We get to drive a lot of cars in this business. That’s both a blessing and a curse. It exposes us to new and different ways of getting people and things from Point A to Point B, but it also introduces us to new objects of lust and desire.
Such was the case when we procured a Mazdaspeed3 from the press fleet for a comparison test a couple of years ago. As fate would have it, we fell in love—the kind of unintentional, unplanned love that caused our eyes to glaze over and sent our heart fluttering.
Actually, at one point we believe we slipped the Speed3 a note: “Do you like me? Check ‘Yes’ or ‘No.’” The car never replied, but we knew the answer.
You never know when true love will strike, and that was certainly the case here. Unfortunately, we were so blinded by infatuation that we forgot to look beyond the honeymoon phase of the relationship. While kids were obviously out of the question, we found ourselves with a new project car and no formal project car plan.
We were in a tough spot. See, usually when we begin a project, the car has some clear needs that have to be addressed. Even newer vehicles have rough edges that need polishing, or polished edges that just beg for a mirror finish. With the Speed3, however, the combination of its overall goodness and our love blindness made changing it a difficult proposition, indeed.
As the days wore on, the luster faded. Only slightly, though—we still loved the car just the way it was. However, we did discover some things that could use some fine tuning. The most glaring issue came in the power department.
“Wait a minute!” we can hear you exclaiming, even though you’re there and we’re here. “I thought the Mazdaspeed3 had insane, punch-you-in-the-throat, chirp-the-tires-on-upshifts kind of power?” Well, it does, sort of.
While Mazda rates the turbocharged and intercooled 2.3-liter MZR engine at 263 horsepower—which is probably on the conservative side—the power curve falls off noticeably at around 5800 rpm. Now, the car has ample power and torque down low, where a short shift can keep it in the fat part of the powerband. Still, it seemed to us like a case of untapped potential.
We were also becoming disenchanted with the shifter. It might have suited another car just fine, but the shifter’s vague notchiness was out of character with the slickness of the rest of the Speed3’s controls.
Our mission had become a bit clearer: We wanted to address a few of these minor shortcomings without negatively affecting the overall goodness of the car that made us fall in love in the first place. Think of it as charm school rather than reform school. We’d start with the most glaring issues: the power drop-off and the shifter.
Push My Buttons
Say what you will about technology, but long ago we accepted the fact that computers can make tuning an engine easy. What was once the black art of carburetor and ignition tuning has been supplanted by the black art of fuel injection and engine management tuning. One is performed with a screwdriver and wrench while the other is done with a laptop and USB cable, but the goal is the same: Extract the maximum amount of performance from a given system within a certain set of parameters.
Both methods adjust similar things like fuel mixture, timing and turbo boost, but the clear advantage of the newer digital methods are obvious once the tuning part is done. First, adjustments made in the digital world stay adjusted. Ones and zeroes don’t tend to slip or come loose the way mixture screws and linkage adjustments do.
Plus, those ones and zeroes that define a certain tune can be very easily transmitted and duplicated. All you need are the proper tools to facilitate the tuning and conversation process.
Our tool of choice was the Cobb Tuning AccessPort. The AccessPort is a handheld device about the size of a small cell phone that connects to the car’s OBD port via a USB cable. The gadget retails for $695 and comes preloaded with several tunes representing digital setups that are designed to optimize certain hard part configurations.
For example, there’s a stock tune, which is exactly what it sounds like. There’s also a performance tune, which optimizes settings for premium fuel but assumes no other hardware changes. Then there are tunes that take advantage of certain aftermarket hardware like cold-air intakes, exhaust systems, downpipes and alternate intercoolers.
Although we did perform some chassis dyno pulls with our Speed3, this doesn’t seem to be the hot yardstick for these cars. Mazdaspeed3 dyno runs, as we have found, can be erratic and somewhat inconclusive.
After talking with some engineers at Mazda, we discovered that the root of this inconsistency can be traced to the top-mounted intercooler. Airflow through a pair of vents molded into the hood cools the intercooler, and it expects a certain velocity and volume of air. Those conditions rarely exist at the local chassis dyno shop.
Some shops have assembled large arrays of industrial-sized air movers that are then placed in front of the car during dyno testing, and this seems to produce repeatable results. However, note that we’re talking about six to eight of those barbecue grill-sized turbine fans, and we simply didn’t have access to such a setup.
Our definitive test would involve actually driving the car on pavement. Crazy, we know.
While zero-to-60 runs have been homeroom fodder for generations, we prefer to do third-gear, 30-to-70 mph runs. We time them with our Vericom data acquisition computer. This method runs the engine through nearly the entire powerband while eliminating the potential for human error—the launch and upshifts have been removed from the equation.
Basically, it’s a lot like making a dyno run, except instead of the dyno you’re using the road; and instead of the dyno operator, there’s a guy standing beside the road wondering why you keep driving back and forth over the same stretch of pavement.
We started with baselines runs, and our initial 30-to-70 mph sprints took 5.45 seconds. While the car felt strong, we definitely noticed a loss of power starting at 5800 rpm. Those last few miles per hour seemed to take an eternity to reach.
Then we tried the AccessPort’s Stage 1 tune. This setting optimizes the use of premium gasoline, and provides a bit more fuel and timing advance. Maximum boost is also raised, in our case going from 17 psi to 18.5 psi. Basically, this tune is designed for use on an otherwise stock car, so it’s a “clean hand” modification—no wrenches or screwdrivers are involved.
The AccessPort also allowed us to monitor a few dozen parameters, as it provides a wealth of real-time data, including ignition timing, turbo boost, injector duty cycle plus air intake temperatures both before and after the intercooler. Basically, if the ECU monitors it, the AccessPort will display it. It also features fuel economy and performance meters, while a dyno function is under development.
Installing the tune is decidedly simple. Plug the AccessPort into the car’s OBD port—it’s right on the dashboard—and select the desired tune. The unit holds the user’s hand the entire way, and the whole process of switching maps takes less than five minutes.
And it’s a five-minute workout that makes a difference, too—a serious difference. Our post-tune 30-to-70 mph times dropped from 5.45 seconds to 4.80, and from behind the wheel we noticed both better throttle response and improved thrust.
Sadly, the power continued to fall off at 5800 rpm. The drop was no longer so dramatic, but we still had a definitive plateau past which the engine did not like to rev.
Open it up
Seen from below, our AEM cold-air intake eliminated a lot of the complex and restrictive stock plumbing, resulting in a big performance boost.
Another chat with the Mazda engineers pointed us toward the intake plumbing. The Speed3’s stock intake plumbing is configured to reduce some of the intake noise, and as a result it’s a labyrinth of twists, turns, boxes, hoses and trap doors that all conspire to choke the engine of airflow above that 5800 rpm plateau.
Several firms offer intake systems for the Speed3, and all have their merits. Mazdaspeed even sells their own version, and it doesn’t affect the car’s factory warranty.
We ordered the AEM piece, which is very similar to the Mazdaspeed version. It’s beautifully powdercoated and looks clean when installed. Suggested retail price is $342.84, but we have seen street prices a bit lower.
Installation is straightforward, if a little jumbled by the sheer complexity of the stock intake system. There’s a lot of stuff to take out, and it needs to be done from both above and below the car. The driver’s side wheel well liner must also be removed to gain access.
Thankfully, the directions are excellent and feature photos and illustrations that ensure the right stuff is removed and replaced. A decent hobbyist will find the job easy—but not quick. It helps to have a friend involved, especially for the reassembly phase. Properly aligning the new intake tube and tightening everything requires three to four hands. Unless you’re Vishnu, call a friend.
Now for the payoff. While the effort put forth in installing the AEM cold-air intake was far greater than doing the digital tune—actually, changing the radio station takes more effort than that—the payoff was similar.
We reinstalled the stock tune and saw 4.85-second 30-to-70 runs—about where we were with the stock intake and the Stage 1 tune. However, the top-end performance came alive: The post-5800 rpm falloff was gone, and the engine pulled cleanly and strongly to the redline. So strongly, in fact, that we hit the rev limiter during a couple of through-the-gears runs; we were caught off guard by the newfound power.
Now we were very encouraged to try some of the other available tunes. We uploaded the Stage 1 + CAI tune, which is optimized for an aftermarket intake installation. Our results were slightly better—4.78 seconds to accelerate from 30 to 70 mph—but we had hoped for better gains.
We called Cobb Tuning, and they suggested that we go back to regular Stage 1 tune. This turned out to be the magic formula, as our 30-to-70 times dropped to a scorching 4.40 seconds. Traction had now become an issue.
So, what happened? Cobb says that each tune is optimized for certain parameters, but every customer situation has its own little nuances like elevation, humidity and local fuel condition that can lead to some configurations working better than others. No matter what the explanation, we had cut a full second from our 30-to-70 mph pulls.
There was still the issue of our stock shifter. It was a little notchy, which is not always a bad thing, but it was also a bit vague, which is always a bad thing. It felt like the stock shifter didn’t have enough leverage to properly deal with the resistance in the transmission.
Now, here’s where we say that we’re not always fans of aftermarket shifters. Some are great, true, but others just turn a lousy shift action into a shorter lousy shift action. Also, no shifter will ever affect the inherent characteristics of how a transmission slides into gear. That’s up to the linkages on the transmission plus the gears and synchros themselves. The best a shifter can do is provide a driver with different leverages and angles to handle the same situation—and that’s the genius of the Cobb Double Adjustable Short Throw Shifter. Yep, Cobb Tuning bailed us out again.
As its name suggests, Cobb’s Double Adjustable Shifter is adjustable for overall height plus shifter leverage and throw. It’s also a solidly built piece of equipment that feels like it’s been hewn out of granite. Cobb gets $274.95 for each one.
As with the intake, installation is relatively simple. It’s also time-consuming, as a lot of interior trim must be removed to gain the necessary access. Some fine work must be done as well, such as taking out the plastic cup that the shifter ball rides in, and removing and attaching the linkages in the right order. Those without dexterous fingers might want to enlist professional help.
After installation, we fiddled a bit with the adjustments and found a setting we really liked—a slightly shorter throw, but with a slightly taller-than-stock lever to provide a bit more leverage. It really did the trick. The notchiness was still there, but the vagueness was gone, replaced by a firm and positive shift action.
Still in Love?
So, the car is faster and shifts better, but is it still wonderful? Well, yeah, actually. We’ve managed to make the Speed3 faster by a considerable margin and improved the shift feel, but with no real caveats that we have noticed.
Next we’ll approach the handling, and that’s where we get real nervous. We want to give the car improved grip and response to match its newfound power, but we don’t want to ruin its pleasant demeanor.
Also, just installing the intake bumped us into the SCCA’s STU autocross class (with the stock computer tune in place). We also need to make sure the rest of the car is up to the task, don’t we?
Cross your fingers. We hope the scorching affair continues.
View comments on the GRM forums
I own a 2007 speed 3 and have it decently setup for DSP but run it in SMF as there is a decent amount of competition in that class locally.
I have a couple suggestions for you to improve the shift feel. 1 take the stock counter weight off the shift linkage this is that huge weight under the stock aitbox. Also a good stiff rear motor mount. The JBR88 rear mount has surprisingly little added NVH and makes shifts much smoother.
Also to get better performance above much over 6k it really needs a different turbo. But there are a couple things you can do to make it not loose so much up top. Get a bigger SRI style intake 3" inner diameter or better. This will also allow you to extract more power out of the k04. The next is a 3" downpipe with either 1 high flow cat or remove all of them. Then get a custom tune. There are some great etunners out there. One has tuned the highest hp mzr out there right now. I hope I see a response to this. I love mazda speeds and glad to see someone like grass roots writing about them
You'll need to log in to post.