Project Elantra: Huffin’, Puffin’, and Blowin’ Up

kuhol manga page

Last time in the annals of Project Elantra, we talked about its base engine setup. We hope that the revelation of the Hyundai equivalent of the “Vitara piston” mod piqued your interest as to what the rest of the setup contains. And maybe perhaps buy Elantras to set up and race with.

If you’ve been reading the site recently, you’d know that the car’s engine has catastrophically failed while participating in an amateur circuit race a few weeks back. While in the middle of writing this article, we did not expect that our engine would conk out on us. You can expect the next Project Elantra update to cover the rebuild (no, we’re not going to swap the engine for a Mitsubishi 4G63T). In the meantime, do read up on the litany of power-producing baubles present in and around the engine. Which are currently probably scattered about the workshop we sent the car to.


But before more revelations, we have had reactions conveyed personally that perhaps we share too much about the Project. Since it is still active as a race car, we should be more coy with what we detail publicly, especially about the engine. We say that’s hokum.

First of all, we’re not into street racing. Who cares if everyone knows our setup? We race in fields that are safe, legitimate, and where the weight of performance is skewed favorably towards driving ability than vehicle setup (not that the latter’s unimportant, mind). We don’t need to keep secrets.

Besides, not one of our present competition, if he/she is in a right frame of mind, will follow our setup. The fastest guys in slalom are running naturally aspirated motors in chassis platforms that have racing pedigrees and are stacked with decades-worth of tuning experience. We’re running a turbo Elantra that’s not much in the running (yet) as the fastest front-wheel driver and has reliability issues, which will get licked, slowly but surely. Unless we start winning with the car, having the fear of copycats will just be a waste of time.

We have some ideas on how we want to take the Project further, and those are the trade secrets. However, most of them will be conspicuous once deployed. We mentioned this before – we race because we enjoy the challenge of fielding an uncommonly raced marque and model. It has taken seven years and untold difficulties to get it this far. We still enjoy it – we’re not gonna stop now.

There are limitations to what we can do with this chassis. If one wants to win in slalom, and have a super-seven figure budget, here’s a cocktail to numb your brain over. Mix one part Mazda Miata with equal part Chevrolet LS1 engine. Serve and enjoy.

Image credit:

Turbo Inside

Our chosen forced induction device was a Garrett GT2871R turbocharger that was a leftover from a previous project. This small frame turbo is around the size of a T28 and it fits well inside the engine bay, in the narrow area behind the radiator and fan shrouding. Like the engine, the turbo is rated to support up to 400hp that is partly due to the 71mm compressor wheel and the improved aerodynamic design of both compressor and turbine fins that is a revolution compared to the T28. Boost threshold levels are lower than the equivalent T28 as the turbo is fitted with a ball-bearing center cartridge.

We first heard of the GT series of Garrett turbochargers from Sport Compact Car’s article about the Disco Potato. The unfortunately-named car was a Nissan Sentra with color-shifting brown paint, hence called Disco Potato. It had an SR20DET which was plumbed with, at that time, a prototype turbocharger that eventually became the GT28RS, now called the GT2860R. The turbo had a very low boost threshold when fitted to the Potato’s engine, and had gobs of torque for the majority of the powerband. Thus, the turbo maintained the OEM turbo response while at the same time producing a way decent chunk of power, impressing those who were fortunate enough to sample it.

Not knowing any better, we purchased the same turbo with a bigger, 71mm compressor wheel, thinking it was one notch better than the Disco Potato Turbo. Still, when fitted to Project Elantra, we can’t help but be chuffed by the turbo’s performance. Even with a horribly low compression ratio and an exhaust leak from the manifold-to-external-wastegate flange, the turbo spools at around 40% of the rev range. Below 3000rpm the car feels sluggish due to the very low compression ratio. Past that rev range though, you can feel the turbo winding up, and at 4K, the car suddenly goes insane. The engine just pulls until redline, a sensation that never, and we mean never, gets old.

We had the GT2871R’s integral wastegate deactivated and went with a Turbonetics Deltagate II external wastegate as per the advice of or tuner, Art Rodriguez of Haltech Philippines. The Haltech Platinum Sprint 500 controls the boost solenoid he plumbed in-series to the wastegate signal hose. He says he programmed a boost pressure of 17psi, but the boost gauge indicated 21psi during a few highway fourth gear blasts. It may just be due to the manifold leak, but we are not complaining of the overboost.

We are kicking ourselves at the foot, however, for our turbo’s untimely demise. More on that next time, but that issue uncovered the only downside to the GT2871R, and other Garrett GT-series turbochargers for that matter. When they break, they are expensive to fix. Anything wrong with either the compressor, turbine, or the center bearing, and you have to replace the entire center housing rotating assembly (CHRA). That is not cheap. A core exchange is possible, but you have to ship your old CHRA to the US. Shipping is also not cheap.

We’ll update you on our turbo issue next time. Till then, do light a candle for our fallen turbo.


Pipe Dream

Some sections of both charge and exhaust piping have been with us for like five or six years. The previous project not only handed us down the turbo but also some charge piping, a generously-sized intercooler, the Sheepdog blow-off valve, turbo downpipe, and a complete exhaust system right down to the Sheepdog turbo muffler. Certainly revisions had to be done to make components made for a 4G92A turbo kit adapt to an Elantra. Parts interchangeabilty? Yeah right.

We featured the exhaust manifold in the previous installment. We will be throwing it away and having a new one made. More on this at the next installment. From the turbo back, we reused the downpipe from the old setup, cut and rewelded the 2.5″ exhaust tubing with 24″ resonator to fit, and managed to install the JASMA-certified muffler. These were not done in one step, but in several stages and reworks over the years. As it stands right now, the car’s exhaust emits a quiet, deep tone that we prefer instead of the banshee wail belted out by some of our NA competition. The car is actually louder on the inside as we removed the majority of sound-insulating materials from the interior.

The intake and charge piping has had a more dynamic change-this-adjust-that life. We’ll not bore you with the histogram. We have tried to optimize the charge piping path and design over the multitude of revisions. The trick is to layout the piping to as short a path as possible, to as smooth a path as possible, and with minimal changes to the cross section of the pipe as possible. We have also reworked it when components in the engine bay were rearranged and reintroduced.

To make the charge piping as short as possible, we resorted to cutting holes at the radiator support structure, at either side of the radiator, so that the pipe has a shorter shot to and from the intercooler. All bends were made from mandrel-bent elbow pieces and joined using TIG welding to ensure a smooth, constant flow of the compressed air. The compressor outlet and the intercooler inlet were both 2-inches in diameter so 2″ stainless pipe was used in this section. From intercooler to throttle body, the pipe diameter was stepped up to 2.5″ to allow for the pipe to be properly coupled to the TB.

We reverted to a humdrum lead-acid 2SM from Imarflex, because it was cheaper delivered to our doorstep than buying a Motolite from the neighborhood. We put the battery back to the standard location, but the conical air filter was in the way. We had to relocate the filter to the only position available, which is the vertical space above the transmission. We also took the opportunity to buy an oversized fake Simota air filter to gobble up the real estate, making the engine bay seem all the more packed.


Quaife ATB Differential. This unit is for a Porsche. Image credit:

Quaife LSD

The presence of a Quaife limited-slip differential in Project Elantra is the second miracle of the powertrain, after the OEM low-compression piston/rod hack. Would you imagine, an LSD in a claptrap Korean car? For some reason, a 4G63 front differential from an Eclipse will quite well where the stock open diff used to be.

However, this particular LSD (and we surmise, any aftermarket LSD for the 4G63 tranny) would require the purchase of a genuine Mitsubishi part. The speedo gear, with part number MD717879, had to be ordered from the US just so that the speedometer will work again.

The Quaife LSD is described as a automatic torque biasing differential by its makers. What it means is that torque is sent to the driven wheel that has the more traction. This action is done via helical gears instead of clutch plates or viscous liquid.

For a front wheel drive vehicle, an LSD helps to quell torque steer, increase traction in accelerating, and help a driver improve lap times by allowing him to apply more throttle while cornering after the apex. You’d expect to severely understeer if you do it on a FWD without an LSD, but with Project Elantra, you can point and squirt with aplomb. We need more circuit racing experience to fully exploit this handling characteristic.

Why we went for Quaife is that the unit has a lifetime warranty and has no special maintenance requirements other than is needed by the entire transaxle. This is unlike clutch-type LSDs whose clutches wear out. However, the Quaife has no LSD preload effect, and thus becomes useless if one drive wheel goes off the ground. Unless we’re rallying, not a worry.


Tranny Tricks

All them horses are transmitted to the wheels with the 1.6L Beta transmission that came stock with the car. This gearbox features the closest ratios and the highest final drive produced by Hyundai for the first-gen Betas. It’s a bit crunchy when shifted to third, but pretty smooth since the gear oil change to Redline GL4 75W-90.

We surmise that the ‘box is similar to FWD units in some Mitsubishi cars, more specifically the 4G63/4G92 of the Pizza Lancer generation, where the tranny is on the LHD driver’s side of the car. Lending additional support to this hypothesis is that we fitted an Exedy tri-puck racing clutch for a 4G92M and a Quaife LSD for an Eclipse in the transmission.

With a low level of certainty, perhaps the Beta will accept the 4WD transmission of a Lancer Evolution 4 to 6. Looking at the bellhousing of both transmissions from pictures of the Internet, they seem to match up fine.

Even if the bellhousing be the same, other issues arise in the quest for 4WD. For one, the Elantra’s front subframe will not make room for the driveshaft. Either the Evo front subframe has to be adopted or a new custom subframe has to be made. At the rear, it is clear that the Evo uses a different suspension design than the Elantra. Adapting an Evo rear floorpan and the entire rear suspension/axle assembly has to be done. Sounds hard for a guy who can’t weld to do. We can’t weld.

If you want us to find out for sure what it would take to convert an Elantra to 4WD, all we need is Php 200,000 to buy a surplus Evo drivetrain and do the conversion. That should satiate the desires of who are asking. (Which are none.)


CV Fail

Our major problem with the drivetrain is with the constant-velocity (CV) joints. We have broken the passenger’s side CV joint twice, and we detect imminent failure of the driver’s side. The combination of massive torque, sticky tires, limited-slip diff, low ride height and obtuse driving spells early death for the clearly overworked CV joints.

The problem is not with the CV joints themselves. Whether they be OEM axle assemblies or replacement items, if the axle end is forced out of the CV case through torque and skewed suspension geometry, any CV joint will break.

We did the simplest fix for this problem. We raised the front ride height slightly, bringing it to a height that is about more than an inch lower than stock. (It’s great to have a stock Elantra to compare with.) We hope that brings back the geometry closer to what the car’s designers engineered, and that the tendency to break CV joints will be vastly reduced. That, and not run sticky tires for Slalom anymore.

Next Time

You’ve seen the car in jackstands in the very first Project Elantra article. You’ll again see the car broken and immobile. And what we’re doing to get it running. Oooh, exciting. Until then, as always, keep it slow out there.

Leave a Reply