Stuttgart engineers did one heck of a job with the overall design of the Porsche 951. By increasing boost, matched with the correct amount of fuel, we can get easy horsepower gains. But trying to outdo these engineers in the hardware department comes with the risk of losing performance.
It's widely known that the stock 951 exhaust headers were designed for torque. The question for Project 951 was whether we would get more top-end horsepower and more midrange torque with a given exhaust header. In Part 4 (ec, 04/04), my 951 got the torque but lost a few ponies with the B&B headers (granted, the power came back with the addition of the B&B cat-back exhaust).
This month covers my quest to get some good top-end power gains without sacrificing the torque the car has now. Enter Speed Force Racing.
SFR, owned by car nuts Tim Richards and Keith Ta, is a tuning and fabrication shop that was started by Richards more than 7 years ago. Since then SFR has produced some highly anticipated products for 951 enthusiasts, including fuel systems, mass air flow kits, MAP sensor upgrades, intake manifolds, larger intercooler and turbo upgrades, supercharger kits, intake manifolds and exhaust headers. I decided to test SFR's Stage 2 header system.
SFR offers two different header systems for the 951. For just under a grand, the Stage 1 header system is offered in 304 stainless steel or 321 stainless (for a couple hundred bucks more). It will fit 951, S2 and 968 Porsches, and it's a 4-into-2-into-1 setup, similar to the factory unit and also the B&B header system that was on Project 951 up until this point. However, SFR uses a different merge collector, which reportedly keeps from killing top-end performance, unlike what I saw with mine.
In Part 4 of this project, I saw a dramatic increase in torque with the B&B unit, which was great. However, the slight loss in the top end caused me to search for a freer-flowing exhaust before possibly ending up with the factory pieces on the car once again.
The SFR Stage 2 header system is a completely different ball of wax. Richards designed this header system with the big power, full-T4 turbo 951s in mind.
"The 'Tri-Y' factory header is designed for torque, but this header frees up the top end so we can make power there," said Richards. This 4-into-1 header system has a merge collector that Richards feels is second to none. Fabricated by Burns Stainless, the merge collector was designed o that the exhaust pulses go through the equal-length runners and sequentially shoot through the collector for better spool-up of the turbo.
Although the Stage 2 system is also offered in 304 stainless for $1,800, I opted for the pricier 321, constructed of a steel made from nickel, titanium and a bunch of other alloys that allow quick cool-down. Because of this, there's no need to have them coated, but SFR will polish your 321 stainless headers for an additional $200 if you desire that extra "bling" under the tail. The Stage 2 system is also offered for the S2 and 968 Porsche models.
A header install can be a pretty big job, especially if the technicians have got the factory headers to remove. SFR technicians Ed Filburne and Brad East made it so that it will be much easier to remove the headers next time by replacing the studs with bolts, eliminating the need to maneuver the runners around the studs during header removal and installation.
SFR's Stage 2 header system includes a cross-over pipe designed to fit factory or aftermarket wastegates on 1987 and newer 951s that have the 90-degree bend that comes off the wastegate inlet pipe. For us '86 owners, however, SFR plans to fabricate the 90-degree bend itself so it can save its customers a hundred bucks over getting it from the dealer.
Welding this piece on the factory Inconel material is a touchy job, according to SFR. "Most guys who weld on Inconel find that it cracks after it cools. You have to know what you're doing with Inconel," said Richards.
The crossover and piping leading to the turbo (also known as the "up" pipe) is a multi-piece system held together with V-band clamps for a gasket-less seal. This setup also makes it much easier to access the clutch bell housing and helps complete such previously lengthy jobs as changing the dreaded oil pan gasket.
Initially Richards was concerned that Project 951 would be losing some low- and midrange torque with the 4-into-1 headers, since it is essentially supposed to shift the power curve to the right. However, SFR also offers its 3.0-in. turbo downpipe to free up the turbo. An aftermarket 3.0-in. exhaust for the 951 isn't a true 3.0-in. exhaust until the downpipe is changed to that diameter as well. As seen in Part 5 with the Lindsey Racing system, a 3.0-in. exhaust system can give tremendous gains throughout the rev range over stock. But with the SFR 3.0-in. downpipe, we were additionally hoping to see no losses in spool-up time with the new Stage 2 header system over the previous setup. That way we get the best of both worlds--an additional increase in top-end flow without sacrificing the midrange.
When considering its size, the $400 pricetag for the downpipe may sound like a lot, but in order to maintain a good seal Richards felt it was wiser to keep the costly flange in the design. Said Richards, "I designed it so the stock crush ring is used instead of the gasket, which holds a lot better than copper. This makes it tolerable to heat for years without leaking. Otherwise, the diameter of the gasket would be so close to the edge that the gasket would get blown out quite easily."
Before buttoning down the exhaust system, I gave Richards the okay to open up a tapered section of my Lindsey Racing piece since we were going to the 3.0-in. downpipe. The Lindsey piece was designed to fit with the stock downpipe, which is why it was tapered, but now it was the last piece to modify for a true 3.0-in. turbo-back exhaust system. Richards did this by simply placing it on his jig, cutting the tapered section off and TIG welding a 3.0-in. section back on.
At last the header system was installed and the car was ready to be tested. One thing I was already excited about was that SFR fixed a small yet audible exhaust leak around the number two exhaust runner (where I previously had to have a new stud drilled) thanks to the new bolts and gaskets SFR supplied. Now the engine compartment didn't sound like the little go-karts at Disneyland every time I blipped the throttle.
The sound out the tailpipe was noticeably different. Before, the car rumbled but it was somewhat broken up, even at light loads. Now, the new rumble was deeper, crisper and more pronounced, and it was time to hit the dyno (or in this case, dynos).
Speed Force Racing does its testing just down the street from its own facility at a place called the Dyno Shop. This place is literally what the name implies, housing no less than five dynamometers, four of which are load-based Dyno Dynamics dynos, including a four-wheel-drive unit, and an inertial Dynojet 248C. This facility offers all types of services, ranging from dyno testing and tuning, to performance parts installs, to boat tuning and smog services.
The Dynojet was used for this test since it's the more widely used dyno in the industry, and it's the one we mostly use. Apparently, during our baseline test with the B&B header on, the car had been running lean and peaked with a 277.9 whp run at 17 psi, as indicated by the Autometer boost gauge. A quick increase in fuel pressure brought the car's air-fuel ratio down a few points-to a high 11-and the car's power went up to 289.2 whp. We felt this would serve as a better and fairer baseline than the initial run.
The next afternoon, we returned with the new headers installed. Once the air-fuel ratios were checked and adjusted to a similar mixture as our baseline, we put the car through a fourth-gear pass. With the much-improved design, the power went up to 306.8 whp, with a peak gain of 31.5 hp and 27.3 lb-ft of torque at 6070 rpm. The improvements were significant, ranging anywhere from 10- to 30-whp increases from 5350 rpm through redline, essentially flattening the torque curve in the top end. Also worth noting, the new graph was considerably smoother than our baseline despite the shop temperature being about 7*F hotter than on the previous test day.
In the end, even though we were expecting the power curve to shift to the right, we actually didn't lose much midrange at all. This just went to show that the SFR Stage 2 headers, coupled with its 3.0-in. downpipe, can give the 951 much better top-end power without sacrificing anything below--and I'm all for compromise-free gains!
Author's note: Special thanks to Rob Bynum, Jon Cutico and Jonny Lacanilao for their assistance in getting Project 951 underway so quickly.
At a Glance
Installers: Speed Force Racing
Estimated Time: 14 hr @ $75/hr = $1,050
SFR Stage2 headers: $2,200
SFR 3-in downpipe: $400
Downpipe crush rings: $20
header gaskets: $10
I'll be talking about some of the new things you see on there some other time.
SFR's Stage 2 headers feature an equal length design that shoots the exhaust gasses sequentially through the merge collector for optimal flow to the turbo. The crossover and up-pipe sections are sealed with v-clamps. This system is a work of art.
SFR's Stage 2 headers feature an equal length design that shoots the exhaust gasses sequen
Installation was made much easier by removing the header studs on the cylinder head first. Should the header ever have to come back off, the header was secured with bolts for a much easier removal.
Installation was made much easier by removing the header studs on the cylinder head first.
Since I had an earlier model 951, the installation of the cross-over is a little more time consuming. A 90-degree bend from my factory crossover pipe was cut to use in the new wastegate piping and to be able to use the stock flanges for proper fitments.
Since I had an earlier model 951, the installation of the cross-over is a little more time
Richards spot-welding the piece together once the proper fitment was made.
Once the mock fitment was made, the unit was pulled off and TIG welded. Richards also took the piece and bead-blasted it for a cleaner appearance.
Once the mock fitment was made, the unit was pulled off and TIG welded. Richards also took
SFR's 3-in. downpipe was used to free up turbo exhaust gases.
Stock 2.5-in. inlet of the down pipe (left) versus the SFR 3-in. downpipe inlet (right).
The multi-piece crossover design makes installation much easier. Now the turbo uppipe can be mounted onto the turbo before reinstallation of the turbo.
The multi-piece crossover design makes installation much easier. Now the turbo uppipe can
Turbo installed. I used my old heat wraps to keep the heat from rising up into the intake. Just like before, the hot side of the turbo will also be heat shielded.
Turbo installed. I used my old heat wraps to keep the heat from rising up into the intake.
This is the tapered section of the Lindsey test pipe that was designed to fit the factory downpipe. Since I'd now be using SFR's new 3-in. downpipe, SFR hacked this section away and TIG welded another 3-in. piece to keep the full turbo-back exhaust a true 3-in. system.
This is the tapered section of the Lindsey test pipe that was designed to fit the factory
SFR Stage 2 headers are a tight fit but they fit nonetheless. This is the color the 321 stainless steel turns into after it's been heat cycled-sort of a dark gold color.
SFR Stage 2 headers are a tight fit but they fit nonetheless. This is the color the 321 st
HP: At 17 psi, Project 951's first baseline run with the B&B headers, LINK AFM, Kokeln Stage 5 turbo and Lindsey Racing exhaust was pretty lean. The red graph shows the increase in performance when Richards increased the fuel pressure a bit to fatten up the mixture, which we used for our new baseline. SFR Stage 2 header and down pipe combo (blue) showed a gain of 31.2 hp at 6070 rpm without changing boost. These runs were also done with 91-octane pump fuel.
HP: At 17 psi, Project 951's first baseline run with the B&B headers, LINK AFM, Kokeln Sta
Torque: With a peak gain of 27.3 lb-ft of torque over our new baseline run (red graph), the torque curve was essentially flattened up top.
Torque: With a peak gain of 27.3 lb-ft of torque over our new baseline run (red graph), th