Last week in Part 3 of this series we showed how the aerodynamic pressures acting on cars can be directly measured – perfect for finding the best location for spoilers, wings, bonnet vents, scoops and engine intakes. This week, we’re going to look at testing performance.
Why Do It?
There are two good reasons to test performance.
The first is so that you can make comparison of your car’s performance with other cars.
I remember the third car I ever owned – a ’77 BMW 3.0si. A fuel injected, 3 litre straight six with a manual 4-speed trans, the road test time was a 0-100 km/h in (I think) 7.7 seconds. One night, out the back of the country town in which I lived, I ran some 0-100 km/h times – and couldn’t believe how slow they were! It was then I realised that magazine performance times are normally run with complete disregard for any mechanical sympathy. High rpm clutch dumps? Sure....
Unless you’re in some type of competition (real or imagined), it’s the second reason for testing performance that is much more important. And that’s to see the direction that modifications are taking you.
The car I owned after the BMW was a near-new Holden Commodore VL Turbo. Back then, the worth of modifications to the car was largely unknown – people said what results they got from modifications, but figures weren’t available.
My auto trans VL Turbo ran a standard, measured 0-100 km/h in 8.4 seconds (magazine time was 8.0 seconds). I had a new performance exhaust fitted, then ran another test. The 0-100 km/h time remained exactly the same – 8.4 seconds...
That was a long time ago, but I can very clearly remember the shock with which I discovered that some vehicle modifications do nothing for real-world performance. In short, the money I’d just spent on the new exhaust was completely wasted. (What was wrong with the exhaust? Stuff like the wrong cat, wrong mufflers, wrong bends – you live and learn... especially in not trusting all workshops!)
Careful testing of on-road performance is vital if you’re to make cost-effective, worthwhile modifications. It can also be done at zero cost, and in many situations is much more valid than chassis dyno testing.
Huh? Better than using a dyno
Let’s take a look.
Chassis dynos measure the amount of power being developed at the wheels over a range of engine speeds, recorded at steady-state full throttle. That’s great, but it often tells you little about on-road performance – and invariably, little about the subtleties of good driveability.
For example, what does a dyno chart tell you about the action of a turbo boost control? In short, almost nothing at all.
1) The dyno acceleration rate in each gear is not representative of what is achieved on the road. This one’s a killer because boost overshoot on transients is hugely affected by the rate of engine rpm increase.
2) Very few people do full throttle gearshifts on the dyno. You know, race up through the gears – to the redline, change gears, to the redline, change gears. Again, it’s in just these conditions that you look for boost overshoots and/or slow increases back to peak boost after each gear-change.
3) I have never seen anyone do a full-bore launch from a standstill on a dyno. And how quickly boost can be brought up in these conditions – ie controlling wastegate creep – is a major aspect of good boost control.
4) On a dyno the intercooler never works as well as on the road, so the actual peak boost is likely to be different in those systems that don’t use boost pressure feedback.
5) On the dyno people never bother trialling all the different combinations of throttle position, load and engine rpm that you’ll find in half an hour on the road.
And it’s not just boost controls. Most dyno operators don’t start a run at very low rpm, so the engine revs being used most often in daily driving are not tested. The ‘area under the power curve’ (especially at part throttle) is ignored, while aerodynamic pressures (eg on the engine intake and intercooler) are all wrong.
A dyno is a great tool. But unless it’s inside a climate-controlled wind tunnel, can replicate all the different acceleration rates and inertial characteristics of the driveline, and operates on real world drive cycles, it’s only a starting point.
What a dyno is good for is engine management tuning for maximum power – and even then, this must always be followed by extensive road testing.
In short, when you’re doing on-road testing, you do not want an ‘on-road dyno’ – instead you want to test road car performance!
Stop Watch Testing
By far the best tool for assessing vehicle performance is a stopwatch and the car’s standard speedo.
With care, stopping and starting a hand-held stopwatch can be carried out with a consistency of less than 1/10th of a second. And it should be noted that any performance mod that results in less than a tenth of a second gain isn't very successful!
Practise starting and stopping a stopwatch while you view the second hand on an analog clock or the digital seconds display on another watch. Time a number of 10 second increments on your stopwatch and see how close you actually get to 10.0 seconds.
I have just stepped away from the keyboard to do this and have got the following results: 9.94, 10.10, 9.98, 9.92, 9.94 and 9.96 seconds. You can see that when the single figure furthest from the median is excluded, the timing is only 0.6 per cent inconsistent!
Obviously, with the g-forces pushing you back in your seat and the speedo needle whipping around while the road races towards you, the in-car timing won't be this accurate, but using a digital stopwatch is far better than most people think.
Using a stopwatch to measure acceleration performance in this way does not require that the speedo be accurate. If you wish to compare your results with figures gained in other cars, then of course the speedo must be correct, but in the vast majority of cases you will be simply comparing the performance before and after making a modification. Unless you change the tyre diameter, final drive ratio or some other gearing aspect, the speedo accuracy won't change during this period.
Using a stopwatch requires the use of an in-car assistant.
So what can you measure?
The most important real world performance times are those measured ‘in gear’ or ‘rolling’. These can be done in two different ways.
A pure rolling time is one where, for example, you drive along in second gear at 50 km/h. You put your foot flat to the floor and then click the stopwatch start as the speedo needle moves past 60 km/h. You then again click the stopwatch at 90 km/h - and then back-off after that. So you’re measuring how quickly the car will accelerate from 60 to 90 km/h when it has already started accelerating.
The other way of doing this is to drive along at a constant 60 km/h and at the same time as you boot it, press the stopwatch. Again stop the watch at 90 km/h (or whatever speed you nominate). This measurement not only takes into account acceleration, but also response – the time it takes for the car to start to accelerate.
Note that the speed increments described here (60 – 90 km/h) can be used almost anywhere on suitably speed-rated roads – from a freeway ‘on’ ramp to a 100 km/h limited country road. The actual speed range you choose to use, and the gears you do it in, are up to you.
These two ‘rolling’ measurement techniques are exceptionally useful – far more so than the standing start times I’ll cover in a moment. In the real world of road driving, I’d trade off a drop in power for better rolling throttle response. Why? Because a faster-responding car will always get the jump on one with (say) 10 per cent more power but a laggier response. (This is yet another thing a dyno does not show!) A responsive car is also much easier to corner.
(The most responsive car I have ever driven is the 3.5-litre manual transmission Mitsubishi Magna Sports – that car had simply neck-snapping response, and was far more responsive than – say - a V8 Holden Commodore. Don’t underestimate the importance of throttle response.)
Back to rolling acceleration times. I did some modification of a diesel Peugeot 405. The car was a pretty slow (but very economical) machine and so I wasn’t expecting to revolutionise performance. But by making changes to fuelling, the intake and exhaust, I dropped the ‘accelerating’ 80-100 km/h time from 10.7 seconds to 5.5 seconds – so after the modifications, the car took about half the time to accelerate between these speeds in 4th gear.
80-110 km/h in 3rd gear improved even more dramatically - falling from 15.1 seconds to 5.9 seconds.
Importantly, the standing start time 0-100 km/h difference (from 15.3 seconds to 13.4 seconds) was much less impressive – but on the road, the car felt transformed. On the dyno, the measured gains were minor...
Another test was done on a different car, comparing the supposed power gains of an engine oil additive. Testing was carried out using an ‘accelerating’ 40 – 80 km/h in second gear. In standard form the 40-80 km/h times were 3.80, 3.82 and 3.86 seconds. I then added the oil treatment which - amongst other things - was claimed to give up to 15 per cent more power.
The times with the treated oil in the engine were 3.75, 3.76 and 3.79 seconds. This gives a 'before' average of 3.83 seconds and an 'after' average of 3.77 seconds - a statistically insignificant six one-hundredths of a second difference. Graphing the results makes this (lack of) change very clear…
Incidentally, the test results of the oil additive didn’t surprise me. I’ve seen similar (non) results from special spark plugs, special ignition leads, replacement ‘sports’ airfilter elements, etc.
The greatest variation in standing-start acceleration times is that caused by differing launches.
My standard R32 Skyline GT-R used to get to 100 km/h in low sixes if launched and driven as you would if you owned the car; on a cold night clutch-dumped at the redline by a maniac (me in a don’t-give-a-shit mood) it sprinted to 100 km/h in high fours.
A high-powered manual RWD car (and any manual FWD or constant AWD car) is very difficult to launch consistently. In both types of two wheel drive cars, wheelspin will occur if too many revs are used on launch, while a constant four wheel drive car will very easily bog down. For this reason, accurately measuring performance changes with these cars is better done from a rolling start.
On the other hand, all cars with automatic transmissions lend themselves very well to consistent 0-100 km/h timing.
Set the watch to zero (it helps if it beeps when started and stopped) and then rev the engine to the launch rpm. In an automatic car, load the engine with the brakes while applying some accelerator.
Release the clutch (or brakes in an automatic car) as you press the stopwatch and then accelerate hard through the gears. Prepare to press the 'stop' button as the speedo needle sweeps around, hitting the button a fraction before you see the needle actually pass over the mark.
As with all types of acceleration timing, if you are after the most accurate figures possible, make three or four runs in each direction. You will quickly see if the figures form a pattern or are simply all over the place. Don't confuse timing inaccuracies with the performance of the car changing during the testing; as the engine heats up, power (especially in a forced induction car) will often decline.
An example of standing start testing is that carried out on a standard six cylinder EF Falcon.
The automatic car recorded a 0-100 km/h time of 9.2 seconds in standard trim. With the airfilter removed, this improved to 9.1 seconds. The airfilter was then replaced and the bonnet 'popped' to the safety catch. This let more air get to the factory intake snorkel and the time stayed at 9.1 seconds even with the filter back in its box. The bonnet was then returned to its closed position and the intake snorkel to the airbox removed, allowing the engine to breathe hot air. The 0-100 time then lengthened considerably to 9.5 seconds. This Falcon engine is fitted with a dual length intake manifold that changes from long to short runners at a certain engine speed. For the final test it was permanently held in its short runner position, resulting in a slow 9.9 second 0-100.
Many years later, I did some further testing on a modified EF Falcon 5-speed manual, trialling variations in cam timing. In that case, the 0-100 km/h time did not change with the cam timing in either of two positions – it stayed at 7.5 seconds.
Rolling and standing start stopwatch testing will clearly show performance gains (or otherwise) from cold air intakes, extractors, exhausts, camshafts, turbo boost controls and engine management changes. If you cannot record an improvement by the use of stopwatch, it’s because there isn’t any improvement.
It’s not sexy like a dyno graph, but using a stopwatch is much more useful and costs nothing at all...
Next week we’ll look at using other cheap instruments to measure performance