Yesterday, 03:16 AM
A recent article about a Mustang dyno run for the Alfa Romeo Giulia Quadrifoglio exposed a large flaw in traditional dyno thinking. On this website most readers know that you can not just grab a graph from one type of dyno and compare it to another dyno and expect the numbers to be uniform.
Unfortunately, the average person does not understand this which is what led to someone putting out an article stating the Giulia Quadrifoglio is not making its stated output which sets understanding of dyno runs back instead of helping educate.
Let's start with the idea that cars suffer from 15% drivetrain losses. Where did this idea start? Why is it 15%? It originated with the Dynojet and became a rule of thumb on automotive forums for manual transmission cars. The percentage changes based on if the car is automatic (20%) or even all wheel drive and automatic (25%).
The problem here is that while the conversion factor works well to extrapolate crank horsepower from peak Dynojet figures at the wheels that is not the way drivetrain losses work. First of all, losses are not static. There is no one set figure for losses.
What does this mean? It means that your losses in 5th gear at 8000 rpm will be different from 2nd gear at 2000 rpm. The loss figure will change based on rpm, load, and even what accessories are running. Stating all manual cars suffer from a single set figure is flawed for this reason.
Not to mention that figure only works on the Dynojet which is inertia based. BMW M cars used to follow the number exactly back when they were naturally aspirated:
E46 M3 S54
Displacement: 3.2 liter
Torque: 269 lb-ft
Wheel Horsepower: 280
Wheel Torque: 235
E92 M4 S65
Displacement: 4.0 liter
Torque: 300 lb-ft
Wheel Horsepower: 350
Wheel Torque: 254 lb-ft
85% of 414 horsepower is 351.9. Notice the E92 M3 is almost exactly at 15%. DCT models would actually dyno slightly less than manual examples due to the wet clutch dual clutch having higher losses.
For the E46 M3 85% of 333 horsepower is 283.05. Again, the E46 M3 is almost perfectly at 15% losses on the Dynojet. It's almost scary how consistently close BMW was to this figure.
Now, in comes the turbo era. Let's see what happens when we get to the F80 M3.
F80 M3 S55
Torque: 406 lb-ft
Wheel horsepower: 427
Wheel torque: 429
How is the S55 engine producing more horsepower at the wheels than it is rated at the crank by BMW? Does the F80 M3 have a magical transmission with negative losses that somehow adds power?
Of course not. The DCT transmission in the F80 M3 works the same way as the DCT transmission in the E92 M3. What happened was the 15% Dynojet loss rule went out the window when the turbo power war started. BMW is able to claim a lower number to not push an output war on paper yet delivers far more where it matters, to the tires.
The drivetrain losses are still about the same for the transmission and rear wheel drive layout as the previous generation despite it no longer lining up with the crank horsepower figure which is practically meaningless.
How does this all look on a Mustang load bearing or eddy current variant? Well, very different. While the Dynojet is spinning a drum of a set weight and calculating the result based on how quickly that weight is spun the Mustang dyno creates load.
Why is this important? Because it simulates real world conditions which is great for tuning. A tuner can provide all kinds of different scenarios. The spool on the Mustang for a turbo motor will look different than on a Dynojet. Often time real world spool is much quicker than what is shown on a Dynojet graph due to load which generates exhaust gases more quickly for the turbochargers.
Note: Dynojet operators can optionally eddy current load control as well which muddies comparisons further.
The other day BimmerBoost posted an article on VF-Engineering's F87 M2 ECU flash software and provided Dynojet runs. Let's take a look at the same M2 running the same Stage I software on a Mustang:
Now the Dynojet:
269 rear wheel horsepower for the Mustang and 331 rear wheel horsepower for the Dynojet on the baseline figure. That is a difference of 62 horsepower at the wheels or 18.7 %.
If you take that 18.7% difference and apply it to the tuned run on the Mustang which shows 287 to the wheels you get 354 horsepower. Almost exactly what the Dynojet tuned run shows.
This is not an exact science as you can see. However, if you were to dyno cars all day on the Dynojet and on this Mustang you would see the runs follow very closely to the ~18% difference between the two.
So is that it? Just convert Mustang runs by 18.7% to get a Dynojet number? No. Things actually get very muddied due to the games tuners play.
Here is an example. The new Porsche 991.2 Carrera S on a Dynojet:
380 horsepower to the wheels. Now here is a 991.2 Carrera S on a Mustang from AWE-Tuning:
Why does the Mustang dyno show more power now? Because it is being corrected to read higher. People do not like low numbers and high numbers are better for marketing. BoostAddict asked AWE-Tuning what their correction factor is but they would not share it with us. You can see it is reading VERY high and higher than a Dynojet though.
In this instance it is better to focus on the delta which you will often hear tuners say. That means the difference between the two runs as comparing the peak figures to other runs on other dynos is pointless due to the unknown correction factor.
Good for marketing, bad for those who want relevant figures to compare with. It can also be very bad for consumers. Why is that you ask? Because some people will dyno without a correction factor and then change the correction factor to show larger gains.
A good example of someone who plays with the numbers is Vivid Racing: http://www.bimmerboost.com/content.php?4252-Inflated-numbers-Vivid-Racing-claims-407-wheel-horsepower-for-exhaust-and-VR-ECU-tuned-E9X-M3-S65-on-a-Mustang-dyno
They claim 407 wheel horsepower for an exhaust and canned tune on a Mustang dyno for the E92 M3. Anyone who has an E92 M3 (like me) and has had these modifications and dyno'd the car knows this is impossible. It looks good for someone who doesn't know that though and wants to buy parts thinking they make a ton of power, right?
Always keep in mind when you see figures from a company or tuner you should question them. Independent runs are the best to go by as the person usually isn't trying to sell you anything but just share their results. Also remember the Mustang can be configured like most dynamometers to read whatever the operator wants it to read.
If vendors and tuners did not play with dyno runs for sales we would have a great resource to work with for comparisons and so forth. You can not trust all tuners to do so. Fortunately, VF-Engineering has no dog in the fight and provided us with a variety of runs showing the difference between their Mustang and their Dynojet.
Yes, they have both. That is the proper way to do it. One dyno for tuning and one for marketing. Rather than correcting figures with whatever percentage generates the most sales they show the uncorrected runs on both machines so people can learn from and reference the pulls.
BMW 335i (N55) Stage TWO
An 18.2% difference between the peak horsepower on the tuned figures. There will be variance from run to run, day to day, fuel to fuel, gear to gear, etc.
Ultimately, remember, while dyno runs can tell you a lot about a motor or modifications they can also mislead people. Even worse, people can read them incorrectly and spread misinformation or manipulate them to sell parts.
A dyno is just one, albeit important, piece of the performance puzzle.