• Home
  • traffic
  • Test bench: the pitfalls of performance measurement

Test bench: the pitfalls of performance measurement

Rossen Gargolov
Technology of modern dynamometers
Subscriptions & booklets

W he measures, measures dung - the old basic law of measurement technology demands to always question everything that has been measured. Certainly rightly, and of course that also applies to us - for sport auto.

Porsche 911 Turbo S as an occasion

Since our measurements are public and the manufacturers are directly affected by them, will remain Smallest mistakes never go undetected or without comment. This is what happened in spring 2014: In Supertest of the Porsche 911 Turbo S (sport auto 2/2014) the performance measurement caused frowns in Weissach. 607 PS instead of the specified 560 PS - a full deviation of 8.4 percent.

Our performance measurements have never been questioned before, no matter what values ​​the test stand spit out. But one shouldn't be so presumptuous as to exclude measuring. Porsche insisted on clarification, and we took this as an opportunity to kneel deeper into the matter.

Four questions had to be answered: How do roller dynamometers work? What do you have to pay attention to? Where are possible sources of error? And finally: Did the Porsche have too much power, did we measure or - more precisely - calculate too much?

Like a volcanic eruption on Edge of the crater

Let's start the story with Adam and Eve: What does a dynamometer do, how does it work? On-site meeting with the market leader Maha in Haldenwang. Michael Pleinies, the company's test bench expert and training manager for test bench technology, straps a Porsche 911 Turbo S onto the in-house test bench. 'Des ham wa glei, be scho seha.' Protective goggles on, ear plugs in - and please don't stand around stupidly behind the car during the measurement, because if there are still small stones stuck in the tire tread, they will be catapulted backwards at up to 300 km /h.

That sounds like somewhat exaggerated? Are you kidding me? Are you serious when you say that! Anyone who has experienced a professionally performed performance measurement live knows that the process harbors a certain drama: The full acceleration from the initial speed of a good 50 km /h to the achievable top speed in the test run is roughly like looking at a volcanic eruption from the edge of the crater. The turning of the wheels, the revving of the engine, the whirling of the rollers increasewith increasing speed to a deafening cacophony of hissing, whistling, screeching, whimpering and yowling - until at the end you voluntarily move a few meters away from the action in order to protect your hearing.

The tension belts are stretched to tear The elemental force of the rotating acceleration is even more haunting than sitting behind the wheel of a super sports car. As an onlooker, you don't even want to imagine what happens if a tire blows during the measurement. After a good 30 seconds, the spectacular spook is over. The measurement, however, is not - because now the drag power is determined when rolling.

Two types of roller test benches

Two hours later Michael Pleinies gives a lecture on the current state of the art in the Maha canteen. Wear-free eddy current brakes work in the roller test bench, which generate a counter-torque to the wheel force by means of electromagnetic force on the roller - this is how the vehicle is braked and the engine is loaded. As physics laypeople, we take it that way without being contradicted.

More practical is the hint that there are two types of roller test stand, the double roller test stand and the single roller test stand. The question of which concept will prevail has already been clarified: 'The single roller test bench is clearly the way forward for the future,' says Pleinies.

What are the differences? In the double roller test stand, the wheels turn between two rollers, a support roller and a braking roller, the diameter of which is smaller than that of the top rollers, which can be up to a meter in diameter. The wheels are guided in a stable manner in a lowered roller prism, but the system has disadvantages.

The rolling behavior of the tires is hardly comparable here with the conditions on the road. Because modern sports cars are getting stronger and stronger, there is a risk of too much slip on the double roller dynamometer. And roller slip is a big issue when it comes to reproducible performance measurements. 'On the double roller, you have to work with aids to minimize slip, for example by lowering the car, spraying the tires or working with additional weights on the drive axle,' explains Pleinies. However, the higher flexing work between the tire and the roller does not mean that the measurement is less precise, as this is compensated for between the wheel power and power loss measurement.

Less flexing work on the top roll

The future, however, belongs to the top roll test bench, in which the wheel runs on the top (or top) of a roll. 'That is much more realistic than with the double roller, you have less undesirable loads on the tires and the drive train,' explains Pleinies. The power transmission chain between the tire and the roller works much better, which means that the tire temperatures hardly exceed 65 degrees

Single roller test stands from the MSR family from Maha can cope with top speeds of 320 km /h and can absorb outputs of 260 kW to 1,100 kW or 1,500 PS per axle as a continuous load. The two roller sets with a diameter of 30 inches are each equipped with an eddy current brake and an electric motor. The 40 kW electric motors ensure that the vehicle axles run synchronously through electronically controlled coupling of the roller sets.

Enough gray theory. How does a correct measurement have to take place in practice? Every performance measurement starts with a thorough check of the test vehicle. 'The vehicle is checked for leaks or leaks,' says Moritz Müller, who uses his Maha test bench to carry out the performance measurements for the super test by sport auto in Backnang. The focus is on the tires, because the measurement means extra stress for them: 'The tires have to transmit the engine power, so the air pressure has to be right, and the camber shouldn't be too high, because that increases the grip on the roller and thus the measurement negatively influenced. '

Some things are self-evident, but should be mentioned again here: The fuel filled in plays just as important a role in the measurement as its temperature. If the fuel is low in octane and if it is also hot, this can have a significant negative impact on the measurement. The car should cool down before the test bench run because the engine tends to knock (uncontrolled self-ignition) after city traffic or full-load boom on the motorway, which causes the software for the full-load map to reduce performance, also because of component protection.

Reference values ​​from the road helpful

At sport auto there is an extra check on the road that is not part of the regular scope Performance measurement is a part: The OBD (onboard diagnosis) surface is used to determine reference values ​​such as intake air temperature and boost pressure for the test bench run.

'This ensures that we are working with correct data during the measurement,' explains Moritz Müller . 'If the intake air temperature is 50 degrees while driving at medium load and 200 km /h, for example, then we should work with similar parameters on the test bench - otherwise you will get incorrect readings.' For this purpose, the values ​​for throttle valve position, ignition angle and governed speed are also determined for naturally aspirated engines. To do this, the correct test gear is selected, i.e. the gear stage with which the measurement is made.

'The measurement is carried out in a high gear, but at the speed level so that the acceleration starts at around 50 km /h can ', explains Moritz Müller. In the 911 Turbo S, fifth gear is the optimal level, which allows a speed range of 50 to 240 km /h to be perfectly covered when measuring.

In the second stepthe vehicle is fixed on the test bench with tensioning straps in order to optimally adjust it on the roller. After the car has been tied up, conditioning takes place on the test bench. In addition to a test run in which it is checked that no tension occurs and the car is properly positioned, basic parameters must be selected in the software - such as the wheelbase or the expected acceleration performance.

Can the test driver influence the result?

In a 560 hp Porsche 911 Turbo S, the ramp value is 2.0 m /s, which is the average acceleration capacity of this vehicle corresponds. For this purpose, the weight value of the rotating masses is preselected, i.e. 80 kilograms for 19-inch wheels in the present case.

The measurement itself is divided into two parts: In the first step, the wheel power is determined via the acceleration in the test gear. If the speed limit is reached in the test gear, the clutch is disengaged or, in the case of automatic transmissions, the neutral gear is engaged and the drag power is determined when coasting. When measuring wheel power, the selected test gear is of particular importance for two reasons.

First, the vehicle must be able to accelerate cleanly from a starting speed of around 50 km /h. Automatic transmissions also have a minimum speed for each gear, which must be adhered to because otherwise the transmission will automatically shift down.

Rossen Gargolov
Despite stories to the contrary, the test driver hardly has any leeway to intervene in the wheel performance measurement.

Second The maximum permitted speed of the test bench must not be exceeded in the test run. Most single roller test stands today can cope with top speeds of up to 300 km /h.

Despite stories to the contrary, the test driver has little leeway to intervene in the wheel performance measurement, as Michael Pleinies from Maha explains: 'It is important that the driver accelerates quickly as soon as the starting speed is present in the test gear. However, full throttle does not mean triggering the kickdown, because otherwise the gearbox will downshiftEnd of revs, the test bench does the rest. When accelerating there is no possibility for the examiner to influence the result. '

Adding wheel power and towing power results in engine power

During the coasting process, i.e. the measurement of the towing power, the test driver could theoretically help by braking and thus pushing the result. 'But that is noticeable immediately,' says Pleinies, 'because the towing power curve has a parabolic course, and you can do it not to slow down a parallel force in order to increase the curve and thus the performance. The result would always be a rather atypical curve shape! '

The addition of wheel power and towing power gives the engine power. Why?' Power is mass times acceleration times speed, 'explains Pleinies.' There are mass fractions on the test bench and on the Vehicle. The test bench and vehicle are like a sandwich when taking measurements, the engine has to drive everything and all masses are a burden on the engine. The load of the brake on the rollers forms the static component, for this the centrifugal mass must also be accelerated as a dynamic component. '

The largest The enemy of measurement is slip

In plain language: On the roller dynamometer, the drive train of the vehicle and the rotating components of the dynamometer together create a number of losses that load the engine during the wheel power measurement. These static and dynamic losses are over The engine power is the result of the addition of the wheel power and the drag or power loss.

The terms wheel and power loss cause confusion time and again: Many see the measured wheel power and consequently believe that this is it Performance on the wheel, but that's wrong. The vehicle and the test bench form a unit: if you remove the car from the test bench, the losses and masses are also gone from the test bench d the wheel power on the test bench is smaller - therefore the drag power must be measured in order to add it up!

Modern test benches can actually measure the total losses exactly and reproducibly at any time. If so, if the framework conditions are met. And there are some: those that depend on the test bench and the environment, and those that are made mandatory by law.

A big enemy of measurement is slip: 'With sports cars, there is always slip on the Role, 'Pleinies knows,' and especially with turbo engines when the boost pressure builds up fully. But as long as that stays in the range of six to eight percent, that's within limits. ' Second, a precise measurement can only be made under realistic conditions. If the car is accelerated to over 200 km /h on the test bench but is not receiving adequate cooling, the measurement is not worth the paper onwhere it stands.

50,000 cubic meters per hour as the lower limit

'Cooling is the alpha and omega of the measurement, especially with turbo engines ', says Pleinies. The charge air must be cooled realistically, and oil and water coolers must also have a good and clean flow. To do this, the air also has to flow through the engine compartment, the transmission also needs cooling.

'In the event of heat build-up, the intake air temperature rises and the electronics reduce power,' says Pleinies. Curious: Sometimes all the values ​​for the engine are correct, but the end result still doesn't match because the gearbox control unit reports temperatures that are too high. As a result, the control unit reduces its performance. 'Often the gearboxes are completely packed, in V-engines with side catalytic converters it sits directly between them - it gets brutally hot,' said Pleinies.

When the car is lowered, there is no more air flowing under it - the cooling is zero. The cooling air fans in front of the car are also important: 'The goal is 50,000 cubic meters per hour with an air flow of 130 to 140 km /h as the lower limit for sports cars,' says Pleinies. There is no upper limit: '90,000 or 150,000 m /h, plus a flow of up to 200 km /h - that is no longer uncommon today! The electricity meter then runs a tad faster!' Professionals also use so-called component blowers, which are used in a mobile and variable manner, for example to target the inlet ducts for the charge air cooling.

Even if everything should fit, there are still external environmental factors such as air pressure and room temperature that must also be taken into account, and where, in some cases, legal requirements also have to be observed.

'Our test stands have an environmental module that records the environmental data plus the intake air temperature,' says Pleinies. Air pressure, room temperature and also the air humidity are recorded, because the EEC norm 80/1269 calculates the air pressure with dried air.

If old standards meet new technology

This EEC standard dates back to 1980 and was therefore developed for engines that did not have a map. At that time, the influence of environmental factors such as air pressure and intake air temperature was very far-reaching, especially with the vacuum cleaner. Therefore, the extrapolation was made - depending on where the test bench is. 'Today, however, we have map-controlled motors that partially adapt to such factors independently,' says Pleinies.

'The software in the engine control unit also includes other variables, for example component protection or emissions. The correction factor should Generally, of course, it's best to be as small as possible. '

Ergo, the temperature in the measuring room should be 20 to 25 degrees. 'If you can do that, you have less correction.' That leaves the air pressure, theof course depends on the altitude of the test bench. You cannot compensate for this artificially, otherwise you would need thick steel walls so that you could 'pump up' the space, so to speak.

In theory, the useful output could decrease by up to one percent with a height increase of 100 meters. With a test rig tolerance of two percent, this seems negligible - but between 700 meters and sea level there is still a full seven percent!

Minimal deviation of 3.6 hp in the 911 Turbo S

So, come back We are slowly returning to the original question, namely the measurement of the Supertest Porsche 911 Turbo S. It was measured at that time in Haldenwang near Maha, and not in Stuttgart. Haldenwang is at an altitude of 757 meters - therefore the calculation of the performance according to the EEC had to be corrected. The problem: the Porsche engine automatically corrects the altitude and thus the ambient air pressure in a range of up to around 1,000 meters. As? 'The engine control unit has a pressure sensor and detects the altitude,' says Christian Kunde, Head of Basic Engine and Combustion Processes at Porsche.

'This means that the ECU knows exactly how much boost pressure is required for the desired torque. The adjustment is made Via the VTG system of the turbocharger. The VTG is the control element to regulate the boost pressure. The compressor needs more speed to fill the air at altitude, and this larger amount of air has to be supplied by the turbine - and that is done via the Target charge pressure control regulated. '

Moritz Martiny, Head of Vehicle Application Boxer Engines at Porsche, adds:' The curve at that time in the Supertest was created because our turbo engine was treated like a naturally aspirated engine. With a naturally aspirated engine, you do that for one to enable standardized specification of the power as the engine would have at sea level, but this correction according to the EEC standard may not be applied to a modern turbo engine the. For older turbo engines, the correction may be justified because they work without the so-called target boost pressure control, which reacts to altitude difference and air pressure. '

To cut a long story short: What happened back then was a double correction. The control unit registered the altitude and adjusted the system via the VTG for the altitude, at the same time Maha corrected the result according to the EEC norm for the altitude - in this case double moping is no better!

The final measurement of a 911 Turbo S on three Maha test benches at different altitudes in Haldenwang, Stuttgart and Recklinghausen resulted - this time without correction - minimal deviations of 3.6 hp. Even Porsche was surprised that the measured values ​​were in such a narrow corridor of around one percent. 'At Porsche if the boost pressure is controlled by means of VTG, the filling remains the same, although environmental conditions such as the air pressure change, ”Michael Pleini sums up it. 'Do with itthen a correction according to EEC no longer makes sense. '

BMW and Audi are proceeding similarly today. Technical development has thus run away from the legal norms. The result: correctly measured, but incorrectly calculated.


Leave a reply

Name *