The proliferation of roller dynamometers, also known as “chassis dynamometers,” makes it very practical to measure a vehicle’s power and compare it with other cars.
Many assume that results obtained on one roller dyno can be compared with those from another dyno. Also, that the power indicated by a chassis dynamometer reflects the true power of the car.
We must consider the question of measurement accuracy, and how sensitive it is to small changes made to the engine.
In this article, we will try to analyze what lies behind the readings delivered by chassis dynamometers.
Types of Chassis Dynamometers
Currently, there are two types of chassis dynamometers:
The purely inertial ones and the rollers that have a dynamometer brake (electric or hydraulic)
Purely inertial rollers can only be measured in wide-open throttle condition, at maximum acceleration and always in acceleration. On the other hand, rollers with brakes can perform stabilized tests at constant RPM, in intermediate throttle positions, and can even simulate certain loads such as aerodynamic forces.
How Do Inertial Rollers Work?
Inertial rollers estimate the energy applied to an inertial flywheel (the roller itself) and the speed at which it is accelerated. Knowing the inertial mass of the roller, it will tell us the power that is being applied to accelerate it. From this power, other data are derived, such as the applied torque.
Inertial rollers can only calculate power during acceleration. They cannot perform tests under constant load. This reduces the possibilities of use for certain applications such as injection mapping and timing advance under different load conditions. It also happens that because the weight of the roller cannot be changed, we are always tied to a fixed acceleration speed, which depends on that weight and the power of the engine. This means that if the roller is very light or very heavy, the real conditions of the engine are not simulated, for example, to achieve sufficient pressure in the turbo.
How Does a Roller with Dynamometer Work?
In this case, the roller behaves in an inertial manner during acceleration, but they are usually designed with less inertia since most of the power delivered by the vehicle is transferred to a power absorption unit, such as an eddy current dynamometer or a hydraulic one. This dynamometer can dissipate that delivered power and even oppose resistance to the vehicle so that it does not accelerate and remains at constant RPM.
At the same time, this power absorption unit creates a resistance torque that opposes the vehicle’s torque. That torque is measured by a load cell, which transmits the information to a computer. The reading of torque and RPM will allow us to calculate the power absorbed by the dynamometer brake, which is proportional to the power delivered by the engine (but is not the same).
Something important to keep in mind is that dynamometric measurement can only be performed at constant RPM. If the RPM is increasing, it means that a percentage of the power is being used to accelerate the inertial masses, so that power will not reach the dynamometer brake and will not be measured in the load cell. This causes the dynamometer brake to indicate less power in acceleration situations than in constant RPM conditions. The most modern equipment combines inertial and dynamometric measurements to compensate for this situation and allows obtaining similar readings in conditions of acceleration and load at constant speed (or in steps).
What is the Difference Between the Power Indicated by a Chassis Dynamometer and that Indicated by an Engine Dynamometer?
There are many conditions that influence the accuracy of measurement in any type of dynamometer, both in chassis dynamometers and engine dynamometers:
- the temperature of the test room
- the airflow in the intake
- the barometric pressure
- the system calibration
But in chassis dynamometers, there is a set of additional variables that affect the measurement and are very difficult to control and compensate for.
The main factor is transmission power losses. For example, the gear in which the test is conducted affects to a greater or lesser extent, with the direct (1:1) transmission or as close as possible to this condition usually being where there will be less loss in the gearbox. Other factors are the temperature of the gearbox or differential fluids, the acceleration speed, and the parasitic inertias introduced by the transmission elements that add to the roller’s own inertia, brake friction, the way the vehicle is tied to the roller, the weight that appears on the axle, and the deformation suffered by the tires, for example, due to different weights, inflation pressures, different rubber materials, and their slippage on the roller.
There’s a very interesting analogy for this, which is that measuring on a chassis dyno is like wanting to know your weight while dressed.
Sometimes you weigh yourself with shoes, other times barefoot, sometimes in underwear, and other times with a coat. You’ll never know your real weight.
Although software makes use of different estimates and additional measurements, it will never be possible to know the engine power using a chassis dynamometer, since these losses are not exactly measurable.
Some people use a rule which is to take around 10% of the power as loss in the drivetrain. But this is obviously an estimate with a lot of error, since I can have two similar vehicles, with the same weight, but one with 200HP and another with 500HP. Using the percentage, I will get very different losses, when reality says that if the vehicles are similar, it is likely that the losses are similar between them. Transmissions can differ abyssmally in losses, both in their construction and adjustment, as well as in the quality of their parts. This will also make the loss values different.
Trying to reach the engine numbers is a fantasy that distracts us from our true objective. The objective is to put more power to the ground, and that’s where we should focus. We can increase that number by improving engine power, it’s true, but also by performing optimization and improvement work on the parts that make up the vehicle’s transmission.
If you’re still interested in reaching accuracy values, the only way to achieve it is by having an engine test bench and a roller with an absorption dynamometer. You would have to perform a stepped test at stabilized RPM on both benches. Take into account that the air supply to the engines is adequate in both, the oil and engine temperature, the climatic conditions, and the conditions of the test room (ventilation and contamination), as well as the same exhaust system and length of exhaust pipes. That would be the only way to know the real losses in the transmission. Complicated, isn’t it?
What are the Main Factors that Affect the Accuracy of a Chassis Dynamometer?
There are different ways to look at accuracy. On one hand, there is the accuracy of the measurement system itself (sensors, electronics, software), and on the other hand, there are external elements that cause changes in the power that reaches the ground and is measured by that measurement system.
We start from the basis that the power read by the system is repetitive and precise (this is where you should find out which brand gives you that precision). Let’s now consider the external elements.
Among these elements, the transmission elements greatly influence a roller dynamometer. The contact between the tire and the roller is the first point where errors can appear. The type of rubber, its pattern, inflation pressure, material, rubber temperature can affect if one is not careful with the test conditions. The weight on the tire or the tension with which the vehicle was tied will cause the tire to deform more and therefore consume more power, which will disappear from the final reading, generating errors.
If we move towards the engine, factors such as intake air, both in flow and pressure, can significantly affect. If the exhaust gases are not evacuated from the room, that contamination re-enters the intake and also causes the engine to develop less power. Even the way the vehicle driver presses the accelerator can influence the final result if care is not taken.
In the transmission, the elements heat up, the oil becomes more viscous, the bearings heat up.
In vehicles with electronic control units, there are numerous control points where the computer intervenes and makes changes to the engine according to conditions. Starting from the most basic correction, which is the adjustment of the mixture according to barometric pressure and intake air temperature, to other corrections such as turbo pressure, intercooler temperature. Nowadays, in vehicles with traction control, if there is no way to deactivate it completely, on single-traction rollers, it cannot be measured properly since it will detect slippage in the wheels that do not rotate and automatically disconnect the engine power.
For this reason, multiple tests are always sought to be performed on the vehicle to obtain a profile of the repeatability of the tests. If between tests I get differences of 3HP, then that is my limit to detect changes; I cannot achieve greater precision than that. The more controlled the environment and the test, the smaller that number will be, the more repetitive the tests between themselves, and the finer the changes that I will be able to achieve in the engine. Even so, it is likely that small changes, of less than 5% of the total power, are difficult to detect in typical installations, since it is difficult to detect if the improvement is due to real changes made to the engine, or to momentary effects of the environment or conditions.
How Much Does the Transmission Ratio Affect?
There are differences between gears. Each has its own moment of inertia, its own friction, and there will also be an effect on the slippage of the tires on the roller.
In lower gears, there is a greater reduction, so the gearbox tends to be less efficient, producing more losses and consequently, less power reaches the wheel. In addition, there tends to be more wheel spin.
In theory, the same power should be obtained in any gear, but there is another factor to take into account. The speed at which the vehicle accelerates is directly affected by this ratio, so that the faster it accelerates, the more energy is needed to rotate the engine flywheel, and that energy does not reach the wheel. That’s why it is recommended to perform the tests as close as possible to a direct 1:1 transmission.
How Do Tires Influence the Results?
Tires should be considered as part of the test conditions, and if you change the test conditions, you change the results. This change can range from physically changing the tire (width, diameter, pattern, etc.), to changing its state, for example, changing the inflation pressure or the vehicle weight on the axle. Wheels influence 1 to 3 percent of the total measured.
Another factor to consider is that tires expand due to centrifugal force and this generates a change in the total transmission ratio.
How Can You Cheat on a Chassis Dynamometer?
There are many ways, some are intentional and others are the result of poor test conditions. The more test factors are controlled, the more accurate it will be.
For example, changing the way the vehicle is tied can greatly affect the result. The greater the downward force, the greater the deformation in the tires and therefore, less power will appear on the ground. Changing tire pressure, testing with the engine too cold or too hot, changing parameters in the software in general, such as lying about the moment of inertia, moving the engine fans away or closer can also affect the measurement.
Although dynamometers that have a dynamometer brake are potentially the most accurate, they are also vulnerable to alterations. The first point to consider is the correct calibration of the dynamometer’s load cell.
It is said that purely inertial dynamometers are the most repetitive, since power is only affected by the moment of inertia value of the rotational mass. This is very true, and really very small changes can be observed, however, it must be taken into account that repetitive does not mean accurate. This is because the power indicated by that type of dynamometer in many cases does not represent the real power of the vehicle. If what is sought is precision, the best results will be achieved in dynamometers with absorption brake.
Contrary to popular belief, my advice is to be wary of test benches that deliver very high numbers. It is most likely that a roller that is delivering rather low values is more accurate. Always ask for as much information as possible about the test (test conditions, temperature, humidity and ambient pressure, SAE correction factor used, moment of inertia of the roller). In any case, it must be taken into account that roller dynamometers are not useful for comparing numbers between one installation and another. The differences are important. The chassis bench should be considered as an element to make incremental changes and be measured on the same vehicle, in the same test place.
Real World Data
An article published by Hot Rod magazine in May 2004, compared the readings of the main dynamometer manufacturers in the United States, measuring the same vehicle in different installations.
The results, as expected, are very disparate from each other. For example, dynamometers with absorption brake marked differences of about 18 HP less than purely inertial dynamometers.
Power indicated by the manufacturer: 260HP (at the engine)
Measured powers (at the wheel)
- Dynapack (absorption): 208.9 HP
- Dynojet (inertial): 226.9 HP
- Mustang (absorption): 219.1 HP
- Superflow (inertial): 225.0 HP
Conclusions
As can be seen, not even the leading brands in test benches can agree on the indicated power. These are even values very far from the manufacturer’s data. The recommendation is to always perform the tests in the same place, and under the same test conditions. In this situation, it makes no sense to compare results between test benches, nor does it make sense to try to get the roller bench to mark the engine power, since it is equipment to measure power at the wheel and the power loss is not measurable (not even in a deceleration test).
However, our experience tells us that rollers are very sensitive elements and if they are well built, and have a high-precision measurement system, they are capable of measuring very small power differences, which allows making small changes to the engine and detecting them. At the same time, it is equipment that behind a very simple operation, hides a complexity in all the elements that influence the measured power, and it is those factors where the bench operators can make the difference between a serious and precise measurement or an approximate measurement that ends up generating more doubts than certainties. When working with reliable and precise tools, the training and suitability of the operator ends up being the most critical element of the roller.