What's that you say? You -- the consumer, the enthusiast -- want more comparisons between the 2010 Mazdaspeed 3 and 2010 Volkswagen GTI?
Well, then. Betcha you never saw this coming.
You don't need a dyno to know that the fiery red Mazda has more gumption under its perforated bonnet than does the VW. After all, a quick look at the spec sheets will tell you that. While you can get an idea of a car's potency from its manufacturer's rating of power and torque, the dyno tells us so much more. Besides, dynos are cool.
We rounded both fully broken-in examples from our long fleet and headed to MD Automotive's Dynojet chassis dyno in Westminster, CA.
As usual, these cars are running 91 octane because that's "premium" fuel in CA. See, that's funny, because I call it "crappy."
First up, the GTI. Its 2.0-liter direct-injected 2.0-liter four-pot is rated at 200 hp and 207 lb-ft at the flywheel. It's a smooth-revving, smooth-sounding smoothie with smoothness. On the dyno, it behaved in a similarly mannerly way, exhibiting minimal run-to-run variation.
Hey-oh, look at that. The GTI puts out as much power at the wheels what Volkswagen claims its makes at the flywheel, and a skosh more torque. Note that this is not some outlier dyno run -- it's representative of several pulls that were all repeatable within a few horsepower.
Ignore the noisy portions of the traces which are simply artifacts and not representative of the GTI's actual power production. In your mind's eye, strike an arc through the noise. This mental smoothing (there's that word again) exercise clips the peaks, one of which is the peak power. Call it 199 hp and 214 lb-ft.
The GTI's torque rolls off gently starting at 5200 rpm, but the dyno trace tells us that it's still worthwhile to rev the GTI out to the rev limiter. This is because best acceleration relies on maximizing the time spent at or around peak power as you go through the gears.
The Mazdaspeed 3? Short-shift it for best results. To wit:
Rated at 263 horsepower and 280 lb-ft of torque at the flywheel, our Mazdaspeed 3 long-termer put 243 horsepower and 265 lb-ft to the ground. Er, rollers.
The Mazda's 2.3-liter inline four is all midrange, too. The Mazda absolutely falls on its face at 5,500 rpm and there is little sense in revving it past 5,800 rpm. It's done.
As a result, the ECU ramps boost down and closes the throttle at high engine revs, effectively packing up its marbles and heading home for a nap. Volkswagen is almost certainly using the same approach but it's far more subtle and doesn't artificially truncate the rev range. Speaking of which, the Mazda's tach is wildly optimistic, indicating a 7,200-ish rev limit when in reality it's 6,600.
What's happening -- more accurately, what's NOT happening -- past this point is Mazda's way of reducing dangerous horsepower. Okay, it's actually a measure to prevent turbo speeds from getting too high. The company fitted a wee turbo in an effort to give acceptable boost response and a low boost threshold but the tradeoff is excessive backpressure and spiraling turbo speed at high rpm.
Still, the Mazda kicked out healthy numbers and is clearly stouter than the Volkswagen but the grunt gap between them is not as quite as vast as the manufacturer's specifications suggest.
It'd be premature to claim that VW is intentionally under-rating its mill. That's because there's no universally accepted procedure regarding how manufacturers treat the intercooler on the engine dyno when they're developing the performance rating of a boosted engine.
A conservative company would assume some worst-case heatsoak situation for the intercooler. This would result in lower output on the spec sheet because intercooler discharge temperatures in the real world would often/always be much more favorable for making power.
--Jason Kavanagh, Engineering Editor
vt8919 says:
06:24 PM, 10/ 5/10
Pretty cool read. Got a question, though: Noticed the ABS light was on in the Mazda. Do you have to disable it because it confuses the sensors when only two wheels are moving?
zoomzoom22 says:
06:43 PM, 10/ 5/10
Is that an NSX I see?
clarkma5 says:
06:43 PM, 10/ 5/10
VW's various iterations of their modern turbo 4 (starting with the AWD 150 HP 1.8T, through the AWP 180 HP 1.8T, then onto the EA111 2.0T and now the EA888 2.0T) have always been a bit underrated, typically putting out exactly what you see on this dyno...wheel power and torque matching what the spec sheet claims at the crank.
Back in the 1.8T days, this was to give some headroom to the pricier 6-cylinder options in the price range (the VR6 in the A-chassis cars (first the 12v then the 24v) and the V6 in the B-chassis cars (first the 12v then the 30v)) and nowadays I believe it's to give some headroom to the Audis (211/258 is the A4 2.0T's rating, remember). It also has a lot to do with the acceptance of the turbo motors I think, why they've always been praised as "feeling like 6s"...they feel like more than their spec sheets indicate they should because they, in fact, are.
Not to dismiss the Mazda's motor in any way, it's legitimately putting out its claimed figures if not a little bit more, and it's a grunty motor, though the powerband is unfortunate. They supposedly flattened the torque curve out since the first gen as well with this one, makes me wonder how bad the previous MS3s were in this regard. Also, let us not forget that it is possible that the MS3 engine will perform better on the open road with real airflow into its airbox and through its intercooler than the GTI will.
And one last thing about these motors: my experience with the VW motors, great though they are, is that VW has always chintzed on the intercoolers (225 HP 1.8T from the 1st-gen TT notwithstanding) and the end result is heat soak in hard driving and/or high ambient temperature which really blunts performance, both in terms of absolute power output as well as in lagging the response. So it's not all roses in Germany's corner.
saxdogg says:
06:52 PM, 10/ 5/10
I watch Powerblock on Spike and I think it was the show "Horsepower" that just did a feature on a power programmer for the MS3. They got like 70 HP out of the tuner at the higher RPMs. Don't know what it would do long-term to the car, but it was interesting.
huyracing says:
07:08 PM, 10/ 5/10
these modern GTI's can get by fine on a track with a larger k04 turbo on the stock intercooler, where mazda's trickery is because the intercooler isn't big enough... and any car can suffer from heat soak, turbo cars generally suffer more because they run hotter. i have spent a lifetime driving cars with forced induction and it is nice and far faster than without, but that is the downside when it gets hot out.
adavis2493 says:
07:37 PM, 10/ 5/10
I'd take that Porsche on the lift in a second.
f1ndler says:
07:58 PM, 10/ 5/10
Thank you for doing this test IL. Although, I knew that VW and as well as BMW and Mercs don't "lie" to their customers about real numbers as Japanese car manufactures, you guys proved that it is not a myth at least regarding VWs.
Those who objected me before can bite their elbows now:)
liquoredonlife says:
09:08 PM, 10/ 5/10
Where's the video that the article mentions?
hb22 says:
09:54 PM, 10/ 5/10
Too bad VW changed its packaging :(
santiagofdz says:
10:58 PM, 10/ 5/10
@clarkma5
I'd expect the Mazda to do better on the open road with the scoop gulping down air the way it was designed to. The way the graph looks, it's very similar to the way the 1st gen MS3 I drive feels on the good ole' butt dyno. Power starts to come on at 2.5 K RPMs, hits you at around 3k and stays on until somewhere around 5.7K RPMs. Supposedly it was just a minor tune and the air scoop from the 1st gen to the 2nd gen, so that graph looks pretty consistent with that claim.
lt1boy says:
12:10 AM, 10/ 6/10
Gotta love the MS3. 243 whp equates to 285 hp at the crank. Both manufacturers are underrating their engines, although the GTI is underrated by a bit more.
However, you really can't believe any dyno tests from IL, since their long-term SRT-10 dynoed more rwhp than when they dynoed a ZR1.
qrmu says:
12:50 AM, 10/ 6/10
These results look mostly the same as other GTI and MS3 dyno run graphs I've seen. Usually the 2.0 TSI torque plateau starts at 1700 rpm though, not as late as 2500 rpm.
Makes you wonder why MS3 real redline isn't at 6000 rpm. With that tach it would still show 6500 rpm.
breif says:
04:40 AM, 10/ 6/10
QRMU, would you buy a sporty car with a 6k rpm tach? It's all marketing.
As an MS3 owner, I agree, the powerband could be extended a bit. These cars LOVE intakes, though. The power increase is noticeable, and some people claim a 25hp increase on a dyno.
dagmar3 says:
05:30 AM, 10/ 6/10
Dear Inside Line:
If you folks are ever sitting around and wondering if you should post more dyno runs and oil spectrum analysis stories or more posts about the latest press release on the Chevy Spark or some other dull econobox, please - fill your bandwidth with dyno runs and tech stories.
Thanks.
dagmar
cdyer says:
05:34 AM, 10/ 6/10
Too bad VW has decided to go bland, even though their 20% growth is because of fun/unique cars like this (wsj.com yesterday). VW finally takes the US market seriously, and they go the wrong direction...as long as they save the GTI I'm coo' wid it.
eldaino2 says:
06:08 AM, 10/ 6/10
I remember getting my head torn off for suggesting that the gti had the better,smoother powerband and that the
Mazda literally fell on its face after 5500rpms or so.
Thanks Edmunds, for this 'revelation' and using my specific wording :)
vtirocz says:
06:17 AM, 10/ 6/10
Good article, but couple things to note:
1. Dynojet chassis dynos are notorious for giving high readings when compared to something like a Mustang chassis dyno. From articles I've read, I'd say the Dynojet is a bit optomistic, however, this test can still be used for comparison between the two cars which is what you were doing.
2. There is an SAE spec/procedure regarding intercooler effectiveness for engine dyno testing. If the mfg claims are "SAE certified", then they should be following that same procedure. Not sure if they are forced to follow it though. If I find the SAE spec #, I'll post it.
breif says:
06:31 AM, 10/ 6/10
vtirocz,
Dynojets are VERY optimistic compared to Mustangs. I was just at an even where significantly modified Speed3's were not matching those stock numbers on a Mustang.
Unfortunately, since it does read high and feeds peoples egos, the Dynojet has become the de facto standard.
lt1boy says:
06:59 AM, 10/ 6/10
Next up: a dyno test of the new Chevrolet Cruze. Sponsored by Chevrolet. Get used to more.
kosmo69 says:
07:22 AM, 10/ 6/10
i'd still take the GTI.
Too bad both cars are FWD. I hate that.
jhatmaker24 says:
07:26 AM, 10/ 6/10
I concure, adavis2493.....
norsairius says:
08:14 AM, 10/ 6/10
@ f1ndler:
You're right, Japanese car makers do lie to their customers about power numbers. Fuji Heavy Industries/Subaru rates my '09 WRX at 265 hp/244lb-ft of torque at the crank.
The car has been seen putting out around 230-240 hp/240 lb-ft of torque at the wheels. Taking into account drivetrain losses (with AWD I'm guessing 15% at best, more likely around 20%), it seems the engine is making more at the crank than quoted.
Those liars! (sarcasm) However, there are probably other examples you can cite that I've not yet heard of or read about.
That said, I think there are too many factors to take into consideration to believe that any dyno results would be conclusive. I believe that chassis dynamometers are meant to be precision instruments for tuning so that you can lay down a good baseline and then make more runs as you continue to make tuning adjustments to see power gains, etc. So they may not be accurate, but if they're precise, that's what matters.
It doesn't hurt to see what your car can do on a dyno though, as I'm curious to see myself what kind of numbers my car puts down even though other WRX owners have done it already. Every dyno will read differently for every car. Results may be relatively close, but they'll be different.
church123 says:
08:15 AM, 10/ 6/10
The Dynojet does not read high folks. The Mustang dyno reads artificially low. There is simply no physical way a car is losing 25-30% of its output through the driveline as some Mustang dyno readings would have you believe. If that was the case every car would need a transmission oil cooler to prevent cooking the tranny fluid in day to day driving.
Southwest Research Institute (SwRI) did a study about 7-8 years ago (available from the SAE, called "Listening to the Voice of the Consumer - Chassis Dynos) where they looked at the Dynojet, the _old_ Mustang inertial setup and the Superflow inertia dyno. They also measured actual losses through the transmission on a transmission dyno with the tranny out of the car. Measured differential losses, etc. And of course measured the output of the engine on an engine dyno. Turns out the sum of all the losses was 14-16%. And on the Dynojet, guess what? The losses were 15%. On the Superflow they were 14% and the Mustang 16%.
The newer SAE spec means that current factory numbers are rated somewhat lower, mainly because of test conditions, not actual engine output, but on cars since 2007 or so the difference will be smaller.
The problem with the Mustang dyno is that in the eddy current setup the operator has control over several load variables, and this can cause the readings to vary. Unfortunately, most operators do not change these load variables to match each car properly. But again, it simply isn't possible that cars are losing so much hp through their drivetrains. If you want another way to confirm it, simply look at calculated wheel torque on a Mustang and calculate predicted acceleration in 3rd gear using that number. I guarantee you your car will pull at least a tenth of a g more than what the Mustang predicts, indicating that the measured number is too low.
BTW, you have to test the MS3 with the hood closed and a strong fan blowing into the hood scoop. Otherwise the intercooler airflow gets all messed up. At a minimum someone needs to hang a blower right on top of the intercooler with the hood up.
church123 says:
08:19 AM, 10/ 6/10
Oh, and f1ndler, the Japanese aren't lying about their numbers. If anything, VW is lying about theirs, by the strict definition of lying. So you're still wrong, but you're probably used to it, eh?
iskch says:
09:25 AM, 10/ 6/10
Good numbers for the GTI. True, as a MS6 owner after 5500 RPM is nap time for the Mazda engine and they don't like 91 gas (In okie land is like like CA - sucky 91 gas). Those Mazda Turbo engines love 93 octane gas.
huyracing says:
09:30 AM, 10/ 6/10
VW's lying is because of the R32. The R32 with the VR6 only made a tad more than the GTI at the time. The FSI motor did not make as much as the current TSI motor, but it was still close to 200whp. so they lied about both cars to make it seem like a bigger gap in performance.
ms3omglol says:
09:45 AM, 10/ 6/10
The MS3's "puke and die" powerband is helped tremedously with ECU manipulation. The hypertech power programmer for my 2008.5 MS3 fixed this problem. I'm kind of surprised that Hypertech nor Cobb has anything out yet, seeing as this powertrain is more or less identical to the previous model.
http://www.hypertech.com/PerformanceGraphs/mazda/62003_2006_2009_Mazdaspeed3_6_2_3L_Turbo_CAI.pdf
f1ndler says:
10:31 AM, 10/ 6/10
to church123:
Lol, you're funny:) How is VW "lying"? Even with the graph in front of your eyes you still can't let your old believe go. Learn to lose buddy, it's one of the greatest virtues.
mrb5091 says:
10:38 AM, 10/ 6/10
f1ndler,
I'm not church123 but I agree with his point, VW is lying by underrating the output of their engines, while Mazda is closer to the truth here.
The standard for measurement is power at the crank, no one claims otherwise as far as I know.
f1ndler says:
10:56 AM, 10/ 6/10
Yes, you are absolutely right about engine output, but what is the point of telling the customer the power of the engine if 15-20% is not where you need it to be - at the wheels?
dadasracecar says:
11:04 AM, 10/ 6/10
The biggest issue with the mazda is the throttle plate closing. THis was the case on the first gen cars and is still present. Custom Performance Engineering, www.cp-e.com, were the first to recognized this and they reflashed the ecu on the first gen cars to keep the throttle plate open. That coupled with the addition of up to 8 degrees of spark advance at the top end resulted in smoothing of that power curve. The stock turbo is basically maxxed from the factory so peak power doesn't increase but just holding existing power out to redline results in gains of 100+ hp at 6500 rpm or so - see graph,peaks at 243, falls to 110 at redline...
One issue that creeps up however is that the mechanical cam driven high pressure fuel pump has difficulting keeping up with the fuel demand of the car with the open throttle. I would speculate that the saving the turbo is a seondary aspect of the throttle closing. The primary issue resolved with closing the throttle is that the fuel pump can deliver adequate fuel. I've personally seen fuel pressure drop from 1800 psi to below 1000 at 16 psi with an open throttle plate. The car goes lean and may cut ignition. Replacing this pump with an aftermarket version from CPE or APR solves this issue.
For the first gen cars a fuel pump and the throttle plate flash make the biggest gains. After that it's bolt ons to increase performance.
mrb5091 says:
11:11 AM, 10/ 6/10
f1ndler,
I agree with you that power should be measured at the wheels - you're buying more than just the motor after all.
At this point, though, I doubt it will change. Not only is the standard of using brake horsepower as the metric of choice common and well accepted, but I doubt manufacturers would want to lower the ratings of their vehicles.
The average car buyer wouldn't understand why one year's model had 240 horsepower but the next year's is under 200.
racingrealtor says:
01:10 PM, 10/ 6/10
I've never had the opportunity to drive the MS3 but I've driven the MS6 and it certainly didn't feel as torquey as the graph says it should have. Maybe the AWD system ate it up?
johnmval says:
01:58 PM, 10/ 6/10
Great little read guys. It's always interesting for me to see what an engine is "really" producing in terms of horsepower and torque. Of even more interest is when I actually own one of the vehicles being tested...so thanks again.
santiagofdz says:
03:42 PM, 10/ 6/10
@racingrealtor.
And I've never driven the MS6 nor a turbo equipped CX7 for that matter, but in the MS3 it is plainly evident when the engine is in the power band and it does feel kind of violent at first, but like with all things you get used to it. In my case though I was upgrading from a car with 110HP, so it was always going to be a big of a deal to suddenly get more than twice the power.
racingrealtor says:
04:01 PM, 10/ 6/10
@ santiagofdz
True! I went from a Sentra SE-R (165 HP) to a Hemi Magnum (340 HP with ram air, CAI, exhaust and tuner).
It felt as if I upgraded to a rocketship!
A few months later it doesn't feel as fast....and I wish I had the funds for an SRT version.
joefrompa says:
09:45 AM, 10/ 7/10
A redline and tachometer should accurately reflect the physical limitations of reliable operation of rotating components at a given engine speed. These engines/tachometers do that, AFAIK.
It's important, for example, for Mazda to provide a tach up to 7200 rpms. With some modifcations, that engine can easily produce power to 6500-6600 rpms....and the pistons/rods/crank can handle it. So why not allow the stock tach to still be of good use?
Joe
jkavanagh says:
10:20 AM, 10/ 7/10
(catching up, as I've been away since this went live... )
Yup, I placed a small, dedicated fan on the MS3's intercooler during its dyno pulls to replicate the airflow that it would receive on the road. The GTI's receives airflow from the main fan, so there's equal footing here.
Also, regarding Mustangs vs Dynojet et al. Forget the actual numbers and look at the difference between runs on the same dyno. You should see a similar difference between these two cars on any given dyno, even if the absolute numbers vary widely.
@vtirocz, it's true that SAE J1349 includes some guidelines on intercooler setup, but they're loose enough as to be a source of considerable variation.
For example, manufacturers can elect to use a straight 60 degrees C (140 F) intercooler outlet temp OR they can use an instrumented acceleration test of a full vehicle to determine transient intercooler performance (not heatsoaked, plus gear and vehicle speed are not defined!). The former is cheaper, easier and conservative with respect to real-world use. Were I a manufacturer, I'd go for the 60 degree thing too.
jkavanagh says:
11:59 AM, 10/ 7/10
@joefrompa, the tach was reading 7,200 at the same moment the dyno's speed pickup (which reads rpm from the engine directly) read 6,600.
Most manufacturers ensure their tachs err in a similar regard to this, but I've not seen a tach as far off of reality as this Mazda's.
rlyon says:
02:18 PM, 10/13/10
Yea, but you can chip the GTI and get so much more from that engine :)
zackd44 says:
08:20 AM, 10/15/10
In MS3's do you shift 2nd early to hit the power band or hold it to get 60 in 2nd gear? It would take longer if u had to shift, i would imagine, into 3rd but power falls off at high end so? anyone??