the airbox, the scoop must be sealed to
the top of the box. This is a critical part of making the design work – some testing was carried out without this
seal in place (even with the air feed tubes working) and the intake air temps
were still way high.
Sealing a bonnet opening to an airbox (or intercooler for that matter) is
normally very difficult because of all the variations in clearances caused by
curved bonnets, bonnet ribs etc. A simple but very effective approach was
After the opening had been cut in the bonnet, a flat piece of sheet aluminium
was cut out, with a centre opening the same shape as the airbox opening. The
plate was then pop-riveted to the bonnet ribs, so that when the bonnet was
closed, the plate came down directly over the airbox. (In this case I didn’t
have quite enough material available so the plate was made in two pieces.)
The gaps in between the plate and the skin of the bonnet were then filled
with foam rubber strips which were later spraypainted black. This gives a flat
surface, airtight sealed to the bonnet opening.
The airbox had already been constructed with a flange formed from aluminium
angle running around its top edges and so when a foam rubber strip was glued to
the aluminium bonnet plate, a positive seal could be gained with the airbox when
the bonnet was closed.
This photo clearly shows the system – the bonnet scoop opening that seals
down onto the top of the airbox. It’s important to stress again that this
sealing process is one of the most important aspects to get right.
For those people who want to tie both the intercooling and airbox ideas in
with the diagram above, here is the wider view.
Initial testing concentrating on measuring intake air temp (with the probe
placed just before the throttle body) and the pressure within the airbox itself.
In addition, the temp of the airfilters was also measured by hand.
A 0-1 inch of water Magnehelic pressure gauge was used to measure the
pressure within the airbox. This is an extremely sensitive gauge (we've
covered them before so to find out more, do a site search under
'Magnehelic') which is easily able to monitor pressure variations of this
type. Testing showed that at 100 km/h, a pressure of 0.6 inches of water
was present within the airbox - that is, the airbox was under positive
pressure. Even at full throttle at this speed, positive pressure was
maintained within the airbox - an excellent result. Positive pressure could
also be measured at lower speeds - eg 0.2 - 0.4 inches of water in cruise at
These measurements showed that the two air feed supply pipes were not only
providing enough air for the engine’s needs, but they were providing more than enough!
The temp of the filters was also near to ambient. Feeling by hand indicated
that they were a bit warmer than outside air, but only by 5-10 degrees C or so.
IOTW they were just warm, whereas before the airbox was built, they were always
stinkin’ hot! The measured intake air temp near the throttle butterfly was about
15 degrees C above ambient – still warmer than with the original single filter
layout, but an acceptable trade-off with the much greater filter area mounted
inside the engine bay.
With careful design it’s possible with only hand tools and cheap materials to
make an air intake system that is highly effective at providing a heap of cool,
force-fed air to the engine. Even including the price of the filters and bonnet
scoop, this entire set-up cost well under AUD$200 and gives results better
than nearly all aftermarket filter installations.
The sharpness of the radius of the lip that surrounds a bellmouth is
important. If it is too tight, the airflow ‘unsticks’ on the transition around
the corner and the intake flow is reduced.
Flow Fans and Ducts (R. Alan Wallis, 1983,
John Wiley and Sons, ISBN 0-471-87086-2) the minimum radius of the bellmouth lip
is specified as being best between 0.25 and 0.3 times the diameter of the tube.
So for example, a 7.5cm tube should have a bellmouth that has a radius of
curvature that is 1.9 – 2.3cm. (Where did these figures come from? 7.5 x 0.25 = 1.9, and 7.5 x 0.3 = 2.3).
To turn it into data that can be more readily used, the diameter of a disc that can be nestled
inside the outer face of the bellmouth lip should be 0.5 – 0.6 times the
diameter of the tube. Most bellmouths are in fact smaller than this – these
represent figures to be strived for.
In this pic you can see that the Australian currency 20c coin just fits
inside the bellmouth of this 50mm tube. The coin is about 19mm in diameter so
this bellmouth has a lip that is 0.38 of the diameter of the tube – it’s a good
However, these stainless steel bellmouths (made from eggcups – see Driving Emotion),
are much tighter – the washer is about 11mm in diameter and the tube opening is
also 50mm, giving a relationship of 0.22 Not so good – although still far better
than just a bare cut-off tube.