Project Ford Probe GT
Continued head-banging yields slightly improved results
By Dave Coleman

Followers of Project Probe GT will remember that our previous attempts at improving performance met with little success. While the suspension modifications (Intrax springs, Tokico Illumina shocks and 235/40-17 Yokohama A520 tires on Axis Touring Cup wheels) were quite successful at improving the Probe's already impressive handling, engine mods seemed almost pointless. At the end of our last installment after a series of different intake configurations and ECU tuning misadventures, the Probe was equipped with an HKS exhaust and a Split Second ARC2 that replaced the restrictive air flow meter with a massive hot-wire air mass sensor and K&N filter from a Mustang Cobra. Neither had delivered the performance we expected, but both offered enough promise that we expected them to be beneficial when we finally found whatever was limiting horsepower.

Our next step in the quest for power came not from attempts to increase airflow, as with our previous improvements, but from reducing power loss from belt-driven accessories. In other words, we installed an underdrive crank pulley from Unorthodox Racing. The theory behind underdrive crank pulleys is sound, universal, and virtually impossible to thwart. In other words, it had to work. No amount of anti-tuner design work on the part of Mazda's engineers (who designed the Probe's engine) could prevent a properly designed underdrive pulley from improving power. The concept is simple. The accessories on an engine (alternator, water pump, power steering pump, air conditioning compressor) take power to drive. More importantly, the faster you turn them, the more power they consume (this is more true with the pumps than it is with the alternator). By making a crank pulley that is slightly smaller in diameter than the stock one, the drive belts can be slowed slightly, which slows the accessories, reducing the amount of power they consume. Voila! There is more power available for roasting tires, blowing off Mustangs, or whatever other uses you have for it. Given the car's lackluster response to airflow improvements, something that had to work sounded like a good idea to us.

We chose an Unorthodox Racing pulley, which was available anodized in a variety of bright colors. Since we couldn't actually see it after it was installed, though, we just chose silver. The pulley is a simple, direct replacement for the stock pulley, requiring only a change to shorter belts to make up for its smaller diameter. Installation should have been a snap, except that the instructions didn't say what size belt was needed, only that it had to be a little shorter. When St. Andre's did the installation, it took about twice as long as they expected, since they had to go back and forth to the store trying on belt after belt. Back at Jackson Racing [(714) 891-1113], we finally had a dyno session we could be happy with. Right off the bat, the Unorthodox pulley showed a consistent gain across the powerband, with peak power up 4 hp, and peak torque up 4 lbs-ft. It was a rare and wonderful moment of satisfaction in an otherwise grim and confusing project. Four horsepower might not sound huge, but it is just what we expected, and this was the first time the Probe's engine has turned expectations into reality.

With a new, brighter outlook on the Probe, a spring in our step, and visions of horsepower back in our heads, we turned back to the engine with renewed vigor. The power gained from the Unorthodox pulley was not actually an increase in engine output. Though power at the wheels was increased (and that is what matters in the real world), that increase came by reducing external power losses, not by the engine actually making more power. So in reality, we had not gained any ground with the engine, we were just in a better mood (and we had four more horsepower).

Looking back at our changes so far, we have gotten negligible gains by making attempts to increase airflow. There are two reasonable explanations for this. The first possibility is that although both the HKS exhaust and the Split Second air mass sensor conversion opened the door to allow significantly more airflow, no more air actually flowed because there was another door closed elsewhere in the engine. The closed door could be restrictive exhaust manifolds, a restrictive intake manifold, poor port design in the heads, or extremely mild camshafts. If it is either of the first two, we will know before this project is over; if it is head or cam related, it may be hopeless to make power within the budget of the average enthusiast. Of course, the problem might not be airflow related at all. The second possibility is that both our changes so far have increased airflow, but the ECU has dialed back ignition timing, delayed the opening of the secondary butterflies in the intake manifold, or used any number of evil tricks to prevent that extra airflow from becoming extra horsepower. Though this seems unlikely, every engineering team is motivated by different performance goals and different external forces. Some will design engines that overshoot their performance goals and then choke them down with easy-to-remove external plumbing. Some will choke it down internally with heads optimized for non-performance goals and others will do the same with tuning, possibly even making the ECU adapt itself to maintain an even performance level. Only time and continued testing will tell us where the performance is hiding.

Given our difficulties getting more power from the ECU so far, we turned to the next level of obvious flow restrictions—the exhaust manifolds. Externally, at least, the Probe GT's engine appears to be a perfect example of modern engine design. Besides the basic architecture with a deep-skirted aluminum block, integral main bearing bracing, dual overhead cams, four valves per cylinder, etc. The engine also features a variable tuned length intake manifold designed to provide good high rpm power as well as good low-rpm torque. It is surprising, then, to see the cast-iron lumps that pass for exhaust manifolds on this car. If they were made for four cylinders each instead of just three, they could easily pass for 1955 Chevy Bel Aire manifolds —they look that crude. Another well-designed V6 with crude exhaust manifolds comes to mind when looking at the Probe's engine. The 3.0-liter Acura NSX engine is a beautiful (and powerful) piece of design work, but it is strangled by log exhaust manifolds similar to those on our Probe. When our sister publication, Street Power, tested DC Sports headers and exhaust on a 3.0-liter NSX back in April of 1997, they found 21 hp! While we certainly weren't expecting that much of a gain, we did have high hopes for a set of headers.

When our new Brospeed headers arrived in the mail they consisted of an impressive assortment of parts. Merging the exhaust gasses from the two banks of a transverse V6 is no small feat, but doing so while keeping all those exhaust gasses happy, allowing all the O2 sensors and EGR ports to work, and fitting within the tight confines of the Probe's engine compartment makes for a serious challenge. Because of the limited space, there is no room to make the headers long enough for proper tuned-length operation. Ideally, the primary tubes would have to be well over a foot long to get the maximum pulse tuning benefit, but even with relatively short primaries, the Brospeed headers looked to be much more free-flowing than the cast-iron logs they would replace. Each header's 3 to 1 collector is relatively long, leaving lots of room for the gasses to merge smoothly. The secondary pipes connecting the two banks were also redesigned to be of equal length and to merge smoothly.

Unfortunately, since our car was a '97 model, and the headers were designed for the '95 model, installation was far from simple. Although the ‘93 to '95 and '96 to '97 models are largely the same, the addition of OBDII on the '96 model caused some changes, many of which happened on the exhaust. Since we expected some problems, Brospeed recommended that we take the Probe to Pro Automotive, Fastrax Turbo's repair shop, since they frequently work with Brospeed verifying fitment of new ports.

Last time, our installation of the HKS exhaust was complicated by the fact that the '97 catalytic converter was closer to the engine than the '93 to '95 model. St. Andre's made an adapter that went between the rear of the catalyst and the front of the exhaust to make up the space. Not surprisingly, the relocated catalyst was a problem with the header as well. Before tearing out the stock parts, we lined the rear flange of the Brospeed collector/downpipe with the front flange of the catalyst and found that we would be off by at least 6 inches! Most of that mismatch was eliminated by removing the St. Andre's spacer and bolting the catalyst directly to the exhaust, the rest would have to be taken care of by cutting a short section out of the rear of the collector/downpipe. Since we would have to add an 02 sensor fitting to that rear section anyway, we weren't too concerned about the extra work (the extra O2 sensor is part of the OBDII system that is designed to verify proper functioning of the catalytic converter).

In case you were thinking of installing headers on a Probe yourself, take our advice: Don't. You would probably he better off just spending the day sitting naked, slathered with honey, on an anthill. If you need proof that do-it-yourself header installation is a bad idea, just look at the rear exhaust manifold. Go ahead, try to see it. The rear manifold is buried so deep behind the engine that it is virtually impossible to see, let alone remove. To make matters worse, the EGR pickup is on the rear manifold, as if just getting to the manifold bolts weren't difficult enough. Just take the car to a qualified shop and don't worry if you hear lots of banging around and swearing while they work on your car, that's part of the removal procedure.

On the street the engine seemed stronger, but that could have been a result of the improved sound. The Probe's engine note is absolutely fabulous, but too subdued. Every time we make the engine louder it sounds even better. The only exception to that rule is the slight hiss we get now on coastdown. It appears that one of the flanges on the front end of the collector/downpipe isn't quite angled properly, causing a slight exhaust leak. We haven't figured out an easy solution to that problem yet, so for now we are just living with it.

On the dyno, well, while we can't really say that our previous experience with the Probe hadn't prepared us for disappointing results, you are never really ready when something that absolutely should help... absolutely does not. The power and torque curves before and after six agonizing hours of header installation were virtually identical. The peak values actually show a 1 hp and 1 lb-ft loss with the headers, but that appears to be a fluke, since output is the same everywhere else on the curve. The headers, in fact, did nothing one way or the other.

When asked about the disappointing results, Brospeed's Dan Brown suggested that the problem probably is hidden somewhere in the changes made between the '95 and '96 model Probes. The '95 model that they designed the header on, Brown says, saw about a 10 percent increase in power with the headers. They have also tested their exhaust system on '95 and '96 models and found the exhaust didn't help the '96 Probe nearly as much as it did the '95. They had not tried the headers on the '96 since that year had close-coupled catalysts built into the exhaust manifolds, and Brospeed's policy is not to remove catalysts. Strangely enough, our '97 Probe had manifolds that were designed to have catalysts in them, but they were hollow. Apparently the close-coupled catalysts were deemed an unnecessary expense after a year of production, so the expensive working guts were left out.

The most interesting thing we noticed when reviewing Brospeed's test results from the '96 Probe was the fact that they say only about a 2 hp gain with their exhaust, and another 2 hp by completely removing the muffler. What this suggests is that there is little power to be gained with the exhaust on a '96-'97 Probe, no matter how good the exhaust is. That could explain our lackluster results with the HKS exhaust. In all fairness, none of the exhaust pieces we have tested on this car were designed for the '97 model, so we shouldn't be too surprised when we don't get stellar results.

So what now? There is only one possible airflow restriction remaining that is easy to remove. The intake manifold is strangely convoluted inside, with the air being forced around an inconveniently placed set of secondary butterflies. Besides that, the runners are quite long, and have a relatively rough surface texture. If the many rough edges on the convoluted innards of the manifold are a problem, Extrude Honing will fix it. If not, at least we'll know for sure. Besides further attempts at finding our power, we also want to do a little more work on the handling. While Probe has a steady-state grip stock, and excellent steady-state grip now, it is still a serious understeerer at the limit. We are hoping to try getting the rear wheels a little more involved concerning. Until next time, we'll keep banging our heads.

©2000-2001 Performance Probe, Inc. All rights reserved.