Project Mazda 323 GTX: Part 4

Project 323 GTX lives on as one of the most enjoyable heaps any of us have ever driven. In fact, after this month's round of upgrades, it's got a speed to ugliness quotient matched only by the most awful muscle cars. However, as most of you know, the GTX is quick in a very different way. What muscle cars do on straight, smooth pavement, the GTX does in the dirt. Give it five minutes on any gravel mountain road and you'll be begging for more. Very fun.

Since the last installment of Project 323 GTX , we've spent little time concerned with anything except power. Granted, we've been through a few drivetrain and suspension parts (minor repairs mostly due to playing in the dirt). But for the most part, we've focused on safely turning up the boost on our low-dollar, made-for-the-mud hot rod. This month we'll talk about the simple, cost-effective nature of making power with the GTX. It's time. We're sick of screwing around. Let's turn up the wick.

Fast, Simple Power
Believe us when we say this is the easy part. If you've been following Project 323 GTX since the first installment, you know how many hours are involved in replacing wheel bearings, transmissions, clutches and suspension components. As rewarding as these changes are to the car's dynamics and drivability in the dirt, they did nothing in the way of making good, old-fashioned horsepower.

Here's the good news: Adding 40 hp to your GTX can be done safely, and relatively cheaply, in one evening. There is no bad news, no disclaimer and no "but" that goes with this statement. Sure, if you get stupid and don't pay attention to boost control adjustments or spend disproportionate amounts of time at high boost, something will probably give. But, if you use your head and have a little common sense, the GTX stands up like a champ to these upgrades. According to Mike Welch of Road/Race Engineering, modified GTXs have been known to last as long with these power mods as they do stock. And given the age of our GTX, it seems there's been plenty of time for this theory to prove itself.

Intake and Exhaust
The GTX's stock airbox (and airflow meter) is notoriously restrictive. The only solution we know of for the airflow meter would be to go to a proven, stand-alone engine management system such as Electromotive's TEC II or another similar speed/density system which relies on a MAP (Manifold Absolute Pressure) sensor and engine speed to calculate necessary fuel delivery. This is not an option on Project 323, as cost is a major factor and these systems cost more than our exhaust, intake, boost controller and ECU upgrades combined.

So, when it came time to upgrade our intake, Welch pulled out an old HKS adapter that allows a mandrel bent piece of steel tubing to be fitted to the inlet of the GTX's airflow meter. Within minutes, the crew at RRE was cutting, welding and painting the new intake bits. Two brackets were welded to the steel tubing and a K&N open-element filter was clamped onto the end. The brackets fasten the intake firmly to the fender wall and frame rail and locate the filter in the left front corner of the engine compartment. Welch said that similar intakes can be had for about $150.

We were initially concerned about dirt and dust entering the engine when using an open-element filter such as the K&N. After all, our adventures in the gravel place unusual demands on anything designed to seal, filter or protect any part of our GTX. However, we soon realized that an extra measure of protection was available at the nearest department store. With a nylon stocking stretched over the filter, we were considerably more confident about playing in the dirt with our new, easy-breathing intake.

HKS' exhaust was next. When we bought it, our car was fitted with a pathetic pinch-bent muffler shop chop job from the catalytic converter back, so moving up to an real aftermarket exhaust was an absolute must. Besides, there's no point in turning up the boost without removing the rest of the horsepower plugs.

HKS' cat-back exhaust uses 60mm aluminized steel from the cat back and fit easily into place using the original hangers. Breaking Project 323 tradition, the bolts from the catalytic converter back loosened without breaking off--a pleasant surprise.

Despite rumors to the contrary, this exhaust is still available from HKS. While it obviously doesn't sell very many, HKS still has two complete systems in stock at the time of this writing. According to Ben Paske, the exhaust will not be discontinued as long as there is enough demand to justify keeping a few in stock. If demand increases, it may take HKS a few months to get new systems from Japan, but they will still be available. The exhaust is part number LET-Z04 and retails for $495.

Boost Control 101
Road/Race's manual boost controller for the GTX is a simple combination of silicon hose, a needle valve, a few T-fittings and one critically sized restrictor. The boost controller taps into the car's stock system in three locations. The pressure source is taken from the stock intercooler piping--exactly where the stock unit sourced pressure. The pressure source then tees into two lines--one to the wastegate actuator and another to the needle valve. A restrictor orifice is placed in the pressure source line to maintain the correct pressure ratio to both lines downstream of the tee fitting, ensuring the wastegate opens at the desired boost. By creating a leak in the wastegate actuator line, the bleed valve effectively increases boost by the amount being leaked. From the bleed valve, air is routed back into a breather hose after the airflow meter to avoid confusing the mass-air system. In all, an inexpensive, effective method of controlling boost.

The boost controller comes from RRE with all appropriate instructions and warnings. Boost is adjusted by turning the needle-valve knob a quarter-turn counter clockwise at a time and carefully watching the boost gauge (something we'll deal with in the next installment). Boost should be evaluated at full throttle under load (third or fourth gear works best). Peak boost should not go beyond 16 psi at about 4000 rpm. With a stock turbo, it will fall rapidly as engine speed increases, eventually stabilizing at 12 or 13 psi by redline.

Manual boost controllers get a bad rap. That's what Mike Welch thinks and we agree. Look at the price of any electronic unit on the market, compare that with the $60 combination of silicon hose, bleed valves and T-fittings available from RRE and the decision is a no-brainer, especially on a car like the GTX.

There's a good argument for the precision of an electronic boost controller when running near the ragged edge on many cars. The GTX, however, seems to be a bit of an exception. Permit us to explain our reasoning, before you pick up your pens or get crazy with your e-mail. Our GTX's relatively small (and somewhat tired) IHI turbo doesn't have the efficiency (size) to flow much more than 15 psi of boost to redline in anything other than first gear (and even then, it struggles considerably). Nor does the car's fuel system, even with an ECU upgrade, have the capacity to keep up with more than 15 lbs of boost at redline. It does, however, have the capacity to move enough fuel to provide a considerable torque and power increase to 16 lbs of boost, and slightly beyond, at about 4000 rpm.

Here's where things get tricky. Normally, the less-than-precise boost spike common to manual boost controllers is seen as a drawback, since it risks lean-run conditions at high boost and rpm. On the GTX, however, that boost spike comes between 3500 and 4000 rpm where the car's fuel system (with RRE's chip upgrade) can increase injector pulsewidth enough to provide sufficient fuel. Look carefully at our dyno chart. See the torque peak just below 4000 rpm?

Notice that it's 55 lb-ft higher than the stock curve. That's a good thing. While the turbo and fuel system can't maintain that output all the way to redline, having this much torque at 3600 rpm makes the GTX much more impressive in the dirt where torque is essential. Rallycross driving is made considerably easier, as are our romps in the mountains. Peak power is also up about 30 hp while a maximum gain of 40 hp is available at 3800 rpm.

Chip Upgrade
Also absolutely necessary when turning up the boost on the GTX is some means of fuel management. After all, without more fuel, all that extra boost is going to be spent destroying your engine rather than making your car go faster. This issue has been addressed by several means in the GTX. HKS' PFC F-Con, a piggyback fuel computer designed to remap the GTX's stock fuel curve, is one option. We didn't take that route. We preferred the less expensive (and less complex) RRE ECU upgrade. The chip is installed in the factory ECU (in our case, we replaced the entire ECU) and allows the GTX's stock fuel system to operate safely at boost pressure approaching 16 psi.

According to Welch, the chip handles only fuel enrichment. For ignition control, the GTX relies on a centrifugal advance and boost retard built into the distributor, combined with a separate knock control system outside the ECU. The fuel enrichment capabilities of the upgraded ECU apply only at full throttle. The car's stock fuel system calculates injector pulsewidth based on engine speed and signals from the airflow meter and throttle position sensor. However, because the flapper-door style airflow meter is pinned wide open at high engine speeds, set algorithms are used to calculate stock injector pulsewidth. Here's where the upgraded chip comes into play. The new chip re-maps those algorithms to allow longer injector pulsewidths, making more fuel available at higher boost levels. Simple and effective--like everything else we've done on this car.

According to Welch, the fuel cut on a stock GTX is activated at 11 psi. With the fuel cut bypassed, most GTXs can run to 12 to13 psi safely. The ECU upgrade allows 15 to 16 psi depending on the car's condition and octane of the fuel used.

Cool Air, More Power
Most GTX owners know by now that the tiny stock intercooler is severely undersized when it comes to big boost. Plus, running big boost means adding heat to the system; adding heat to the system means losing power. That's not exactly our goal, now is it?

There's a simple solution for all this. The first-generation Ford Probe/Mazda MX-6 intercooler offers larger volume and a more efficient design. Couple those traits with the fact that it's a near bolt-in and it's the perfect candidate for replacement of the GTX's pathetic stock intercooler. We found a Probe donor car at a local junkyard and even had the joy of cutting its front bumper in half to get to the goods. A little hacksawing and $25 later and we were again on our way to Road/Race for another power upgrade.

Fitment requires trimming the stock inlet and outlet pipes and re-beading the shorter pieces to allow for proper sealing. Since the new intercooler fits in the same location as the stock unit, very little is required in the way of custom fabrication. We bent and drilled new brackets and asked RRE to weld aluminum alignment pins to the bottom of our intercooler so it could use the same holes to locate its base as the stock unit did. Road/Race used pieces of a stock Eclipse intercooler hose and a few hose clamps to plumb intake air back to the stock intake pipe across the valve cover. We inserted one of the pieces we cut off the inlet tube and used a few more hose clamps (see photo) to make the fit perfect. The stock intercooler hose is used from the turbo outlet to the new intercooler.

On the road, the Probe intercooler exhibited substantially cooler on-boost intake temperatures than the stock unit. It was also less susceptible to heat soak after idle. Testing was conducted up hill in third gear. Intake temperatures were recorded cruising at neutral throttle at 3250 rpm. Intake temperatures were then recorded at full throttle at 4000 rpm and 6000 rpm. Boost was also recorded and registered at 16 psi at 4000 rpm and 12 psi at 6000 rpm with both intercoolers. While testing with a 69-degree ambient temperature isn't exactly an intercooler torture test, it was all we could manage in California's fall weather. Our testing revealed the Probe intercooler lowered on-boost intake temperatures almost 20 degrees (relative to the stock intercooler--see chart) under normal conditions and more if heat soak set in. In addition to the expected intake temperature drop, the new unit also proved to be less susceptible to absorbing radiant engine heat than the stock unit. We observed almost no noticeable effect of heat soak with the new intercooler while the stock unit allowed intake temperatures to increase an additional 5 to 10 degrees if given no airflow for a few minutes. Bottom line, the Probe intercooler appears to be worth every penny of its $25 junkyard price. Go get one.

Loose Ends
Before making any of these modifications, it's absolutely critical to be sure that your GTX is in good working order. While these upgrades seem minor and are relatively inexpensive compared with a more modern car, the effects could be catastrophic if any of the ignition problems we've addressed in previous installments weren't handled properly. In other words, be sure that the spark plug wires, distributor cap and rotor are all in good shape before you make any of these changes. We replaced the spark plugs in our car with NGK BKR7E plugs shortly after the boost upgrade to ensure proper burn. Welch noted that plug gap is critical on the GTX when running high boost and should be set between .028 and .030.

Even More Power?
Making more power on the GTX is certainly possible. However, careful consideration of its value is necessary if going beyond this level. First of all, the GTX is now 12 years old. The modifications described above are all relatively well known and accepted as viable means to make more power while maintaining reasonable reliability and cost.

Welch told us that the next logical step would be a larger downpipe (which requires removal of the car's air conditioner), ported wastegate and high-flow catalytic converter. If one is replacing a ruined turbo it makes sense to get the larger downpipe and high-flow cat anyway. With these mods, GTXs have been known to run 18 or 19 psi of boost on race gas. However, once the little B6T engine reaches these power levels, it's only logical to have full instrumentation (exhaust gas temperature, fuel pressure and water temperature gauges) on the car to monitor its health.

As discussed, if making this kind of power is your goal, it makes sense to transition to a full, stand-alone engine management system. Of course, at that point a turbo capable of maintaining boost at high rpm makes sense too. That means a budget beyond what we have for this project, so it's likely we won't pursue power beyond the levels we've reached at this point.

Additionally, more power means more chances of breaking rare and expensive GTX drivetrain parts. Axles, transmission parts and clutches seem to be likely candidates for failure with increased power and four-wheel drive grip. The last price we heard for a Group A rally transmission for the GTX was about $5,000. That's enough motivation for us to leave things alone.

Sources
HKS USA
2801 East 208th Street
Carson, CA 90810-1102
(310) 763-9600

Road/Race Engineering
15542 Graham Street
Huntington Beach, CA 92649
www.roadraceengineering.com