After fitting a high-flow cold air intake to my MY94 Subaru Impreza WRX (see "Stage one of mods to AutoSpeed's MY94 Subaru WRX - the air intake...
") we had the car generating 123kW at the wheels - as measured on a Dyno Dynamics AWD chassis dyno in third gear. As you may recall, our reworking of the intake system yielded a reasonable gain at lower rpm but no real top-end power hike - only around 2 percent more than the standard '95 Japanese-spec engine fitted to the vehicle. Since then we've been wondering how much exhaust backpressure there is....
A Look at the Factory MY94 - 96 Impreza WRX Exhaust System
The '94 WRX exhaust system is very similar to those fitted throughout the rest of the turbocharged GC8 Impreza series.
Starting at the front, the standard Subaru turbo dump pipe is notoriously inefficient. With a flat metal plate butting up near the turbo wastegate outlet, all of the exhaust gasses are forced to mix and jostle to escape down the single 2 3/8-inch OD pipe adjacent to the turbine. This impedes flow of both the turbine and wastegate.
About six inches downstream of the turbo flange you can find a vertical catalytic converter; previous testing has shown this cat flows quite poorly.
After passing through the only cat converter in the system, gasses flow through more 2 3/8-inch OD pipework and a small empty chamber below the cabin. This is probably an expansion chamber intended to eliminate exhaust resonances. Note that extensive heat shielding surrounds the pipe from the turbo flange right to just past the centre chamber.
Along its entire length, the standard exhaust is made from 2 3/8-inch OD pipework, but the cross-sectional area decreases at a number of press bends. The most heavily pressed section is that which passes beneath the axle/rear sub-frame and enters the rear muffler. This muffler itself is a baffle type and works on the principle of reversion (rather than absorption) to quieten exhaust noise; judging from this photo, it appears quite restrictive. The system ends with twin 43mm OD tips.
Although not apparently great from a flow point of view, the standard exhaust is certainly quiet. With a SPL meter (dbA weighted) held at 45-degrees 50cm away from the tips we recorded 58dB at idle, with 62 and 64dB on revs up to 3000 and 4000 rpm respectively. To give you a guide, a standard VX HSV GTS300 has a relatively thunderous 63dB idle.
Potential Gains
While much has been said about the backpressure of the standard WRX exhaust system, we decided to identify if it is really that bad. Once again - as performed on our long-gone Holden VL Turbo - we made a connection to the oxygen sensor fitting at the start of the exhaust and measured backpressure at full load using a handheld gauge.
While we were at it, we also decided to measure backpressure prior to the turbo (using the pre-turbo oxygen sensor fitting as fitted to the Japanese-spec engine). These backpressure readings were then referenced against manifold boost pressure; boost pressure varies depending on atmospheric conditions and (where an open-loop ECU-controlled wastegate is employed) is known to overshoot when a high-flow exhaust is fitted. We 'mapped' all three data sources simultaneously through a series of second gear WOT runs using an in-car movie camera.
In this graph, the highest reaching plots represent pre-turbo backpressure, while the middle plots represent manifold boost pressure and the lower plots represent post-turbo exhaust backpressure. As you can see, backpressure before and after the turbo increases almost linearly with revs. Given that pre-turbo backpressure peaks at 180 kPa and exhaust system backpressure peaks at up to 60 kPa, there is a substantial 120-odd kPa backpressure occurring through the turbine/wastegate. Note, however, manifold boost pressure at redline rpm is around 68 kPA, which means the remaining 52 kPa of pre-turbo backpressure is written off as losses through the turbine and wastegate bypass.
What you can't see in the graph - since we were sampling data at every 1000 rpm - is that boost pressure spiked to a peak of 83 kPa at around 4300 rpm and pre-turbo backpressure was quite sporadic through mid-to-high revs.
So what does this graph tell us about the standard '94 WRX exhaust? Well, compared to - say - a standard Holden VL Turbo with up to 90 kPa exhaust backpressure, the Rex is not that strangled by its exhaust. As you'll see, however, impressive overall gains are still very achievable...
A Cost Effective Approach To Exhaust Replacement
In the case of our WRX we weren't looking for a mild improvement from, for example, just replacing the rear muffler. Since relieving exhaust backpressure is a proven way of enhancing turbo spool-up, reducing fuel consumption and - of course - adding power, we saw no reason to hold back. A full exhaust replacement was what we were after.
As you may already know, the cost involved in bolting a high-flow exhaust onto a WRX varies almost beyond belief. Some aftermarket WRX mufflers alone cost nearly $1000! On the other end of the spectrum, you can find cheap second-hand WRX exhausts appearing on forums and in the classifieds every week. As it just happens, I had already purchased a complete second-hand exhaust system about three years ago; a mild steel 3-inch mandrel system with a 'big mouth' turbo dump, 3-inch cat converter and a small dogleg rear muffler. The whole set-up cost around $300.
Buying a second-hand exhaust is, of course, a lucky dip. The particular exhaust I bought had been removed from a vehicle because the previous owner found it too loud. I, too, wasn't keen on having a headache inducing exhaust, so - rather than fit the system as-is - I thought I'd swap to a more effective rear muffler.
|
How to Find a Cheap Yet Effective Muffler
Picking a quiet muffler that flows well is not easy, but there are a couple of considerations to keep in mind...
First - as a generalisation - those mufflers that are large and relatively heavy are usually effective at suppressing noise. Car manufacturers don't fit large mufflers for fun.
Second, independent testing has shown that quality Japanese straight-through type mufflers perform brilliantly in terms of flow and noise suppression. The large body Trust, HKS and Fujitsubo mufflers (not the cannon-style ones) offer better noise suppression than any other aftermarket muffler we've seen. Like all straight-through mufflers, their flow restriction is also negligible - about 5 percent of an equivalent straight length of pipe.
With this in mind, we purchased a second-hand ex-R32 Skyline FGK (Fujitsubo Gikken) straight-through muffler from a Japanese import wrecker. Its body dimensions would j-u-s-t fit under the WRX floorpan and the cost was only $95 - a bargain! This photo shows the FGK muffler alongside the loud dogleg muffler originally used as part of our 3-inch system - note the difference in body size.
|
With a larger rear muffler chosen for the job, we then had to prepare the mild steel pipework for fitment; sitting around for three years had seen the pipe go rusty brown. A 10-minute rub back with a few scourer pads, however, brought the pipe up 80 percent as new. We paid particular attention to the inside of the pipe between the turbo flange and the cat converter - we don't want rust flakes clogging the cat's honeycomb.
The next stop was the exhaust shop.
Exhaust Fitment
I handed the job of installing the system to Mark Marchesan from Adelaide's Exhaust Technology. The fact that Mark has exhaust'd countless WRXs - not to mention that the system I bought was one of his earlier pieces - gave me a lot of confidence in his work.
Swapping the exhausts is a relatively straightforward exercise. Some of the bolts around the turbo are reluctant to be undone, however, so it's a good idea to douse them in CRC (or similar) prior to arriving at the 'shop; this helps ensure they won't shear.
Once the bolts have been cracked and the oxygen sensor and turbine cooling duct have been removed, the factory exhaust system can be removed from the vehicle.
The front dump pipe section of the ready-made system goes on first. Note that we asked to have a 3-bolt flange attaching the front section of exhaust to the centre - this makes future clutch/gearbox servicing easier. This is a WRX after all...
From the 3-bolt flange back, the existing 3-inch pipework was bolted on in barely a couple of minutes. Note that there is enough room left under the floor to insert a fairly large central resonator if later required.
Mark commented that his current systems are slightly different to the earlier piece being fitted to my vehicle. These days, his systems typically use a twin pipe dump arrangement, fewer welded together pipe sections and - where noise is a concern - a tapered 2 ½-inch rear section. Mark explains that the smaller rear section takes the edge off exhaust noise - something that makes sense given Air Power System's ADR approved power-up kit uses a similar approach.
When it came time to hang the big FGK muffler we were in for a rude shock...
Once Mark cut the 90-degree inlet elbow from the muffler we discovered its leading section of perforated internal pipe was stuffed - it was completely rusted through!
This muffler raises an interesting point. Mark says many of the older aftermarket mufflers he's seen from Japan are in quite poor condition; some are nearly fifteen years old and those that have been stored as individual components (rather than being imported on a vehicle) have been left outside to fill with rain... A good way to check the condition of a second-hand muffler is to bang one end on the ground and watch for any rusted perforated pipe or filling that might fall out. In the case of straight-through mufflers, you can also drag a hooked length of wire through the length of perforated internal tube to see if it's crumbling.
|
Temporarily Putting Up With the Noisy Little Dogleg Muffler...
Since we couldn't fit the FGK muffler, we were forced to temporarily fit the little dogleg muffler that was part of the original 3-inch system.
With the dogleg muffler in service the exhaust was very loud - it was no wonder the previous owner had concerns. A quick SPL meter test revealed just how loud the system was - 68db at idle (up from 58dB standard) and 97 and 100dB on revs to 3000 and 4000 (up from 62 and 64db standard)!
|
A few days later we fronted to Exhaust Technology once again armed with another large body second-hand import muffler; we first made sure that this one was in better condition...
The first step was to cut off the existing hanger brackets and remove the mandrel bent inlet pipe, which was suited to (once again) a R32 Skyline; with these removed, Mark could snugly fit the muffler under the Rex's floor.
Routing the 3-inch pipe the muffler was quite simple in this case; since a 2-bolt flange had been incorporated in the pipework just behind the rear sub-frame, only a new adjoining bend section had to be inserted.
With the muffler set as high as possible - to maintain decent rear ground clearance - Mark fabricated some new metal hangers, which were welded directly to the muffler body. Oh, and - rather than re-use the original rubber exhaust hangers - Mark used some extra heavy-duty rubber/plastic jobs that are a lot more rigid. This rigidity helps prevent mufflers - particularly larger ones - swinging around beneath the car and clouting the bodywork. Job done.
The moment the engine was started we were stunned at how much quieter the exhaust was compared to when it had the smaller dogleg unit. At idle, in fact, the car seemed only a bit louder than standard - but we'll come to that...
Results
No question - the free flowing 3-inch exhaust made a monumental difference to our '94 WRX. Before we tell you about the on-road and on-dyno improvements, though, let's take a look at the direct impact of the bigger system...
As previously, this graph reveals the relationship between boost, pre-turbo and post-turbo backpressure; the highest reaching plots represent pre-turbo backpressure, while the middle plots represent manifold boost pressure and the lower plots represent post-turbo exhaust backpressure. As you can see, there is zero post-turbo backpressure measurable below 3000 rpm (10 kPa backpressure was seen at 3000 rpm previously) and the maximum backpressure is 20 kPa - two-thirds less than when running the standard exhaust system. Pre-turbo backpressure, meanwhile, is also markedly reduced - the engine now feels up to 136 kPa total backpressure compared to our previous high of 183 kPa. Note that manifold boost pressure now comes on faster and spikes higher during spool-up but tapers off to around the same value as previously toward redline. What you can't see is boost peaks at 92 kPa at around 45500 rpm - a 9 kPa 'overshoot' coming from the new exhaust.
And now onto the end results...
Strapped on to Turbo Tune's AWD Dyno Dynamics chassis dyno, the newly exhaust'd Rex showed a 10 percent peak power gain over the previous 123kW ATW run with just the modified air intake; the new power tally is 135kW at the treads. Note that these tests are conducted in third gear. The blue plot shows that boost pressure rises more rapidly from the bottom-end and reaches a 13 psi peak before being hauled back to the 10/11 psi value that continues to redline (the same top-end boost pressures as previously). So, without any more boost being generated through the second half of the rev range, it appears that our 10 percent power gain must be solely thanks to the eased backpressure. Air-fuel mixtures remained ultra-rich as before - in the low 10:1 range and, at times, off the scale.
On the road, the car is more responsive and spools up more rapidly - we can hear the turbine whistling earlier than previously. With boost now spiking to over 90 kPa it's no wonder there's strong mid-range shove and, from thereon, the car feels smoother but its rate of acceleration tapers off toward redline. Again, boost pressure is the key factor here.
Against the stopwatch, the car has shown an improvement in both 60 - 90 and 20 - 100 km/h splits (in second gear). These times have been cut from 2.0 to 1.9 seconds and 6.2 to 5.9 seconds respectively. Note that these times are slightly slower than we recorded after we fitted the revised air intake - blame that on higher ambient temperatures.
Exhaust noise thanks to our second-hand Japanese muffler is now well in hand and easily bearable. At idle, our SPL meter needle wavers between 62-63dB with the flat-four beat and on revs to 3000 and 4000 we've recorded 91 and 92.5db respectively. Note that the SPL at idle about the same as the aforementioned HSV GTS300 and the rev-up SPLs - although still quite loud - are much quieter than with the previous muffler. Despite not having a centre resonator, there are also no exhaust resonances at any rpm/load combination.
All this for around $400 in parts plus $150-odd for fitment! Hmm, now how much power would we get if we hold more boost pressure to redline?
Contact:
Exhaust Technology
+61 8 8272 7500
www.exhausttechnology.com.au
Japanese Motorsport
+61 8 8260 6919
|