This article was first published in 2006.
Arguably, the 1983 Audi 100 started it all – at least in post World War II
times. The Audi was the first car for many decades that combined a relatively
unremarkable appearance with a very low drag coefficient (Cd) of 0.30. Sure
there had been Citroens and Porsches and some others, but the Audi was neither
weird looking nor a svelte two-door sports car. (The Ford Sierra was another
groundbreaker but at the time, it was considered a very strange looking car.
Consider: it replaced the Cortina!)
After the Audi 100, manufacturers around the world turned to new body shapes
giving lower drag coefficients, mostly in search of better open road fuel
economy. In some markets, the change was startling: Australia went from the
VB-VL shape Commodore (a series whose incredibly bad aerodynamics was best
highlighted by what needed to be done on the pictured ‘Walkinshaw’ VL to make
real improvements) to the slippery VN Commodore. Ford released the EA and – much
later – Mitsubishi the TE Magna. Especially in the rear treatment (the back of a
car is more important than the front in gaining low drag), these cars
represented a radical change with their shallow-angled rear glass and high boot
Of the Europeans, especially Audi and Mercedes released long lines of stylish
and slippery sedans. In Japan, the first Lexus LS400 combined underfloor air
guidance panels and very clever front-end sealing with superb rear aero to give
a claimed drag coefficient of 0.29. The change had been dramatic – from cars
typically with drag coefficients of 0.38 – 0.4 dropping in just a few model
cycles to around 0.30. That’s a 25 per cent decrease in drag coefficient,
resulting (in cars retaining the same size) in a 25 per cent reduction in total
drag. (Total drag is worked out by multiplying the drag coefficient by
effectively the height and width of the car.)
Pundits were suggesting that in the near future there would be plenty of cars with
0.25 and lower drag numbers. It seemed like a win/win change: since at speed by
far the majority of power is used to push air out of the way, slipping more
easily through that air results in more power available for acceleration and
less overall fuel usage. And what with the huge investment in wind tunnels, the
experience that aeronautical engineering could bring, and a new crop of stylists
schooled in aerodynamics, well, it was all going to happen.
Except, it hasn’t.
Aero Progress Falls Apart
Perhaps it was the Audi TT that really highlighted the decline. That’s
especially sad as the TT was the product of a company whose aerodynamic
development didn’t just include the 100 but also - through the acquisition of
NSU - the Ro80 and stretched back more than 50 years to the storming Auto Unions
of the ‘30s. If you’ve forgotten, the Audi TT was first released without a rear
spoiler. Just eyeballing the rear of the TT shows some high lift surfaces and
yes, that’s exactly what happened. Initially Audi denied any problems but when a
former champion rally driver was killed when his TT spun at high speed,
suspension modifications and a rear spoiler were retrofitted fitted to all cars.
The rounding of subsequent Audi tails will not have been good for drag,
either – a sharp cut-off provides for clean separation and so makes for a
In fact, you can look across all makes and see that aerodynamic progress has
effectively stopped. Actually, it’s worse than that – with the increase in the
size of cars, retaining similar drag coefficients means the total drag is really
going up, not down. Car manufacturers now don’t even bother stating Cd values,
confident in the knowledge that the general public doesn’t understand the
significance (but there are lots of vehicle specs that the public doesn’t
understand the significance of, yet the car manufacturers still provide them)
and that most people can’t simply look at a car and get a feel for how slippery
To see this in crystal clear detail, start looking under a few current
releases. Almost without exception, no attention at all is being given to
smoothing undercar airflow – despite the fact that over one-third of total drag
is caused by the underside of the car! Oh sure, a few little things are being
done like shields ahead of the front wheels and vestigial deflectors ahead of
the rear wheels. But if you want a real comparison, peer underneath a 1960s
Citroen or Porsche, or under a 1930s Tatra (pictured). These cars have almost
completely flat and sealed floors – not just a few in-fill panels.
It must be said that DaimlerChrysler – especially through its Mercedes models
– has at least kept styling and aerodynamic engineering working hand in hand,
but for most companies, chasing low drag is simply not a big deal. Instead, all
the CFD and wind tunnel work is being devoted to the humdrum – making sure there
are enough cooling airflows, keeping aero noise down and basic stuff like
But What Could We Have?
For starters, it’s bloody stupid how moveable aerodynamic surfaces are being
used in only a few sports cars, where typically spoilers pop-up at speed. All
cars – not just sports cars – could benefit from aerodynamics that alter with
the use of the car. Sound a bit far-fetched? Not a bit of it – and here’s a
It was mentioned above that underfloor drag is responsible for about
one-third of the total drag of a current car. The other proportions are also
easy to remember – about one-third for all the exterior surfaces and the last
third caused by airflow required for engine cooling. (See
Modifying Under-Car Airflow, Part 1.)
Now, answer this question. What proportion of the time that you’re driving
your car do you think the cooling system is working at full capacity – that is,
the thermostat is fully open and the radiator is being completely utilised? One
per cent, two per cent, never? Well, OK then – so how come as much air as
possible is always being shoved through the radiator? Why don’t cars use
moveable aero surfaces that channel only as much air through the radiators as is
actually required? If they did that, the aero drag caused by the cooling system
could be massively cut for a high proportion of the time.
And this thought is hardly new. More than 60 years ago, piston engine
aircraft (they were all then powered by piston engines, of course) used variable area outlet ducts to control the flow of cooling air through the radiator (water
cooling) or the engine cooling fins (air cooling). The Elementary Handbook of
Aircraft Engines (published in 1945) describes the system used with the
Gipsy V12 cylinder air-cooled engine in the de Havilland DH-91:
Control of the airflow is effected in the case of the Gipsy engine by an
air-exit gill located in the most efficient position for minimum interference
and drag, namely, on the underside of the engine nacelle. In the fully-open
position for take-off this gill induces a suction which assists the slipstream
pressure in forcing air over the engine. In cruising conditions the gill is
nearly closed to restrict the airflow to the desired rate, at the same time
greatly reducing drag.
(Note that some early cars used Venetian-style radiator blinds, but these
were for controlling coolant temp rather than for aero advantage.)
So how do current cars control radiator airflow? The short answer is that
they don’t. Air is picked-up from the high pressure zone across the front of the
car (and some cars used purpose-shaped ducts starting at the stagnation point
to do this), is pushed through the radiators and then spills untidily out past
the engine and into the turbulent area between the rough underside and the
ground. Hell, a 60 year old aeroplane was far superior to that!
Or what about using expandable membranes to change the shape of the body and
so reduce drag? If you’re doing 100 km/h, you aren’t going to need to open the
doors, lift the bonnet or get something out of the boot, are you? So why keep
the same degree of access and wear the subsequent aero loss outcome?
Again, over 60 years ago, a long distance bus was trialled that used an
inflatable extending tail. At speed, it extended. When the bus came to a halt,
it retracted. It sounds odd but only because we’ve never considered it. After
all, modern jet aircraft change the shape of their wings enormously depending on
whether they’re taking off, landing or slowing on the runway after landing. Why
have a fixed shape car when straightforward changes could reduce your high sped
travelling bill by 20 percent?
And the change in shape doesn’t even have to be visual. DaimlerChrysler has
patented a system where air jets are pumped through surface holes, effectively
changing the aerodynamic body shape and also giving the ability to control how
the air departs (see
Cutting Edge Aerodynamics).
Smoothing the underfloor, providing variable cooling airflow capability (and
better channelling the outlet air) and potentially more radical technologies
that change the shape of the car could certainly result in a further drag
reduction of 25 per cent, or Cd numbers in the low twos.
The most aerodynamic car ever sold in Australia – the Honda Insight – has a
Cd of 0.25...
...and the current Toyota Prius has a factory number of 0.26.
The most aerodynamic car ever sold (well, leased, anyway) in the world was
the General Motors EV1, with a Cd of 0.195. These cars show what can be done –
but other role models are few and far between.
Right now, car aerodynamic design is going nowhere - and we’re all paying for
that inaction every time we fill up at the pump.