Racing Exhaust System - Header Tech

Since we are on the subject of collectors, let's take a look at some of the more high-tech stuff available. First, the goilet. This is basically a spike that streamlines the entry of the primary into the collector. At first you might think that a goilet would most likely improve top-end output. This, as often as not, proves untrue. The gains with a goilet are small-typically in the realm of 3-5 lb-ft. The gains usually show up in the low- to mid-speed range, and by the time peak power comes around, any advantage is usually gone. If extra top end is seen, it is often because the collector is just a shade too big and the goilet is making the collector look slightly smaller.

Merge collectors look pretty racy (Graphic 1) and were very popular with the Nextel Cup guys until the advent of 4-2-1 collectors (Graphic 2). A merge collector usually has an effect on low- and midrange torque, although improved top end is not out of the question. My own tests show that the merge collector makes a big race cam look a little smaller at low rpm. At a thousand to fifteen hundred rpm before peak torque, a merge collector can (but not always) pull in some 10-15 lb-ft more. Peak torque may also be a little better, but in the main, peak power stays about the same as an optimally proportioned regular collector.

Take a look at the exhaust on a current Cup engine and you will most likely see a 4-2-1 collector. These nearly always bump torque in the rpm range prior to peak torque. Improved top end really depends on how near optimal the more conventional baseline system it's compared against may be. It would appear that there is some potential for improved top end, but I would like to do a lot more testing before I go too far down the road on this subject.

We have looked at the principle dimensional aspects of an exhaust system, but a substantial edge can be taken off a race car's performance if the consequences of heat and muffler flow are not addressed.

First, heat. These days, it is relatively common to see headers and systems coated with a finish that keeps the heat in the system rather than letting it raise the temperature of the area around it. A lot of this is done just for driver comfort. The question remains, though, as to whether keeping heat in the system helps power. Over the last five years I have worked closely with Calico Coatings in Denver, North Carolina, and the results of quite a few before-and-after exhaust tests show that coatings produce increased output, even on the dyno, where heating of the intake air is not an issue. Assuming an operating range of 3,500 to 7,000 rpm, we find that a thermal barrier coating improves output up to, about, or just past peak torque. From there it usually stays near the same.

What the coating seems to be doing is making the header act as if it is slightly smaller. This might be the case, as holding the heat in makes the volume of exhaust greater, which increases the velocity. This seems to be born out by the fact that headers that are slightly too big prior to coating produce greater output throughout the rpm range. To be able to establish the effects of the increased exhaust velocity in the header and the effects of reduced engine compartment temperature (and hence cooler carb air), a chassis dyno test is necessary. Figure 6 shows the results of such a test.

As time goes by, more tracks are pressured into mandating noise limits. Adding mufflers without knowing what is needed is the biggest and most widespread source of power reduction seen at the track. Maybe this warrants a story in its own right, but two factors can save the day. The first thing to avoid is making the mufflers look like an extension of the secondary tuned length. For the tuned length to remain unchanged, there must be a large increase in cross-sectional area right at the collector/muffler junction. You have to re-evaluate the secondary tuned length if a straight-through glasspack is used, as a conventional glasspack stuck on the end of the collector will make that critical tuned length appear longer.

The flow of any muffler system used is also a factor. Too little flow causes backpressure. If anyone tells you that a little is good for the engine, check to see if he or she is a salesperson for bad to indifferent mufflers. What the engine likes is zero backpressure-don't let anyone convince you otherwise. To get zero backpressure, the engine needs at least 2 cfm (but preferably 2.2 cfm) per open exhaust horsepower. So, to avoid any loss of output due to muffler-induced backpressure, a 500hp engine needs one or a pair of mufflers that deliver at least 1,000 cfm at 1.5 inches of mercury (1,170 at 28 inches H2O).

Before wrapping up, I would like to thank the guys at Hooker and Kooks. Their patience with my request has made it possible to do all these tests.

Since the diameter of the collector/secondary hinges on the selection of the primary diameter, it's handy to have some idea of what size is initially needed without having to resort to testing a number of header configurations on the dyno. This chart will get you close to the optimum size needed to put the torque in the best place in the curve. The pipe size required is influenced greatly by the exhaust port flow at full valve lift. With this number we can get the size required from the graph. The blue line represents a size for a superspeedway car, where top end is the only consideration. The red curve is for intermediate-length tracks, which have corners that are fairly wide open compared to the length of the track (that's almost everywhere except Martinsville), and the green curve is for shorter tracks or those with a tight turn in relation to the overall track length.