Vadim Fedorovsky on Exhaust Design
Vadim Fedorovsky on Exhaust Design
By Vadim Fedorovsky, [former] SHO Shop Proprietor
We have all had our own experiences, listened to many 'experts', seen a lot advertisements promising big HP gains and spent a lot of money on the exhaust - only to be disappointed later.
Let's start at the head. As two 1.18 in. exhaust valves open, a charge of hot combustion gases rushes toward the exhaust manifold. The charge is at approx. 2000 deg F and under 75-90 psi. We have flowed the heads on the Superflow 1200 flow bench and the exhaust port flows 190 cfm @ 0.350" valve lift. These numbers can be improved with hand porting and Extrude Honing up to 225 cfm. Either number is quite adequate, (6 x 190 = 1140 cfm total). The Extrude Honing improves flow at all lifts, helping low-end torque.
The next piece is the exhaust manifold. If you have ever had a chance to see one out of the car, you can really appreciate Yamaha's workmanship. The unit is cast and very smooth, with no casting flash anywhere. On the flow bench, exhaust manifold runners flow anywhere from 230 to 240 cfm. They are oval in shape (1.75" x 1.6") and vary in length from 13-15 inches. For most purposes you can call them equal length. As the exhaust pulse reaches and passes the collector of the exhaust manifold (the collector is where all three runners merge together) it creates slight vacuum in the other two runners. At a certain RPM the arrival of this pulse will coincide with exhaust valves opening in one of the other two cylinders and by creating vacuum in that runner will help the newly created pulse flow down that runner, helping the intake charge fill that cylinder and making more power.
What I have just described is called tuning. Now in reality, the process is a lot more complex, with a large number of variables involved. For our purposes all we care about is the exhaust runner diameter and length. The factory manifolds are fairly large in diameter and short in length. They are tuned for high rpm (7000-8000), most likely remaining on the engine due to its previous intended application [ed. note: the GN34].
99.99999% of SHO owners have stock manifolds on their SHO's. Ted Breaux and I have a 36 inch long set of headers built by JBA on our SHOs. Mine are 1.875 inches in diameter, Ted's are 1.75 inches. Neither one is what I would call a bolt on item, as they require a completely new exhaust system to merge into, and are loud. These are race parts! As far as performance, on my SHO they boosted mid-range torque at the expense of top end power. With my 89 all apart at this time, dyno testing will have to wait.
After the exhaust manifold, we have the dreaded Y-pipe. I think most of you have seen it on your SHOs and know what I mean. It is designed purely to get the cats hooked up to the car in the cheapest way possible. Now after the exhaust pulse goes through some pretty restrictive piping [and that damned 90* turn -ed.] into the cat, the cat spreads the pulse across its honeycomb structure as it does its work. By the time that pulse reaches the end of the cat it is considerably slower and cooler. With other pulses coming quickly behind it, the pulsing nature of the exhaust gases becomes more of the constant flow. At this point all that matters is how quickly the exhaust gas can escape from the Y-pipe. in other words it wants to see as little back-pressure as possible.
Now if we could discard the cats and use true duals with a properly setup crossover, we could create some additional tuning as we merge pulses from each side of the engine through the crossover. However, IMHO there are two big problems. First of all, since the engine is transversely mounted, the front pipe will be longer than the back pipe. The crossover will have to be welded on diagonally between two pipes in order for it to be in the same spot on both sides - nearly impossible given the confines of the car. Secondly, the diameter of the pipes, their length and crossover location would have to be tuned to the stock intake and cams design. The minute you decide to go with Extrude Honing the manifold and Stage I/II cams, your tuning will have to be redone.
Besides, fitting two pipes into a tunnel designed for one creates numerous problems of their own. Now Ted, you are going to list me a lot of people who have put true duals on. And are happy with them. Unfortunately, they don't call you a month later. They call me and order our Y-pipe and cat-back. Fit and noise problems are listed as major complaints. As far as power gains of dual vs. single, personally I have not seen any. One of the best installations of duals, that I have seen, on one of my customer's SHOs, was done in T304 stainless with mandrel bends (cost close to $2500) and did not run any stronger than any of our cars with our exhaust on it. The amount of fuel the LPM needed to get that SHO to run at the same air/fuel ratio was the same as ours, telling me that the car did not breathe any better. And you will not pass smog check in many states.
For the cat-back we use 2.5" single mandrel bent pipe that goes through a resonator, over the rear suspension and then splits into two 2.5" branches with Edelbrock RPM mufflers. For a while we used custom made mufflers that were supposed to be as quiet as Edelbrocks. Unfortunately, our fabricator could not maintain the consistency in sound, so we are going back to Edelbrocks.
You have to look at the engine as a system. The intake runners, cam profiles, exhaust manifolds and exhaust system will all have to be tuned to work together as a SYSTEM. There is no one single problematic component that, when improved, will make 30-40 HP. Back in '94 [ye Gods, 16 years ago -ed.] we spent 2 weeks and $2000 in dyno time, testing a lot of components. I wish I knew then what I know now about chassis dynos. If anyone quotes you dyno numbers obtained on anything but a DynoJet 248 Inertia Dyno, do not believe them. I can honestly say I have tried almost all chassis dynos in the LA area, and with the exception of DynoJet, they are about as useful at measuring horsepower as an hour glass is at measuring acceleration. Using a water-brake to load the engine, they require a constant attention from the operator. They are easily manipulated by the operator to make sure you are happy, very inconsistent and are also hard on the engine. In contrast, the DynoJet uses inertia rollers to measure HP. The measurement is done by flooring the car at 2000 rpm and backing off the pedal at 7000 rpm. This is all operator involvement that is needed. The computer calculates TQ and HP curves and prints them on the screen. The dyno is very repeatable, within .5HP and very simple to operate. The first time I got a SHO on one of those, I ran it 3 times in a row. All I saw was a single curve for HP and TQ, all 3 runs were the same. Most performance magazines, like Motor Trend, Turbo, European Car and others now only use DynoJets for their testing. No more BS.
Last week I finally had a chance to test some of our products on a DynoJet. Being only 2 miles away makes it an even easier proposition. The results surprised me. They were totally the opposite of I expected.
Lets start with a stock 89 SHO. 102,000 miles, needs valve adjustment and new plugs and wires. [No graphs available, use your imagination. ed.]
Two Torque peaks due to two sets of runners. 193 lb/ft @ 2900 rpm and 188 lb/ft @ 4800 rpm. HP peaked at 183HP @ 5900 rpm.
Being short on time we installed the next simplest bolt-ons: An LPM. With MOTEC reading air/fuel ratio a bit rich, we leaned the fuel curve out, added timing and changed secondaries opening to 3650rpm instead of factory 3950rpm. 1 lb/ft of torque gain across the RPM range. How can it be? The car feels so much more responsive on the bottom end and with secondaries opening sooner it feels much smoother. Since we are testing at WOT where the airflow is what makes the power, factory calibration is perfectly adequate. On the street, where most of the driving is done at part-throttle, the added timing and leaner fuel curve makes the car feel more responsive and therefore more powerful. A valuable lesson; how your butt feels is not always real. :-)
Since the car is stationary on the dyno we did not test our cold air induction. Opening the air box and letting the SHO breath through a K&N panel filter did not show any gains. The 80 mm mass air was the next simplest bolt on. Since the exhaust was all stock I did not expect any gains. To my surprise we picked up 3HP from 4500 rpm on up. Since the LPM was off, the air meter was mismatched and we lost 10 lb/ft of torque below 3000 rpm. The MOTEC showed lean and sporadic air/fuel mixture. Plugging in an LPM with a correct transfer function for the air meter and fuel curve brought the torque back up on the bottom end and added another 3 HP on the top, for a total gain of 6HP and 4 lb/ft of torque from 4000 rpm on up. The whole curve has come up.
The next simple bolt on were the underdrive pulleys. This was the biggest surprise of all. The theory says 3-5HP, mostly at higher rpm. The car picked up 3-4 lb/ft of torque from 2000 rpm on up. We gained another 7 HP at 6000 rpm.
Next was the Y-pipe. With an LPM recalibration, this was the biggest gainer of them all, with 11HP gain at HP peak. We are now at 199 lb/ft @ 2900 rpm, 197 lb/ft @ 4800 rpm. HP is at 207HP still peaking at 5900 rpm. What is more interesting is the difference between stock and current configuration at 7000 rpm. Where stock car only made 106 lb/ft of torque and 141 HP, it was now making 143 lb/ft of torque and 190 HP, a 49 HP difference. And the cat-back was still stock.
