Single turbos, twin turbos that are simultaneous, twin turbos that are sequential, turbos on in-line engines and turbos on V-engines, turbos that are sized small, turbos that are sized big. If you’re thinking of a custom turbo installation, they’re all ideas that you need to know at least a little about. So what’s the story on the road? We look at how a whole bunch of turbo cars really perform.
A turbo uses the energy left in the exhaust gases after they've come out of the engine. In a non-turbo car, the heat and pressure in the exhaust is usually wasted - sent out of the back of the car with nothing much done with it. On a turbo car, those exhaust gases are channelled by means of a turbo-specific exhaust manifold to the turbine side of the turbo. Here there's a wheel with specially-shaped blades attached to it. The exhaust gases are nozzled down so that they increase in speed, and are aimed at the wheel. This causes the turbine to spin very quickly - at up to 100,000+ rpm.
The turbine is mounted on a shaft, and fixed at the other end of the same shaft is a compressor. This is another wheel with special blades, but the job of this wheel is to blow air into the engine - to develop boost.
At idle, there isn't much exhaust gas coming from the engine. The gases are still all directed at the turbine side of the turbo, but often they simply flow through the wheel, sometimes not even causing it to turn. Likewise, the compressor end of the turbo isn't turning yet - if the turbine is stationery, the compressor must be too, because they're both solidly attached to the one shaft! Even in cars where the turbo is spinning at idle, there's no boost yet being developed.
So what happens when you put your boot into it? The engine revs and load rise, and so the amount of exhaust gases coming out of the engine goes up quickly. This starts the turbo spinning fast. Once the turbo is spinning fast enough, it will develop boost - blowing more air into the engine than the engine can actually breathe. That's why the pressure in the intake manifold goes up. Then the engine will inhale more than usual, and so the exhaust flow will also rise. This spins the turbo even faster - so pushing more and more air into the engine, which in turn produces more exhaust flow....
The biggest problem with matching a turbo to an engine is in coping with acceleration transients – that’s when the power demand is rapidly increasing. A turbocharged piston aircraft engine or big stationary diesel engine changes in power output quite slowly, and so the turbo can be pretty big to provide maximum efficiency. But a car engine needs to have boost over as wide a range as possible, and furthermore, the turbo needs to ‘keep up’ with changes in engine load. This is best achieved by using a small turbo, or when a small turbo can’t provide enough power, two smaller turbos. The lower rotational inertia of the small turbos allows them to accelerate more quickly, reducing turbo lag.
So what’s turbo lag, then? That’s when you nail the throttle, especially at low rpm, and wait-wait-wait for the power rush to arrive. The longer the wait, the poorer the turbo is at keeping up with the changing power demands. A turbo’d car will always have softer throttle response than a naturally aspirated car (well, when both have well-mapped management systems, anyway) but a soft response shouldn’t translate to a power delay.
High power engines with small capacities have less exhaust gas available to initially spin the turbo up (or more accurately, proportionally much less than they have at full power), so these engines are even harder to provide with a responsive turbo. In order that a turbo can accelerate rapidly in speed, turbo manufacturers have used ceramic turbines (less mass so fast acceleration) and ball-bearing turbos (less frictional drag from the bearing). Electric-assist turbos are currently being developed: except to see them on new cars pretty soon.
More than any other factor (maybe packaging constraints aside), OE car manufacturers engineer their turbo systems to reduce turbo lag and improve bottom-end responsiveness. To achieve this, they place the turbo(s) very close to the exhaust port, use a turbo that most enthusiasts would consider undersized, and design the rest of the engine to provide good low-rpm exhaust gas-flow.
European manufacturers – like Audi, Saab and Opel – produce four cylinder turbo cars that have peak torque at incredibly low revs, eg 2000 rpm. In other words, the turbos in these cars are producing plenty of boost (and maybe have even reached their wastegated limit) by these low engine speeds. Lots of macho enthusiasts like to pour shit on this idea but in the real world of driving, this approach is extremely effective.
A torquey, responsive car is easy to drive fast, often provides very good fuel consumption (because revs are so low much of the time) and is very user friendly. As an example of the latter, boost in these cars very seldom arrives with a rush (because the turbo is spinning up as the engine is), meaning that there’s less chance of wheelspin. Simply, that means faster times point-to-point. The Astra, Saab 9000 Aero and Audi A3 turbos are examples of this approach.
In our 2003 test of the Astra Turbo, we said of its engine:
With a dead-smooth idle and hugely driveable bottom-end torque, you could be forgiven for thinking that the top-end wouldn't be that strong. And in a way you're right - with peak power at 5600 rpm, there's not a lot of point in taking the engine to 6800 rpm... which it will still do smoothly and happily.
On the open road the torqueiness of the engine is quite incredible. Helped no doubt by the cold ambient temps experienced during our trip, there was literally no open-road hill that required a down-change if the speed was above 80 km/h. But the bare performance times simply don't tell the story - even though a 0-100 time of 7.3 seconds is quick.
It's in the cut and thrust of urban traffic - where the instantly available torque is just so unusual for a turbo car and where it's simply impossible to be caught off-boost - that the performance starts to assume greater proportions. You can short-change at 3000 rpm and still be quick; take the engine to six grand and you're seamlessly very quick. The Astra turbo specs show it to be as fast as a current Impreza WRX, but in the actual driving, most times it's quicker.
Of the Saab 9-3 Aero, in 2003 we wrote:
"Used to have a bit of lag didn't they, those turbo Saabs?" queried the bloke having a good perve at the new 9-3 Aero as we stood filling it with fuel.
Boy, if only he'd been in the car just a few minutes earlier.
"Nah, mate, not this one," we fire back. "This one drives like a big six-cylinder; it's got instant get up and go." The fuel filler clicks. "The only difference," we add, "is it's a lot more economical..."
The new 9-3 Aero has a fabulously flexible and responsive engine. Throttle sharpness is nothing out of the ordinary, but the low and mid-range rpm punch is what you'd expect if Saab had stuffed a dirty big six under the bonnet.
If you don't reckon a 2.0-litre turbo can give big-cube punch at low revs, you're in for a surprise! From seat of the pants, full boost arrives at very low revs - and that makes sense when you consider the engine's torque curve. Torque doubles between 1000 and 2500 rpm, with a strong 300Nm plateau held from 2500 to 4000 rpm. Its no wonder open-road overtaking is so effortless - just squeeze the throttle and you're steaming ahead.
Of the A3 Audi, we said:
Using a sweet and hi-tech five-valves-per-cylinder 1.8-litre transverse four, the Audi A3 turbo develops 110kW at a relatively low 5700 rpm and peak torque at 4600 rpm - but the latter feels much lower. With variable cam timing, electronic throttle control, a well-matched turbo and 9.5:1 compression, this is one turbo car that never feels like it has a puffer under the bonnet. Instead, the torque delivery is linear, with good power available from the tractable and refined mill, even at low revs.
It's quite easy to skip the in-between gears on the way up through the 'box - the car is happy with a first-third-fifth gear change sequence. The driveability provided by such a good torque delivery is not to be dismissed; this is one car that's very easy to hustle quickly because there're always two gears just right for the occasion!
On the other hand, some manufacturers - like Subaru with some of their earlier WRX STi models - trade off hugely in bottom-end turbo performance for higher peak power figures. In other words, they go for a bigger turbo which – in the case of the Subaru – is mounted a long way from the exhaust ports. The result is a more exhilarating ‘coming on boost’ but – and it’s a bloody big ‘but’ - the off-boost behaviour of the car is terrible. These are the turbo cars that unless launched at high revs, are often 2 or even more seconds slower in their 0-100 km/h times than their published specs. The R32 Skyline GT-R – yes, despite its twin turbos – is another of this school, using very low gearing to give response but still being very slow when caught off-boost, especially getting away from a standstill without a huge launch.
In our 2002 test of the WRX STi we said:
It's pretty sad that the STi gets gobbled up by ninety percent of traffic in normal day-to-day driving. No joke - caught out at anything less the mid-range rpm, the Super Rex is an absolute s-l-u-g.
To give you an example, not long after picking up our test car, we found ourselves stopped at a set of traffic lights with our lane ending about forty metres ahead. The lights change green, we engaged first gear and stomped the loud pedal all the way to the floor - just to make sure we could merge into the next lane with plenty of room to spare. What followed was tragic. Despite its wide-open throttle, the STi proved unable to muscle its way past a humble Mitsubishi Colt in the adjoining lane....
We couldn't even pull out a nose length before we had to abort and hit the brakes...
Embarrassing - not to mention dangerous - situations like this quickly teach you the MY02 STi is a no-goer at anything below mid rpm. It's tractable - yes - but don't expect any useable acceleration
Keep the throttle floored past 4000 rpm, however, and - bang! - the STi transforms into a head-kicker. It slams your head back with the subtlety of an uppercut to the chin.
It’s a different story when you’re starting with an engine twice as large. The Falcon XR6 Turbo, for example, has quite a large turbo that can keep up with modified power outputs of twice the factory engine. Which, since the standard 4-litre engine develops 240kW, is really something. And yet the car in standard form doesn’t drive in a laggy way. Partly helped by the prodigious bottom-end torque that this engine design develops, the turbo comes up on boost smoothly and quickly.
On our 2002 test of the XR6 Turbo, we said:
Awesome, simply awesome. Ford Australia has held nothing back with the release of its new Ford Falcon XR6 Turbo and the result is good enough for it to become the next cult car. Consider the XR6T 'big picture' for a moment; here's a freshly styled full size sedan with a modern tech turbocharged engine capable of 240kW and enough low-to-mid range torque to make a big 5.7-litre V8 cry.
The new turbo engine is fantastic; it offers good throttle response (at least as good as any other turbo car on the market) and it's supremely flexible at all revs. No need to row gears in the XR turbo because with a 450Nm torque plateau available between 2000 - 4500 rpm she'll pull away instantly. With the turbo set to deliver just 6 psi, the new engine can stomp out 240kW at 5250 rpm and a very handy 450Nm from 2000 through to 4500 rpm. Who said all turbo engines are peaky?!
You can't help admire how effortlessly the XR6 Turbo 5-speed can outrun the fabled Impreza WRX and run alongside a 5.7-litre Commodore. Giving it just a gentle launch with two people onboard we hand-timed a 0 - 100 km/h sprint in 6.6-seconds. With a bit of practice, though, we reckon the XR could crack 6-seconds flat - seriously cookin'.
And all these turbo cars that drive so well achieve this performance without equal-length runner exhaust manifolds that fill half the engine bay, huge blow-off valves or any of that crap...
When manufacturers do decide to go to a more sophisticated turbo system, they fit twin turbos. Twin turbos are not particularly suited to four cylinder cars (although some people have fitted them), but instead lend themselves ideally to in-line sixes and V-engines. Because of the way the pulses of exhaust energy can be entrained, the sixes work very well with two turbos, while the V8s suit twin turbos because of the proximity to the exhaust ports with which the turbos can be mounted. However, in the case of a six cylinder engine of say 2.5 or 2.6 litres, you’re really looking at the exhaust flow of only two 1.25 or 1.3 litre engines and so if you fit turbos that are too large for these gas-flows, the result will still be doughy. Especially if the engine is set-up for good top-end power.
The grey import 2.5 litre twin turbo Toyota Soarer (tested in 2001) is a case in point. We drove it after experiencing the sequential twin turbo Supra (we’ll cover sequential turbo systems in a moment) and had expected it to be similarly good:
The gasses exiting its 1JZ-GTE head are constantly divided into two simultaneously operating turbos; this means boost response and low-rpm torque are both a fair way behind the sequentially-turbo'd RZ Supra. You really notice this bottom-end 'hole' when punching the 1600kg Soarer off the line - it feels quite doughy until just before 3000 rpm.
[Compared to the manual transmission car] we found the Soarer more enjoyable to drive when fitted with the 4-speed automatic transmission. That's because the converter multiplies low-rpm torque going to the rear wheels as well as allowing engine revs flare (often by up to 800 rpm). Under increasing engine loads (such as when climbing a gradual incline) the transmission is also very willing to slide back a gear or two, helping to keep the engine on song. The engine and trans are obviously a very well calibrated pair.
The rare 5-speed manual version (which some have suggested is good for 0-100 in the 5s) is not much quicker than the auto without a big clutch dump and brutal shifting. Only under these circumstances will you extract any worthwhile acceleration advantage over the auto. In gears, we also noted that there's bugger-all separating the performance of the two cars fitted with the different transmissions.
What we most dislike about the manual Soarer is that it emphasises the engine's bottom-end torque shortcoming; stomp your foot on the accelerator at low rpm and you have to wait until road speed equates to about 3000 rpm before any real urge become apparent.
Running the turbos in sequence, where initially one turbo gets all the exhaust gases before the turbos are fed in parallel, potentially provides the best of both worlds. However, engineering the changeover is very complex: how do you stop the first turbo dying away when the exhaust valve opens to allow gases to flow to the second turbo? But when it’s done right, this configuration is simply head and shoulders above any other turbo philosophy based around a smallish engine.
Our 2002 drive story of the Supra VVT-i RZ said it all:
The post-1998 VVT-i version of the Supra 2JZ twin-turbo 3.0-litre six is - without doubt - the best production turbo engine we've ever driven.
Stroke it along gently and the VVT-i 2JZ-GTE six behaves as 'proper' as a top-line Mercedes. Despite having 209+ kilowatts on tap from only 3-litres, there's absolutely no hint of lumpy camshafts or an all-or-nothing turbo system. The only on-going reminder of the engine's potency is its throaty exhaust burble from out back. It sounds sensational.
Throttle response (via the ETCS-i electronic throttle) is immediate and comes backed with a progressive snowball of torque. And, no matter what revs you've got the thing lumbering at, there's never a time when you have to row down through the ratios to find accelerative urgency - certainly a rare enjoyment in a turbo car.
On paper, the sequentially-turbocharged 3.0-litre VVT-i 2JZ-GTE is credited with a long-trunk'd 451Nm of torque at 3600 rpm; to give you a comparo, an R32 GT-R makes 355Nm, an Audi S4 twin-turbo 2.7 boosts its way to 400Nm, a Liberty B4 generates 320Nm and an Evo 6.5 kicks out 373Nm.
N-o-w do you get the picture how much grunt the mega-Supra has on-call!?
However, it’s easy to get a sequential turbo engine completely wrong, as our 2001 Subaru Liberty B4 story said:
The newly-released Liberty B4 is powered by an EJ20B 69H - a turbocharged, intercooled, DOHC, 16-valve, 2-litre boxer four, which is of the same basic design as the engine found in WRXs and the Liberty RS of ten years ago. Internal developments have resulted in reduced valvetrain mass and an increased compression ratio (up from 8.0 to 9.0:1), and Subaru has endowed the B4 with two sequentially-staged turbos in order to maintain a good spread of torque.
The B4's focus is on response and flexibility in everyday conditions.
The hottie Liberty is set up to offer good throttle response and backs it up with a strong surge of torque anywhere in the rev range. The primary turbocharger is arranged to quickly deliver boost up to 4000-4500 rpm, with the secondary turbocharger then kicking in to add flow in the higher ranges. It's a cunning system - but is it perfect?
During the transitional phase - where the secondary turbocharger is starting to pump in addition to the primary unit - there's an ugly 'hole' in the torque delivery. Under full throttle, a significant dip in manifold pressure identifies the 4000-4500 rpm transition; we measured a full 0.25-0.3 Bar boost pressure dip.
It's enough for first-time passengers to ask if there's an engine problem....
So where does all that leave us? For driving – as opposed to dyno comps for peak power figures – a turbo that leans towards a smaller size will give far better on-road performance. That’s pretty well the opposite of what you’ll read anywhere else.
So if you’re running a car that is already factory turbo’d, leave a turbo upsize to the last step in the modification process... don’t make it the first. If you’re starting off with a naturally aspirated engine that you’re turbo’ing, seek good advice as to the turbos that would suit the engine - if you want to make peak torque at (say) one-third of the redline revs. Usually, that will mean reaching peak boost by that rpm – and then of course holding it at that value. This approach will limit the top-end power, but the higher average torque through the rev range will make a powerful (sorry!) difference to the real world performance of the car. (Or, if you want to trade-off some more bottom-end response, move the peak torque request up to half of the redline rpm.)
If you need more power than can be provided by a single, reasonably sized turbo, and you have a V or six cylinder engine, use twin parallel turbos. Of course, sequential would be better but we’ve never seen a twin sequential turbo install done in the aftermarket (most people are too busy talking off twins to fit a single huge, laggy turbo!).
And finally, be a little wary of dyno curves that show a fantastic boost curve from a big turbo on a little engine. On the dyno it’s often easier to load-up the engine by the use of a high gear or slow ramp speed in a way that results in a completely different boost curve than you’ll find on the road in first and second gears.