2008 Mitsubishi Evo GSR: Baseline Dyno Run At Dynamic Autosports

100hp difference in readings between dynos? Wow, that's incredible

Yeah, but these aren't simply two different models of the same dyno. They're completely different. Mustang vs Dynapack vs Dynojet will all give you different numbers.

So which one is correct? HP is a set unit of measurement.

Unless they are measureing different things, they should be a lot closer then that.

Otherwise HP numbers are meaningless.

Do all car companies test using the same method and dyno when they give their offical numbers?

^^ Companies tend to get the numbers at the crank, not the wheels. They tend to be a lot more accurate that way.

there are different readings for horsepower. the car companies use SAE approved methods to figure the horsepower. SAE tests the power at the crankshaft, but doesn't account for the power loss due to the transmission.

this dyno is testing the effective horsepower (ehp), or the amount of horsepower that the car is able to convert into forward motion.

there are different types of ways to figure horsepower, mechanical horsepower, metric horsepower, or find the torque and follow some equation based on torque, RPMs, and kilowatts.

also, the term horsepower is somewhat misleading. a horse is able to produce near 15 hp for a little bit, but over a long period of time, is able to produce a little less than 1 hp.

Honestly the Mustang run seems more inline with the acceleration numbers this thing puts down.

Disclaimer - I own a Dynapack and an EVO X. I also believe, like Jason, that dynos are best used for internal comparisons. What matters most is consistency.

First of all, Dynapacks remove the wheels/tires, so a lot of the losses present on a roller dyno (of any type) go away at that point. So you expect results to be higher for sure. Some people don't like that. As a tuner, I love it because we get to eliminate variables. That said, this result does seem really low for several reasons:

1) 95% of the differences between a roller and a Dynapack should be accounted for by the inertia of the wheel/tire package and the rolling resistance of the same. That usually amounts to a 7-9% difference. Rolling resistance/tire drag at speeds under 100-120 mph usually accounts for about 6-10 hp/single wheel depending on alignment (a bigger factor than you might imagine), compound, tire pressure and tire/wheel size. These numbers come from empirical testing and pretty good simulation models used by the OEMs and tire manufacturers (IOW, I didn't derive them myself).

2) Edmund's test of the 09 GTR at RRE and Harman using a Dynapack and a Dyno Dynamics roller dyno a difference of 46 hp or about 10%. This is not unreasonable in my experience.

3) One of the benefits of a roller dyno is that you should be able to closely equate acceleration with the measured torque curve (you can do relative accel comparisons on a Dynapack, but absolutes are tougher because you don't know the wheel/tire effects). And fortunately, Edmund's has already provided us with good acceleration data on the EVO X in its current state.

If you look at the last day of acceleration testing, the X ran 45-60 mph in 2nd gear in 1.3 seconds. That is a change in velocity of 22 ft/sec, or an average acceleration rate of 16.92 ft/sec^2. This is equivalent to 0.529 g of average acceleration over that speed range. We also know that over that speed range in 2nd gear the EVO goes from approximately 5500 rpm to 7500 rpm. Finally, at those speeds, aero drag is minimal, so we should have a good measurement baseline.

Now, look at the torque graph from the Mustang. It is pretty much a constant slope line from 5500-7500 rpm starting at 255 lbs-ft to 155 lbs-ft, or an average of 205 lbs-ft. We also know that in 2nd gear we have a combined gear ratio (2nd x final drive) of 9.1455. This gives us a total axle torque of 1875 lbs-ft. Divided by the rolling radius of the tires, we get 1757 lbs of thrust. Now, comes the only variable we don't know, actual weight of the vehicle. It seems that with driver and equipment, 3700 lbs is within reason? That would give us an average acceleration rate in 2nd gear from 45-60 mph of 0.475 g. This is over 10% lower than what was predicted by the Mustang results. Even accounting for visual estimation error in the torque curve, or a 100 lbs lighter weight, the result is still low, and I don't believe Edmunds corrects their acceleration times which would likely increase the differential as well (since it looks like the test was done in Fontana).

Based upon the acceleration data, I would have expected a roller dyno to produce results 10-12% higher than what the Mustang did, or at least 300 whp and 310-315 lbs-ft of torque.

BTW, if you run the relative numbers on the before and after acceleration results on the cam testing, you'll see that over the same 45-60 mph interval acceleration improved by about 12% and on RRE's dyno average torque over the same range improved by about 10-11%. So the relative gains make sense.

BTW, one other comment. Without knowing the acceleration rate on the Mustang dyno test (or the Dynapack test) it is impossible to make a completely informed analysis relative to real world testing. Grain of salt and all that.

church123,

That has to be one of the most thoughtful comments I've ever seen on Edmunds, and possibly any blog. Three cheers for a cohesive, educational, and interesting comment!

Thank you. I've been dealing with dynos and race cars for quite awhile now and these sorts of discussions repeatedly come up. So this analysis is just a variation on past work.

Nice to have you here Shawn.

church123,

Very interesting and in-depth read. Thanks!

Thanks Jason - wish I had all those cool parts on my X. It definitely feels a little short on power from the factory :). BTW, rereading my post a day later, it should have said:

"That would give us a predicted average acceleration rate in 2nd gear from 45-60 mph of 0.475 g. This is over 10% lower than what was actually measured."

Excellent input, church123!

Roller/wheel dynos FTW! (Yeah. I'm not a fan of hub dynos. Heh.)

Well, hub dynos are superior for getting accurate tuning done. On a roller there are many, many more variables with strap tension (ever seen a roller dyno operator put a strain gauge on his tie down straps to ensure consistency?), tire pressure, alignment, tire temperature, etc. Good operators can and do work around them, and on same day tuning it becomes less of an issue, but nonetheless, the fewer variables, the better.

Think of it this way, if you had an unlimited budget and resources, what sort of dyno would you tune on? Answer, an engine dyno, preferably something like a Heninan-Freud F1 dyno with the ability to electrically drive the engine under decel conditions - price tag $1 million+. That's what OEMs and high end race teams do.

But, these days, the idea of tuning a modern car on any sort of an engine dyno isn't too palatable. Trying to remove the engine, get the engine management set up, etc. is a pain. Sure, if you've got an older carbureted V8, no problem, but not a modern fuel injected computer controlled engine unless you've got $$$. So, you use a chassis dyno. The closest thing you can get in a chassis dyno to an engine dyno type environment is a hub dyno. That's in theory of course. While a hub dyno eliminates variables, there are only 2 manufacturers out there, so they don't tend to do as much with cool software functionality and stuff like Mustang or Dyno Dynamics. But as a pure tool for getting ultra consistent and accurate power measurements, they rock.

Just wanted to add, there is nothing wrong with roller dynos. I've tuned on just about every model of roller dyno myself and the tuner is more important than the dyno. But in terms of ultimate accuracy and repeatability...well, you know :)