Horsepower claims made from DYNO Testing
#1
01-30-2016, 01:31 PM
Horsepower claims made from DYNO Testing
I think members will also find this interesting to read.
DYNAMOMETER TESTING AND
THE MODERN BMW ENGINE
By Steve Dinan
As dyno-testing facilities have become more and more abundant in recent years, increasing numbers of driving enthusiasts appear to be having their BMWs tested. While dyno equipment has become more sophisticated over the years, there still is no substitute for scientific testing procedures and a deep understanding of the factors that will affect the data obtained. No matter how well a dynamometer is designed, manufactured and supported, the information obtained may be meaningless, or at least misleading, if the operator does not have a clear understanding of the procedures necessary to duplicate “real world” conditions and the associated variables. Given the growing number of questions we receive regarding the results drivers are getting from independent tests, not to mention the inconsistencies, I thought that a bit of a technical discussion on the subject might be of interest to BMW enthusiasts.
A modern BMW engine management system is very sophisticated and has the ability to correct for changes in environmental conditions as well as fuel quality. If we go back fifteen years, engine control systems were not nearly so advanced and so power output was backed off from an engine’s full potential in order to ensure longevity. Now that engine control systems have advanced so dramatically, manufacturers can better manage conditions that might otherwise result in engine failure and they can produce more power per cc than ever before. While maximum horsepower has increased, so too has the variability of power output. This is because the engine control systems save the engine from failure by backing off power when conditions are less than ideal. This variability comes from the control system striving to extract whatever power is available, under a given set of conditions
As you would expect, all of our engine tuning products are dyno tested and the results published as part of our product descriptions. I’ll be the first to admit that Dinan’s dyno-numbers typically represent the lower of any claimed horsepower and torque increases for a given product type in the market. And in the case of performance engine management software for later model cars, we seem to be the only BMW tuner that has come to grips with the fact that there is simply no horsepower to be gained from engine management tuning alone. Even when it comes to performance engine components such as Cold Air Intakes, Super-chargers, exhaust systems and the like, Dinan’s published performance data is often lower than that claimed by our competitors. Does this mean that Dinan® products fall short when it comes to horsepower gains? Hardly. We are committed to providing the enthusiast with valid test data that is based upon the results of the very latest equipment, controlled testing procedures and years of experience. Our reputation among automotive journalists as the American BMW tuner that consistently provides realistic and verifiable performance data has been earned by exceeding expectations for nearly 25 years. My philosophy has always been to under-promise and over-deliver, despite whatever shall we say “optimistic” gains are being touted by the competition.
Please don’t get me wrong, I’m not suggesting that our competitors or the many dyno testing facilities are intentionally misleading driving enthusiasts by publishing false or exaggerated data. There are tuners that have an understanding of what needs to be done in order to obtain accurate data. I am of the opinion that even the tuners that publish questionable results are actually publishing numbers that reflect the very data that was obtained during their tests. The disparity lies in the dyno testing procedures employed and perhaps a lack of understanding with regard to the variables that can affect the data obtained. A prime example would be the fact that we get reports from customers who have obtained independent dyno results that range from more, to less, to no gain at all when compared to our published data.
Does this mean that most of the test data is invalid? Unfortunately, in most cases the answer is yes. I believe that this underscores the importance of controlling variables that can directly affect the results, as well as the importance of testing and re-testing in order to obtain a valid average. Truly scientific dyno tests are extremely time consuming, tedious and complex to perform when attempting to duplicate real world conditions in the dyno-room. I offer the following information in an effort to explain a bit about the equipment and the procedures we have developed over the past 25 years of BMW performance tuning, perhaps shedding a bit of light on the subject in general terms.
About Dynamometers
By far the most common types of chassis dynamometers employ large rollers that are turned by the vehicle’s drive wheels. When we relied on this type of equipment, we would replace the stock rubber with sticky, shaved tires in order to reduce slippage on the rollers. The rear wheels would also be set at maximum positive camber, further improving traction on the dyno rollers. The cars would then be tied down with four steel cables in order to further reduce slippage and provide increased safety. Two of the cables were focused on down-force in order to increase traction and the others positioned to hold the car in place on the rollers. In order to accurately read power output, cable tension was evaluated and adjusted in order to ensure that the tension was not too light, causing an artificially low reading due to slippage and that there was not so much tension that drag would negatively affect the readings. All tires slip to some degree, more so as power increases, therefore it is impossible to measure 100% of any gains that are achieved with this type of equipment. Accuracy can be improved with a roller type dyno by connecting a tachometer to the rollers and wheels, or a strobe light with markings so that the measured output may be adjusted accordingly.
Employing the very latest in chassis dyno technology, our current equipment eliminates the tire slippage issue as it connects directly to the drive wheel axles. By eliminating the tires and any related slippage, the results are far more accurate and repeatable, making this type of dyno superior in my opinion.
By far the most significant criticism I have for many dyno facilities is the use of fans that are simply too small for the job at hand. The fan size is so significant that we employ a very large unit that was actually designed for ventilation systems installed in high rise buildings! This powerful fan produces 38,000 cfm of air flow @ 75 MPH, which is still less than the 150 MPH air that a modern BMW might see at redline in 5th gear, but it certainly provides a closer to real-world scenario than the more common fans I have seen used in dyno facilities. It should be noted that 5th gear is used for our dyno testing because it is one to one, meaning that the input and output shafts are connected, reducing power losses and transmission wear. I have seen many examples of dyno facilities where low flow fans obtained from the local hardware store are employed, and even situations where there was no fan at all. A minimum of 15,000 cfm and preferably 40,000 cfm of air flow is required for proper heat exchanging. This type of fan will produce a 40-80 mph air discharge velocity.
A lack of air-flow during dyno testing will almost always alter the fuel mixture in the rich direction as the radiator cannot exchange enough heat, resulting in the computer compensating by retarding timing and richening the fuel mixture to prevent the engine from overheating and detonating. In addition, the intake air sensor will read substantially higher temperatures than that seen on the road with proper airflow. This issue is particularly important to address when testing high output cars like the M5 or M3, and even more so on forced induction cars with intercoolers as the heat exchanger is not able to cool as efficiently because of the reduced air flow. The engine compartment is normally flushed with air driving down the road, particularly at speed, cooling the manifolds and other associated engine components. Cooler engine components and lower air intake temperatures will result in a leaner air/fuel mixture and ignition timing will be advanced, invariably resulting in greater power on the road than on the dyno. In simpler terms, accurate measurements can only be achieved when the dyno tests are conducted in a manner that simulates the car driving down the road, in as much as is possible.I believe that the rather large horsepower gains that are being published by some, particularly with regard to “power chips”, are the result of tuning the cars back to the stock mixture and ignition timing settings, essentially leaning-out the air/fuel mixture and advancing the timing to compensate for the rich mixture and retarded timing experienced on the dyno. It appears to me that this “increase” in power is then included in whatever gains were actually achieved (if any). In reality, these supposed gains are nothing more than a correction for the testing conditions, resulting in an exaggerated performance claim. In addition, many “power chips” create the perception of an increase in power/acceleration as the re-programming will often dramatically increase the speed of the throttle opening on the drive by wire cars, making the engine feel more powerful.Dyno Testing Variables and How to Reduce Them The procedures that we have developed are the result of many years of experience, extensive research and a very real desire to obtain the most accurate data possible. Following is a relatively detailed description of the variables and what we do to reduce the impact on dyno results.Testing the Dinan® S2-M5 The first step is to prepare the vehicle for a “baseline” run. The 91-octane premium pump fuel (the highest octane currently available in California) is replaced with 93-octane fuel, as it is the most common premium pump fuel available in the U.S. In order to ensure that the vehicle being tested is a representative sample, it is inspected for any defects that might affect performance, including tests related to oil consumption, leak down. In the event that a defect is discovered, the car is then repaired accordingly or in some cases even rejected for such testing. The last thing we want to do is test and tune a car exhibiting any sort of issue that might negatively affect the car’s performance in stock or modified form.
Top: Air Fuel Ratio Recorder Bottom: Thermal Couple Recorder
Before we place a car on the dyno, we install sophisticated data collection equipment that has been designed to measure the air/fuel mixture, ignition timing the engine’s air intake temperature; the engine block and radiator coolant temperatures; as well as the engine, transmission and differential oil temperatures. These measurements are conducted on the road, under normal driving conditions and are only recorded once the temperatures have stabilized at what would be considered normal operating temperature. Once the temperatures are stabilized, we record the data from 2,000 rpm to redline at wide- open throttle. Again, the data is collected with the transmission in the gear that is one to one, typically 5th, so that the input and output shafts are connected, reducing power losses and wear on the transmission. The data is recorded several times in order to obtain a solid average.
Sample Data Aquisition Display
Next, the times necessary for the stock car to achieve various speeds are recorded on the road and then loaded into the dyno program in order to simulate road conditions.
The air intake sensor is located under the hood of the E39 M5, absorbing heat produced by the engine. As you can see in Figure 1 below, on the road the sensor absorbs heat from the engine, artificially raising the reading the computer sees to 110° F. As soon as you accelerate at wide open throttle, the ram air coming into the engine flushes out the hot air and cools the sensor. By the time the engine reaches high rpm the temperature sensors are seeing 85°, very close the 80° F ambient temperature that was recorded during the test run.
Figure #1: E39 M5 Intake Air Temperature Measured on the Road
Referring now to the stock radiator and engine block temperature graph below (Figure 2), you can see that BMW’s engineers did an excellent job of maintaining consistent temperatures in both the radiator and block. As the engine revs friction is increased, as well as load and cylinder cycles. All three of these things produce more heat but as the engine revs, the car is moving through the air faster which rams more air through the radiator, exchanging additional heat and therefore maintaining a very stable temperature.
Figure #2: E39 M5 Engine and Radiator Temperatures Measured on the Road
As you will soon see, once we put the vehicle in the stagnant air in the dyno room, all of this changes. The first dyno run we will talk about (Figure 3A – purple line) represents what I would consider the worst dyno testing procedure I have ever seen. The vehicle was placed on the dyno with the hood closed and a small fan positioned in front of the grille, typical of the fans I have seen in most chassis-dyno facilities. The engine is warmed up to normal operating temperature by performing two passes and then the car is left to idle for 10 minutes.
Figure #3A: E39 M5 Dyno Graph
This combination of conditions resulted in the lowest recorded output on the graph, producing 335.7 hp. During the two warm up passes the engine, radiator, intake manifolds, and air intake sensor will “heat soak”, resulting in a reduction of power. Let’s take a moment to look at each graph separately and analyze what has happened.
When the M5 was on the dyno with the hood closed, and there was no ram air or air flowing under the car to evacuate heat from the engine compartment, the under-hood temperature rose significantly beyond what would be seen under normal driving conditions. As you can see in Figure 4 – pink line, the temperature reached 160° F from idling and even after the wide open throttle run was completed the temperature only dropped to 148° F! Comparing this to the road graph (Figure 4-blue line) shows a staggering increase in the temperature reading even though the outside air temperature has not changed at all. Normally when the computer sees a higher air temperature reading, it leans the air/fuel mixture and retards ignition timing in small amounts to compensate for the less dense air. This occurs in order to maintain a proper air/fuel mixture and prevent detonation. However, when the M5 computer sees a very hot value it goes into a portion of the program designed to save the engine from melt down. This mode dramatically richens the mixture and retards the timing, preventing engine damage in two ways: it causes the engine to produce less power, thereby producing less heat; and some of the heat is actually absorbed by the fuel, then carried out through the exhaust. In addition, the rich mixture and retarded timing ensures that the engine will not detonate under these conditions.
Figure #4: Intake Air Temperature Sensor Comparison
The radiator cannot exchange heat as well as it did on the road because there is not enough ram air flowing through it. Looking back at the road temperatures depicted in Figure 5 – blue line, we can see that the radiator stays between 175° F and 178° F. Now compare those readings to the radiator temperature from the dyno acceleration run in Figure 5 – violet line. With the standard dyno fan you can see that the temperature starts at 173° F and climbs to 210° F!
Figure #5: Radiator Temperature Comparison
Since the radiator cannot exchange heat on the dyno as efficiently as it did on the road, the engine block heats up significantly. Looking at the road temperature curve in Figure 6 – blue line, you can see that the engine block temperature stays between 191° F and 188° F. However the engine temperature measured during the dyno acceleration run with a standard type of fan starts at 192° F and ends at 203° F (Figure 6 – violet line). It should be noted that the coolant temperature gauge in the instrument cluster will hardly move in this case, even though the engine management system needs to correct for the increase, as 203° is not a high enough temperature to cause the vehicle to actually overheat.
Figure #6: Engine Block Temperature Comparison
The combination of the engine heating up and the air temperature sensor reading an artificially high value causes the engine management computer to go into the “engine savior mode”. The mixture is richened to an astounding 9.5 to 1 air fuel ratio (see Figure 7 – violet line); whereas the correct mixture measured on the road was 12.2, shown in Figure 7 – blue line. The ignition timing is retarded from a peak value on the road of 27° (see Figure 8 – blue line) down to 15° (Figure 8 – violet line). It is truly amazing how intelligent a modern BMW is. If you were to go back just 15 years, these same conditions would likely result in engine damage!
Figure #7: Fuel Mixture Comparison
Figure #8: Ignition Timing Comparison
We will begin to eliminate these variables one by one so that you can see which aspects are the result of the intake air temp sensor and what portion is attributable to the engine running too warm.
The next dyno run was performed with only one change: the hood was opened!
As you can see in Figure 3B, with this simple change the power increased by approx 35 hp to 370 hp.
Figure #3B: E39 M5 Dyno Graph.
The air intake sensor now absorbs a lot less heat from the engine (see Figure 9 – yellow line). With the hood open, the temperature reached 130° F, as compared to the 160° reading with the hood closed (see Figure 4). After the wide open run was completed, the temperature dropped to 120° F (Figure 9 – yellow line), compared to 148° F (Figure 9-blue line). However, this is still significantly warmer than the temperatures that were measured on the road.
Figure # 9: Intake Air Temperature Comparison- Hood open vs. closed
Looking at the radiator temperature during the dyno acceleration run, using the standard dyno fan and the hood open, you can see in Figure 10 – yellow line, that the temperature is able to cool off more between runs. This enables us to start the run with a temperature of 155° F, ending at 205° F! Once again, this is still substantially warmer than the temperatures measured on the road (Figure 10-blue line)
Figure #10: Radiator Temperature Comparison
Since the radiator still cannot exchange heat as efficiently as it did on the road, you can see in Figure 11 – yellow line, that the temperature starts at 182° F and ends at 197°, still warmer than the road test.(Figure 11- blue line)
Figure #11: Engine Block Temperature Comparison
While we have made some serious progress here, and by now it should be obvious that you should never dyno a car with the hood closed, we are still significantly short of duplicating the normal conditions the car would see on the road. The improvements we have made thus far have leaned the air/fuel mixture from 9.5 to 1 (Figure 7-violet line) to 11.2 to 1 (see Figure 12 – yellow line); however compared to the correct road mixture of 12.2 to 1 (Figure 12 – blue line), the air/fuel mixture is still too rich. Ignition timing has also improved from 15° (Figure 8-violet line) down to 22° (Figure 13 – yellow line). This is still substantially less than the peak road value of 27° (Figure 13 – blue line) that would occur during actual road conditions. I believe that this is how most “power chips” are made, essentially leaning the mixture and advancing the ignition timing back to normal values. This will result in a measured power increase, however this gain is not real because it is merely compensation for the dyno environment.
Figure #12: Fuel Mixture Comparison
Figure #13: Ignition Timing Comparison
The hood will remain open for all subsequent dyno runs. The next dyno run was performed with only one change: disconnection of the stock air temperature sensor and the installation of one at the air inlet. This is done in order to get the sensor to accurately reproduce the temperature of the air actually going into the engine. This will stabilize the engine tremendously and result in the computer making proper corrections for the conditions. As you can see in Figure 3C, the power increased by approx 10 hp, with the run producing 380 hp, with significantly less fall off at higher rpm.
Figure #3C: E39 M5 Dyno Graph
The air intake sensor is now rock steady at 81° F (see Figure 14 – light blue line). This matches the actual ambient temperature in the room at this time and is more stable than the temperatures measured on the road (Figure 14 – dark blue line). This should help to explain why we decided to move the intake air temperature sensor as part of our Cold Air Intake System.
Figure #14: Intake Air Temperature Comparison
The radiator and engine block temperatures remain the same as the previous run since we are still employing the small fan.
While we have made even more progress, we are still significantly short of the normal conditions the car would see on the road. The mixture has leaned out to an 11.7 to 1 air fuel ratio (see Figure 15 – light blue line) when compared to the 12.2 to 1 ratio measured on the road (Figure 15 – dark blue line). Ignition timing has been retarded from a peak value on the road of 27° (Figure 16 – dark blue line) down to 22° (Figure 16 – light blue line), but it is less stable. This is due to the cooler air intake sensor value causing the computer to lean out the fuel mixture and advance timing. Since the radiator still can’t exchange enough heat, the engine runs warmer than it normally would, causing the engine to detonate which in turn set off the knock sensor causing radical spikes (see Figure 16 – light blue line). The remaining richening and retarded ignition timing shown on these graphs are resulting from radiator and engine temperatures, not the engine intake air temperature reading since it has been stabilized.
Figure #15: Fuel Mixture Comparison
Figure #16: Ignition Timing Comparison
The last step in our attempt to reproduce actual road conditions is to employ a fan large enough to exchange enough heat that air temperatures at the end of the run exactly match the temperatures recorded on the road. The only change for this run was to replace the small fan with the Dinan® (Level 1) “Hurricane” fan.
As you can see in Figure 3D below (violet, blue and yellow lines), the power increased by approximately 30 hp to 411.4 hp, backed up by a 410 hp run. When the conditions are truly controlled you can usually produce runs with as little as 0.5 – 1.0% variance.
Figure #3D: E39 M5 Dyno Graphs
The air intake temperature sensor was stabilized on the last run, so now let’s look at the radiator and block temperature graphs. As you can see, the radiator temperature during the dyno acceleration run (Figure 17 – violet line), using the very powerful fan and the hood open, allowed the temperature to cool off more between runs. This enabled us to start the run with a temp of 110° F and end at the exact same value as we saw on the road, 175° F (Figure 17 – dark blue). Even with this huge fan, the largest I have seen on a chassis dyno, we still must start at artificially low numbers so as not to exceed the road value by the end of the run! In other words, even our huge 75-mph fan can’t duplicate the air flow the car would see on the road…but we are getting closer!
Figure #17: Radiator Temperature Comparison
We are able to match the engine temperature recorded on the road at the end of the dyno run as well (Figure 18 – violet line). Figure 18 shows the engine temperature during the dyno acceleration run with the large fan. You can see that the temperature starts at 182° F and ends at 188° F equal to the road values (Figure 18 – blue line).
Figure #18: Block graph
The air/fuel mixture and ignition timing now match the road numbers almost exactly (see Figures 19 and 20 ).
Figure #19: Fuel Mixture Comparison
Figure #20: Ignition Timing Comparison
The important thing here is to match the engine temperature as closely as possible. As we have been able to achieve just a 7° variance (see Figure 18) from the beginning to the end of the run, with an ending value that is the same as the road number, the mixture and ignition timing match the road value. Now we know that we have real-world horsepower number. Remember that we have not added or changed any parts on the car during the course of this testing, with exception of the temperature sensor.
Even still, we fall short of duplicating actual road conditions in one area: we can’t reproduce ram air to the intake system. We are currently developing a system that will produce enough ram air to get us even closer, but until it is completed we will still measure less power on the dyno than the vehicle will actually make on the road!
Engine Wear
Every engine will produce different power output, even if every variable is carefully controlled! This is due to production tolerances as well as maintenance and care during break-in. Most of the variances occur from cam timing errors and cylinder leak-down variances. Cylinder leak-down variances are the biggest variable. A desirable number for leak-down is less than 5%; however, it is very common to see numbers higher than that. It mostly depends on how the engine was broken in and, of course, maintained. The S2 M5 engine used in this test had an average leak-down just over 4%. The same engine had an average leak down of 3% one year ago. Comparing the previous dyno runs on this engine to more recent runs (see Figures 3E and 21), you can see a loss in power of approx 1%. This is the same car on the same dyno! However, due to normal wear, the power has dropped by 6 hp. You must realize that not every engine of the same type produces the same power. You also must realize that not every engine produces the same power throughout its life. BMW engines are very well manufactured with very consistent tolerances. Our dyno test show that almost all engines, without any defect, will be within a 5% window with the vast majority being within a 3% window.
Figure #21: Year old E39 M5 Dyno Graph
Figure #3E: Current E39 M5 Dyno Graph
Drivetrain Temperatures
In addition to variances from the engine itself and the dyno environment, more variables come into play as a result of varying drive train oil temperatures. The graph above (Figure 21) depicts the recorded rear wheel horsepower for the S2-M5. As you can see, as the drive train oils heat up during the four dyno runs, the recorded horsepower increases from 406 to 417; an eleven horsepower gain resulting from nothing more than increased drive train oil temperatures! Remember that the engine was already warmed up to normal operating temperature before the four dyno runs were conducted. The temperatures of the air intake sensor, engine block and radiator were strictly controlled during these runs so that the only variables were drive train oil temperatures. The run that indicates the least output at low rpm was the first one after the engine was restarted (See notation- figure 21). The reason for this is that when the engine is first started, the camshaft and ignition timing have a different program to assist in warming up the catalyst for emissions purposes. As soon as the aft O2 sensor determines that the catalizer is operating, it automatically reverts to the normal program. The colder the catalytic converter, the more runs will produce reduced output. You can see that the oils have reached normal operating temperature for the last two runs as the measured output is very similar. Thermal couples can be used to further improve this accuracy. Many Winston Cup teams are now using chassis dyno’s to reduce drive-train friction. Since improvements in this area are so small in order for this work to be valid they must strictly control the drive train temperatures.
Data from a Stock E46 M3
The M5 engine is extremely sensitive to temperature control when dyno testing. Any increase in power results in an increase in heat and a corresponding increase in sensitivity. Not every engine demonstrates the same levels or types of sensitivities; some have cooling system sensitivities while others are sensitive to ram air volume and fuel octane. Each engine must be tested to determine what conditions exist on the road that do not exist on the dyno and what must be done to correct for them. Since the M5 started the testing with 93-octane fuel, I wanted to provide an example of how octane affects horsepower. The E46 M3 is an excellent example of an engine that is sensitive to octane. It has a very high volumetric efficiency as well as a very high static compression ratio of 11.5 – 1. The engine being tested was a stock M3 engine. It was first warmed up and stabilized using the method described previously for the M5, running 91-octane fuel. As you can see in Figure 22, where the engine was warmed up and the previously discussed testing procedures applied, the stock M3 produced 280 hp (Figure 22- violet line). We then replaced the 91-octane fuel with 93 (available in most parts of the country). The M3′s computer was so quick to determine that the fuel had been improved that it only took four dyno runs for the timing to adapt to the increased octane and raise the power up to 291 hp (Figure 22 – light blue and yellow lines). A gain of 11 hp with just 2 points of octane. The M3 engine is equipped with a very good ram air system. While our large fan is pushing large volumes of air into the ram air duct, the volume and velocity of air seen by the car on the dyno is still less than the engine would see on the road. The power output drops off after 7350 rpm because we simply cannot duplicate the airflow the car would receive on the road in the dyno room, resulting in a loss of power. BMW claims peak power @ 7900 rpm. We have no doubt that if we could accurately reproduce the ram air that the M3 would receive on the road, our peak power would move up from 7350 rpm to 7900 rpm. One of our many engineering projects is to further enhance our ram air capabilities in the dyno room.
Figure #22: E46 M3 Dyno Runs
Stock E46 M3 Ignition Timing Adaptations
Figures #23 and #24 below are taken directly from the BMW factory diagnostic tool, demonstrating how ignition timing adapts to different fuel octane ratings. The same car is represented here, the only difference being the octane rating of the fuel. If you were to add 1° of ignition advance, the engine management system would detect it and retard the timing 1°. You can see that adding timing in the engine management software or “power chip” is futile because the computer will negate the change, as sufficient octane does not exist. However, you can see that adding higher octane fuel is like adding a “power chip” as the system adapts to the better fuel, making more power.
Figure #23: 91 Octane Knock Adaptations Diagnosis
Figure #24: 93 Octane Knock Adaptations Diagnosis
Intercooled Forced Induction Systems
Cars with intercooled forced induction systems (superchargers or turbo chargers) provide an even bigger challenge on the dyno. A separate fan must be employed in order to sufficiently cool the intercooler. Thermal couples must be installed in the inlet and outlet of the intercooler so that the temperature drop seen on the road can be measured. Once the temperature drop has been established, the fan speed must be adjusted until the same drop in temperature is maintained on the dyno as that was measured on the road. If the temperature drop cannot be achieved on the dyno, the error can be corrected for mathematically and the results will be very close.
Corrections
Once we have determined the specific baseline procedure for a vehicle, the car is allowed to sit until early the following morning when the temperature is as close as possible to 77° F, the SAE standard. There are formulas for temperature, humidity and barometric pressure corrections from the SAE; however the correction tables are not completely accurate for a digitally controlled car. This is due to the corrections the car’s computer is making based on the conditions previously discussed. We reduce this error further by performing all tests as close to 77° F as possible. An alternative would be to turn the corrections off in the software but this is a potentially dangerous approach. The most accurate results are obtained when the tests are performed in a climate-controlled dyno-room where temperature and humidity can be completely controlled for each test. By running our tests as close to 77° F as possible, we still must employ the SAE correction tables, but the amount of correction necessary is reduced and accuracy improved. Be aware that not all dynos correct for temperature, humidity and barometric pressure!
Conclusion
If you decide to test your car on a dyno, whether in stock or modified form, be advised that you will not see the same results as BMW or Dinan. Putting the time consuming and tedious procedures aside, any number of things can cause your measurements to be different from those published. Even in a best case scenario, assuming that there is no need to be concerned about calibrations because the performance software has been supplied, it still takes the better part of three days to go through the proper testing procedure and collect the necessary data.
If the engine is detonating or in the “savior mode” because of excessive temperatures, gains can not be measured. In fact if the car sits and heat soaks or cools for an excessive period of time between runs, enough variance can be created that the performance component enhanced car may show less power than the stock car, or even a very exaggerated gain.
By way of summary, following are some of the more significant factors that you should keep in mind when considering dyno testing in general terms, as well as what to look for in the facility itself.
1.) Each dyno will produce different results (even with the same brand of dyno).
2.) The octane rating of fuel varies in different parts of the country (you must have a controlled fuel supply).
3.) Cold weather increases the gains and hot weather decreases them, even with temperature corrections.
4.) Lack of oxygen from exhaust in a dyno room will cause a loss in power.
5.) Slipping tires on the rollers will reduce the measured gains.
6.) Inertia type dynos have a lighter load than the car will see on the road. This is especially true for cars with heavy drive trains because some of the power will get absorbed spinning the masses faster. The inertia correction programs employed in these types of dyno’s are not completely accurate.
7.) Fixed load dynos have a higher load than what the car sees on the road. This excessive load will result in a large mixture shift and the detonation sensor will be activated prematurely.
8.) No dyno can accurately simulate wind resistance, the ram air effect into the airbox or cooling of the intake tract under the hood.
9.) The size of the fan used during testing will change the power output.
10.) Oil temperatures will affect output due to changes in friction.
11.) The air intake temperature sensor will trigger adjustments to fuel mixture and ignition timing.
Dinan® is certainly not the only BMW tuner in the world that understands the variables and complexity of proper dyno testing and tuning. I have a great deal of respect for the handful of BMW tuners around the world that share our passion for accuracy. Unfortunately, this level of dyno testing sophistication appears to be the exception and simply won’t be found in common dyno facilities that rent their time. The purpose of this paper is to provide our customers with a deeper understanding of our test procedures and why they have been developed. We too are constantly learning more about the science and updating our equipment whenever significant improvements in the technology occur in an effort to provide our customers with the most valid data possible. In addition, Dinan also employs an engine dyno test cell, but we’ll save a discussion on that technology for another time.
DYNAMOMETER TESTING AND
THE MODERN BMW ENGINE
By Steve Dinan
As dyno-testing facilities have become more and more abundant in recent years, increasing numbers of driving enthusiasts appear to be having their BMWs tested. While dyno equipment has become more sophisticated over the years, there still is no substitute for scientific testing procedures and a deep understanding of the factors that will affect the data obtained. No matter how well a dynamometer is designed, manufactured and supported, the information obtained may be meaningless, or at least misleading, if the operator does not have a clear understanding of the procedures necessary to duplicate “real world” conditions and the associated variables. Given the growing number of questions we receive regarding the results drivers are getting from independent tests, not to mention the inconsistencies, I thought that a bit of a technical discussion on the subject might be of interest to BMW enthusiasts.
A modern BMW engine management system is very sophisticated and has the ability to correct for changes in environmental conditions as well as fuel quality. If we go back fifteen years, engine control systems were not nearly so advanced and so power output was backed off from an engine’s full potential in order to ensure longevity. Now that engine control systems have advanced so dramatically, manufacturers can better manage conditions that might otherwise result in engine failure and they can produce more power per cc than ever before. While maximum horsepower has increased, so too has the variability of power output. This is because the engine control systems save the engine from failure by backing off power when conditions are less than ideal. This variability comes from the control system striving to extract whatever power is available, under a given set of conditions
As you would expect, all of our engine tuning products are dyno tested and the results published as part of our product descriptions. I’ll be the first to admit that Dinan’s dyno-numbers typically represent the lower of any claimed horsepower and torque increases for a given product type in the market. And in the case of performance engine management software for later model cars, we seem to be the only BMW tuner that has come to grips with the fact that there is simply no horsepower to be gained from engine management tuning alone. Even when it comes to performance engine components such as Cold Air Intakes, Super-chargers, exhaust systems and the like, Dinan’s published performance data is often lower than that claimed by our competitors. Does this mean that Dinan® products fall short when it comes to horsepower gains? Hardly. We are committed to providing the enthusiast with valid test data that is based upon the results of the very latest equipment, controlled testing procedures and years of experience. Our reputation among automotive journalists as the American BMW tuner that consistently provides realistic and verifiable performance data has been earned by exceeding expectations for nearly 25 years. My philosophy has always been to under-promise and over-deliver, despite whatever shall we say “optimistic” gains are being touted by the competition.
Please don’t get me wrong, I’m not suggesting that our competitors or the many dyno testing facilities are intentionally misleading driving enthusiasts by publishing false or exaggerated data. There are tuners that have an understanding of what needs to be done in order to obtain accurate data. I am of the opinion that even the tuners that publish questionable results are actually publishing numbers that reflect the very data that was obtained during their tests. The disparity lies in the dyno testing procedures employed and perhaps a lack of understanding with regard to the variables that can affect the data obtained. A prime example would be the fact that we get reports from customers who have obtained independent dyno results that range from more, to less, to no gain at all when compared to our published data.
Does this mean that most of the test data is invalid? Unfortunately, in most cases the answer is yes. I believe that this underscores the importance of controlling variables that can directly affect the results, as well as the importance of testing and re-testing in order to obtain a valid average. Truly scientific dyno tests are extremely time consuming, tedious and complex to perform when attempting to duplicate real world conditions in the dyno-room. I offer the following information in an effort to explain a bit about the equipment and the procedures we have developed over the past 25 years of BMW performance tuning, perhaps shedding a bit of light on the subject in general terms.
About Dynamometers
By far the most common types of chassis dynamometers employ large rollers that are turned by the vehicle’s drive wheels. When we relied on this type of equipment, we would replace the stock rubber with sticky, shaved tires in order to reduce slippage on the rollers. The rear wheels would also be set at maximum positive camber, further improving traction on the dyno rollers. The cars would then be tied down with four steel cables in order to further reduce slippage and provide increased safety. Two of the cables were focused on down-force in order to increase traction and the others positioned to hold the car in place on the rollers. In order to accurately read power output, cable tension was evaluated and adjusted in order to ensure that the tension was not too light, causing an artificially low reading due to slippage and that there was not so much tension that drag would negatively affect the readings. All tires slip to some degree, more so as power increases, therefore it is impossible to measure 100% of any gains that are achieved with this type of equipment. Accuracy can be improved with a roller type dyno by connecting a tachometer to the rollers and wheels, or a strobe light with markings so that the measured output may be adjusted accordingly.
Employing the very latest in chassis dyno technology, our current equipment eliminates the tire slippage issue as it connects directly to the drive wheel axles. By eliminating the tires and any related slippage, the results are far more accurate and repeatable, making this type of dyno superior in my opinion.
By far the most significant criticism I have for many dyno facilities is the use of fans that are simply too small for the job at hand. The fan size is so significant that we employ a very large unit that was actually designed for ventilation systems installed in high rise buildings! This powerful fan produces 38,000 cfm of air flow @ 75 MPH, which is still less than the 150 MPH air that a modern BMW might see at redline in 5th gear, but it certainly provides a closer to real-world scenario than the more common fans I have seen used in dyno facilities. It should be noted that 5th gear is used for our dyno testing because it is one to one, meaning that the input and output shafts are connected, reducing power losses and transmission wear. I have seen many examples of dyno facilities where low flow fans obtained from the local hardware store are employed, and even situations where there was no fan at all. A minimum of 15,000 cfm and preferably 40,000 cfm of air flow is required for proper heat exchanging. This type of fan will produce a 40-80 mph air discharge velocity.
A lack of air-flow during dyno testing will almost always alter the fuel mixture in the rich direction as the radiator cannot exchange enough heat, resulting in the computer compensating by retarding timing and richening the fuel mixture to prevent the engine from overheating and detonating. In addition, the intake air sensor will read substantially higher temperatures than that seen on the road with proper airflow. This issue is particularly important to address when testing high output cars like the M5 or M3, and even more so on forced induction cars with intercoolers as the heat exchanger is not able to cool as efficiently because of the reduced air flow. The engine compartment is normally flushed with air driving down the road, particularly at speed, cooling the manifolds and other associated engine components. Cooler engine components and lower air intake temperatures will result in a leaner air/fuel mixture and ignition timing will be advanced, invariably resulting in greater power on the road than on the dyno. In simpler terms, accurate measurements can only be achieved when the dyno tests are conducted in a manner that simulates the car driving down the road, in as much as is possible.I believe that the rather large horsepower gains that are being published by some, particularly with regard to “power chips”, are the result of tuning the cars back to the stock mixture and ignition timing settings, essentially leaning-out the air/fuel mixture and advancing the timing to compensate for the rich mixture and retarded timing experienced on the dyno. It appears to me that this “increase” in power is then included in whatever gains were actually achieved (if any). In reality, these supposed gains are nothing more than a correction for the testing conditions, resulting in an exaggerated performance claim. In addition, many “power chips” create the perception of an increase in power/acceleration as the re-programming will often dramatically increase the speed of the throttle opening on the drive by wire cars, making the engine feel more powerful.Dyno Testing Variables and How to Reduce Them The procedures that we have developed are the result of many years of experience, extensive research and a very real desire to obtain the most accurate data possible. Following is a relatively detailed description of the variables and what we do to reduce the impact on dyno results.Testing the Dinan® S2-M5 The first step is to prepare the vehicle for a “baseline” run. The 91-octane premium pump fuel (the highest octane currently available in California) is replaced with 93-octane fuel, as it is the most common premium pump fuel available in the U.S. In order to ensure that the vehicle being tested is a representative sample, it is inspected for any defects that might affect performance, including tests related to oil consumption, leak down. In the event that a defect is discovered, the car is then repaired accordingly or in some cases even rejected for such testing. The last thing we want to do is test and tune a car exhibiting any sort of issue that might negatively affect the car’s performance in stock or modified form.
Top: Air Fuel Ratio Recorder Bottom: Thermal Couple Recorder
Before we place a car on the dyno, we install sophisticated data collection equipment that has been designed to measure the air/fuel mixture, ignition timing the engine’s air intake temperature; the engine block and radiator coolant temperatures; as well as the engine, transmission and differential oil temperatures. These measurements are conducted on the road, under normal driving conditions and are only recorded once the temperatures have stabilized at what would be considered normal operating temperature. Once the temperatures are stabilized, we record the data from 2,000 rpm to redline at wide- open throttle. Again, the data is collected with the transmission in the gear that is one to one, typically 5th, so that the input and output shafts are connected, reducing power losses and wear on the transmission. The data is recorded several times in order to obtain a solid average.
Sample Data Aquisition Display
Next, the times necessary for the stock car to achieve various speeds are recorded on the road and then loaded into the dyno program in order to simulate road conditions.
The air intake sensor is located under the hood of the E39 M5, absorbing heat produced by the engine. As you can see in Figure 1 below, on the road the sensor absorbs heat from the engine, artificially raising the reading the computer sees to 110° F. As soon as you accelerate at wide open throttle, the ram air coming into the engine flushes out the hot air and cools the sensor. By the time the engine reaches high rpm the temperature sensors are seeing 85°, very close the 80° F ambient temperature that was recorded during the test run.
Figure #1: E39 M5 Intake Air Temperature Measured on the Road
Referring now to the stock radiator and engine block temperature graph below (Figure 2), you can see that BMW’s engineers did an excellent job of maintaining consistent temperatures in both the radiator and block. As the engine revs friction is increased, as well as load and cylinder cycles. All three of these things produce more heat but as the engine revs, the car is moving through the air faster which rams more air through the radiator, exchanging additional heat and therefore maintaining a very stable temperature.
Figure #2: E39 M5 Engine and Radiator Temperatures Measured on the Road
As you will soon see, once we put the vehicle in the stagnant air in the dyno room, all of this changes. The first dyno run we will talk about (Figure 3A – purple line) represents what I would consider the worst dyno testing procedure I have ever seen. The vehicle was placed on the dyno with the hood closed and a small fan positioned in front of the grille, typical of the fans I have seen in most chassis-dyno facilities. The engine is warmed up to normal operating temperature by performing two passes and then the car is left to idle for 10 minutes.
Figure #3A: E39 M5 Dyno Graph
This combination of conditions resulted in the lowest recorded output on the graph, producing 335.7 hp. During the two warm up passes the engine, radiator, intake manifolds, and air intake sensor will “heat soak”, resulting in a reduction of power. Let’s take a moment to look at each graph separately and analyze what has happened.
When the M5 was on the dyno with the hood closed, and there was no ram air or air flowing under the car to evacuate heat from the engine compartment, the under-hood temperature rose significantly beyond what would be seen under normal driving conditions. As you can see in Figure 4 – pink line, the temperature reached 160° F from idling and even after the wide open throttle run was completed the temperature only dropped to 148° F! Comparing this to the road graph (Figure 4-blue line) shows a staggering increase in the temperature reading even though the outside air temperature has not changed at all. Normally when the computer sees a higher air temperature reading, it leans the air/fuel mixture and retards ignition timing in small amounts to compensate for the less dense air. This occurs in order to maintain a proper air/fuel mixture and prevent detonation. However, when the M5 computer sees a very hot value it goes into a portion of the program designed to save the engine from melt down. This mode dramatically richens the mixture and retards the timing, preventing engine damage in two ways: it causes the engine to produce less power, thereby producing less heat; and some of the heat is actually absorbed by the fuel, then carried out through the exhaust. In addition, the rich mixture and retarded timing ensures that the engine will not detonate under these conditions.
Figure #4: Intake Air Temperature Sensor Comparison
The radiator cannot exchange heat as well as it did on the road because there is not enough ram air flowing through it. Looking back at the road temperatures depicted in Figure 5 – blue line, we can see that the radiator stays between 175° F and 178° F. Now compare those readings to the radiator temperature from the dyno acceleration run in Figure 5 – violet line. With the standard dyno fan you can see that the temperature starts at 173° F and climbs to 210° F!
Figure #5: Radiator Temperature Comparison
Since the radiator cannot exchange heat on the dyno as efficiently as it did on the road, the engine block heats up significantly. Looking at the road temperature curve in Figure 6 – blue line, you can see that the engine block temperature stays between 191° F and 188° F. However the engine temperature measured during the dyno acceleration run with a standard type of fan starts at 192° F and ends at 203° F (Figure 6 – violet line). It should be noted that the coolant temperature gauge in the instrument cluster will hardly move in this case, even though the engine management system needs to correct for the increase, as 203° is not a high enough temperature to cause the vehicle to actually overheat.
Figure #6: Engine Block Temperature Comparison
The combination of the engine heating up and the air temperature sensor reading an artificially high value causes the engine management computer to go into the “engine savior mode”. The mixture is richened to an astounding 9.5 to 1 air fuel ratio (see Figure 7 – violet line); whereas the correct mixture measured on the road was 12.2, shown in Figure 7 – blue line. The ignition timing is retarded from a peak value on the road of 27° (see Figure 8 – blue line) down to 15° (Figure 8 – violet line). It is truly amazing how intelligent a modern BMW is. If you were to go back just 15 years, these same conditions would likely result in engine damage!
Figure #7: Fuel Mixture Comparison
Figure #8: Ignition Timing Comparison
We will begin to eliminate these variables one by one so that you can see which aspects are the result of the intake air temp sensor and what portion is attributable to the engine running too warm.
The next dyno run was performed with only one change: the hood was opened!
As you can see in Figure 3B, with this simple change the power increased by approx 35 hp to 370 hp.
Figure #3B: E39 M5 Dyno Graph.
The air intake sensor now absorbs a lot less heat from the engine (see Figure 9 – yellow line). With the hood open, the temperature reached 130° F, as compared to the 160° reading with the hood closed (see Figure 4). After the wide open run was completed, the temperature dropped to 120° F (Figure 9 – yellow line), compared to 148° F (Figure 9-blue line). However, this is still significantly warmer than the temperatures that were measured on the road.
Figure # 9: Intake Air Temperature Comparison- Hood open vs. closed
Looking at the radiator temperature during the dyno acceleration run, using the standard dyno fan and the hood open, you can see in Figure 10 – yellow line, that the temperature is able to cool off more between runs. This enables us to start the run with a temperature of 155° F, ending at 205° F! Once again, this is still substantially warmer than the temperatures measured on the road (Figure 10-blue line)
Figure #10: Radiator Temperature Comparison
Since the radiator still cannot exchange heat as efficiently as it did on the road, you can see in Figure 11 – yellow line, that the temperature starts at 182° F and ends at 197°, still warmer than the road test.(Figure 11- blue line)
Figure #11: Engine Block Temperature Comparison
While we have made some serious progress here, and by now it should be obvious that you should never dyno a car with the hood closed, we are still significantly short of duplicating the normal conditions the car would see on the road. The improvements we have made thus far have leaned the air/fuel mixture from 9.5 to 1 (Figure 7-violet line) to 11.2 to 1 (see Figure 12 – yellow line); however compared to the correct road mixture of 12.2 to 1 (Figure 12 – blue line), the air/fuel mixture is still too rich. Ignition timing has also improved from 15° (Figure 8-violet line) down to 22° (Figure 13 – yellow line). This is still substantially less than the peak road value of 27° (Figure 13 – blue line) that would occur during actual road conditions. I believe that this is how most “power chips” are made, essentially leaning the mixture and advancing the ignition timing back to normal values. This will result in a measured power increase, however this gain is not real because it is merely compensation for the dyno environment.
Figure #12: Fuel Mixture Comparison
Figure #13: Ignition Timing Comparison
The hood will remain open for all subsequent dyno runs. The next dyno run was performed with only one change: disconnection of the stock air temperature sensor and the installation of one at the air inlet. This is done in order to get the sensor to accurately reproduce the temperature of the air actually going into the engine. This will stabilize the engine tremendously and result in the computer making proper corrections for the conditions. As you can see in Figure 3C, the power increased by approx 10 hp, with the run producing 380 hp, with significantly less fall off at higher rpm.
Figure #3C: E39 M5 Dyno Graph
The air intake sensor is now rock steady at 81° F (see Figure 14 – light blue line). This matches the actual ambient temperature in the room at this time and is more stable than the temperatures measured on the road (Figure 14 – dark blue line). This should help to explain why we decided to move the intake air temperature sensor as part of our Cold Air Intake System.
Figure #14: Intake Air Temperature Comparison
The radiator and engine block temperatures remain the same as the previous run since we are still employing the small fan.
While we have made even more progress, we are still significantly short of the normal conditions the car would see on the road. The mixture has leaned out to an 11.7 to 1 air fuel ratio (see Figure 15 – light blue line) when compared to the 12.2 to 1 ratio measured on the road (Figure 15 – dark blue line). Ignition timing has been retarded from a peak value on the road of 27° (Figure 16 – dark blue line) down to 22° (Figure 16 – light blue line), but it is less stable. This is due to the cooler air intake sensor value causing the computer to lean out the fuel mixture and advance timing. Since the radiator still can’t exchange enough heat, the engine runs warmer than it normally would, causing the engine to detonate which in turn set off the knock sensor causing radical spikes (see Figure 16 – light blue line). The remaining richening and retarded ignition timing shown on these graphs are resulting from radiator and engine temperatures, not the engine intake air temperature reading since it has been stabilized.
Figure #15: Fuel Mixture Comparison
Figure #16: Ignition Timing Comparison
The last step in our attempt to reproduce actual road conditions is to employ a fan large enough to exchange enough heat that air temperatures at the end of the run exactly match the temperatures recorded on the road. The only change for this run was to replace the small fan with the Dinan® (Level 1) “Hurricane” fan.
As you can see in Figure 3D below (violet, blue and yellow lines), the power increased by approximately 30 hp to 411.4 hp, backed up by a 410 hp run. When the conditions are truly controlled you can usually produce runs with as little as 0.5 – 1.0% variance.
Figure #3D: E39 M5 Dyno Graphs
The air intake temperature sensor was stabilized on the last run, so now let’s look at the radiator and block temperature graphs. As you can see, the radiator temperature during the dyno acceleration run (Figure 17 – violet line), using the very powerful fan and the hood open, allowed the temperature to cool off more between runs. This enabled us to start the run with a temp of 110° F and end at the exact same value as we saw on the road, 175° F (Figure 17 – dark blue). Even with this huge fan, the largest I have seen on a chassis dyno, we still must start at artificially low numbers so as not to exceed the road value by the end of the run! In other words, even our huge 75-mph fan can’t duplicate the air flow the car would see on the road…but we are getting closer!
Figure #17: Radiator Temperature Comparison
We are able to match the engine temperature recorded on the road at the end of the dyno run as well (Figure 18 – violet line). Figure 18 shows the engine temperature during the dyno acceleration run with the large fan. You can see that the temperature starts at 182° F and ends at 188° F equal to the road values (Figure 18 – blue line).
Figure #18: Block graph
The air/fuel mixture and ignition timing now match the road numbers almost exactly (see Figures 19 and 20 ).
Figure #19: Fuel Mixture Comparison
Figure #20: Ignition Timing Comparison
The important thing here is to match the engine temperature as closely as possible. As we have been able to achieve just a 7° variance (see Figure 18) from the beginning to the end of the run, with an ending value that is the same as the road number, the mixture and ignition timing match the road value. Now we know that we have real-world horsepower number. Remember that we have not added or changed any parts on the car during the course of this testing, with exception of the temperature sensor.
Even still, we fall short of duplicating actual road conditions in one area: we can’t reproduce ram air to the intake system. We are currently developing a system that will produce enough ram air to get us even closer, but until it is completed we will still measure less power on the dyno than the vehicle will actually make on the road!
Engine Wear
Every engine will produce different power output, even if every variable is carefully controlled! This is due to production tolerances as well as maintenance and care during break-in. Most of the variances occur from cam timing errors and cylinder leak-down variances. Cylinder leak-down variances are the biggest variable. A desirable number for leak-down is less than 5%; however, it is very common to see numbers higher than that. It mostly depends on how the engine was broken in and, of course, maintained. The S2 M5 engine used in this test had an average leak-down just over 4%. The same engine had an average leak down of 3% one year ago. Comparing the previous dyno runs on this engine to more recent runs (see Figures 3E and 21), you can see a loss in power of approx 1%. This is the same car on the same dyno! However, due to normal wear, the power has dropped by 6 hp. You must realize that not every engine of the same type produces the same power. You also must realize that not every engine produces the same power throughout its life. BMW engines are very well manufactured with very consistent tolerances. Our dyno test show that almost all engines, without any defect, will be within a 5% window with the vast majority being within a 3% window.
Figure #21: Year old E39 M5 Dyno Graph
Figure #3E: Current E39 M5 Dyno Graph
Drivetrain Temperatures
In addition to variances from the engine itself and the dyno environment, more variables come into play as a result of varying drive train oil temperatures. The graph above (Figure 21) depicts the recorded rear wheel horsepower for the S2-M5. As you can see, as the drive train oils heat up during the four dyno runs, the recorded horsepower increases from 406 to 417; an eleven horsepower gain resulting from nothing more than increased drive train oil temperatures! Remember that the engine was already warmed up to normal operating temperature before the four dyno runs were conducted. The temperatures of the air intake sensor, engine block and radiator were strictly controlled during these runs so that the only variables were drive train oil temperatures. The run that indicates the least output at low rpm was the first one after the engine was restarted (See notation- figure 21). The reason for this is that when the engine is first started, the camshaft and ignition timing have a different program to assist in warming up the catalyst for emissions purposes. As soon as the aft O2 sensor determines that the catalizer is operating, it automatically reverts to the normal program. The colder the catalytic converter, the more runs will produce reduced output. You can see that the oils have reached normal operating temperature for the last two runs as the measured output is very similar. Thermal couples can be used to further improve this accuracy. Many Winston Cup teams are now using chassis dyno’s to reduce drive-train friction. Since improvements in this area are so small in order for this work to be valid they must strictly control the drive train temperatures.
Data from a Stock E46 M3
The M5 engine is extremely sensitive to temperature control when dyno testing. Any increase in power results in an increase in heat and a corresponding increase in sensitivity. Not every engine demonstrates the same levels or types of sensitivities; some have cooling system sensitivities while others are sensitive to ram air volume and fuel octane. Each engine must be tested to determine what conditions exist on the road that do not exist on the dyno and what must be done to correct for them. Since the M5 started the testing with 93-octane fuel, I wanted to provide an example of how octane affects horsepower. The E46 M3 is an excellent example of an engine that is sensitive to octane. It has a very high volumetric efficiency as well as a very high static compression ratio of 11.5 – 1. The engine being tested was a stock M3 engine. It was first warmed up and stabilized using the method described previously for the M5, running 91-octane fuel. As you can see in Figure 22, where the engine was warmed up and the previously discussed testing procedures applied, the stock M3 produced 280 hp (Figure 22- violet line). We then replaced the 91-octane fuel with 93 (available in most parts of the country). The M3′s computer was so quick to determine that the fuel had been improved that it only took four dyno runs for the timing to adapt to the increased octane and raise the power up to 291 hp (Figure 22 – light blue and yellow lines). A gain of 11 hp with just 2 points of octane. The M3 engine is equipped with a very good ram air system. While our large fan is pushing large volumes of air into the ram air duct, the volume and velocity of air seen by the car on the dyno is still less than the engine would see on the road. The power output drops off after 7350 rpm because we simply cannot duplicate the airflow the car would receive on the road in the dyno room, resulting in a loss of power. BMW claims peak power @ 7900 rpm. We have no doubt that if we could accurately reproduce the ram air that the M3 would receive on the road, our peak power would move up from 7350 rpm to 7900 rpm. One of our many engineering projects is to further enhance our ram air capabilities in the dyno room.
Figure #22: E46 M3 Dyno Runs
Stock E46 M3 Ignition Timing Adaptations
Figures #23 and #24 below are taken directly from the BMW factory diagnostic tool, demonstrating how ignition timing adapts to different fuel octane ratings. The same car is represented here, the only difference being the octane rating of the fuel. If you were to add 1° of ignition advance, the engine management system would detect it and retard the timing 1°. You can see that adding timing in the engine management software or “power chip” is futile because the computer will negate the change, as sufficient octane does not exist. However, you can see that adding higher octane fuel is like adding a “power chip” as the system adapts to the better fuel, making more power.
Figure #23: 91 Octane Knock Adaptations Diagnosis
Figure #24: 93 Octane Knock Adaptations Diagnosis
Intercooled Forced Induction Systems
Cars with intercooled forced induction systems (superchargers or turbo chargers) provide an even bigger challenge on the dyno. A separate fan must be employed in order to sufficiently cool the intercooler. Thermal couples must be installed in the inlet and outlet of the intercooler so that the temperature drop seen on the road can be measured. Once the temperature drop has been established, the fan speed must be adjusted until the same drop in temperature is maintained on the dyno as that was measured on the road. If the temperature drop cannot be achieved on the dyno, the error can be corrected for mathematically and the results will be very close.
Corrections
Once we have determined the specific baseline procedure for a vehicle, the car is allowed to sit until early the following morning when the temperature is as close as possible to 77° F, the SAE standard. There are formulas for temperature, humidity and barometric pressure corrections from the SAE; however the correction tables are not completely accurate for a digitally controlled car. This is due to the corrections the car’s computer is making based on the conditions previously discussed. We reduce this error further by performing all tests as close to 77° F as possible. An alternative would be to turn the corrections off in the software but this is a potentially dangerous approach. The most accurate results are obtained when the tests are performed in a climate-controlled dyno-room where temperature and humidity can be completely controlled for each test. By running our tests as close to 77° F as possible, we still must employ the SAE correction tables, but the amount of correction necessary is reduced and accuracy improved. Be aware that not all dynos correct for temperature, humidity and barometric pressure!
Conclusion
If you decide to test your car on a dyno, whether in stock or modified form, be advised that you will not see the same results as BMW or Dinan. Putting the time consuming and tedious procedures aside, any number of things can cause your measurements to be different from those published. Even in a best case scenario, assuming that there is no need to be concerned about calibrations because the performance software has been supplied, it still takes the better part of three days to go through the proper testing procedure and collect the necessary data.
If the engine is detonating or in the “savior mode” because of excessive temperatures, gains can not be measured. In fact if the car sits and heat soaks or cools for an excessive period of time between runs, enough variance can be created that the performance component enhanced car may show less power than the stock car, or even a very exaggerated gain.
By way of summary, following are some of the more significant factors that you should keep in mind when considering dyno testing in general terms, as well as what to look for in the facility itself.
1.) Each dyno will produce different results (even with the same brand of dyno).
2.) The octane rating of fuel varies in different parts of the country (you must have a controlled fuel supply).
3.) Cold weather increases the gains and hot weather decreases them, even with temperature corrections.
4.) Lack of oxygen from exhaust in a dyno room will cause a loss in power.
5.) Slipping tires on the rollers will reduce the measured gains.
6.) Inertia type dynos have a lighter load than the car will see on the road. This is especially true for cars with heavy drive trains because some of the power will get absorbed spinning the masses faster. The inertia correction programs employed in these types of dyno’s are not completely accurate.
7.) Fixed load dynos have a higher load than what the car sees on the road. This excessive load will result in a large mixture shift and the detonation sensor will be activated prematurely.
8.) No dyno can accurately simulate wind resistance, the ram air effect into the airbox or cooling of the intake tract under the hood.
9.) The size of the fan used during testing will change the power output.
10.) Oil temperatures will affect output due to changes in friction.
11.) The air intake temperature sensor will trigger adjustments to fuel mixture and ignition timing.
Dinan® is certainly not the only BMW tuner in the world that understands the variables and complexity of proper dyno testing and tuning. I have a great deal of respect for the handful of BMW tuners around the world that share our passion for accuracy. Unfortunately, this level of dyno testing sophistication appears to be the exception and simply won’t be found in common dyno facilities that rent their time. The purpose of this paper is to provide our customers with a deeper understanding of our test procedures and why they have been developed. We too are constantly learning more about the science and updating our equipment whenever significant improvements in the technology occur in an effort to provide our customers with the most valid data possible. In addition, Dinan also employs an engine dyno test cell, but we’ll save a discussion on that technology for another time.
The following users liked this post:
DJS (01-31-2016)
#2
01-30-2016, 02:29 PM
Thanks for taking the time to post that info. Honestly we prefer data in the following order, but due to time of year we'll take what we can get domestically speaking at this point:
1) Track (1/8, 1/4, 1/2, Standing Mile)
2) VBox (still uses real world load and airflow, and starting at 60mph mitigates wheelspin issues for 90% of cars tested)
3) Racing another known car on video (for instance one that you had previously beaten or lost too prior by X amount of car lengths). Yes this still ranks over dyno results for us because when dynoing you don't see real world airflow or load on the driveline, and for some reason DT loss has again become a debate (as far as how much actual loss there is).
4) Dyno for the reasons given above.
Thank You
VMax
1) Track (1/8, 1/4, 1/2, Standing Mile)
2) VBox (still uses real world load and airflow, and starting at 60mph mitigates wheelspin issues for 90% of cars tested)
3) Racing another known car on video (for instance one that you had previously beaten or lost too prior by X amount of car lengths). Yes this still ranks over dyno results for us because when dynoing you don't see real world airflow or load on the driveline, and for some reason DT loss has again become a debate (as far as how much actual loss there is).
4) Dyno for the reasons given above.
Thank You
VMax